Kimini

Kimini
How to design and build a mid-engine sports carfrom scratch
Published and edited by Kurt W. Bilinski
Cover artwork by Ann Bilinski and Kim Nguyen
Cover and back photography by Max Nealon (Professional
photographerhttp://maxandpj.com/)
Copyright 2007 Kurt W. Bilinski
All rights reserved. No part of this book may be used or reproduced in any manner whatsoever
without written permission, except in the case of brief quotations embodied in critical articles or
reviews.
Published 2007 by Kurt W. Bilinski
Kurt Bilinski can be contacted through his web site: www.kimini.com
First Edition: 2007
Printed by Lulu.com in the United States of America
Text copyright 2007 by Kurt Bilinski
Photographs copyright 19952007 by Kurt Bilinski, Max Nealon, and Robert Bowen
Cover photography by Max Nealon, www.maxandpj.com
Dedication
To my wife, Kim, whose love, tolerance, and support made both my car project, and this book,
possible. And to my best little buddy, Cooper, a great listener and supporter of all things fun.
Acknowledgment
I would like to thank the people who encouraged, educated, and enlightened me in my quest. Thanks
to Alan Brickey, who generously shared his years of experience as an expert machinist, fabricator,
and mechanic. When I couldnt figure out how to make something, he always had suggestions. Thanks
also to Armen Assatourian, Steve Coe, Dave Hyatt, Sumant Iyer, Joe Malinowski, Ryan Manes, Cecil
Napoli, Dennis Palatov, and Ron Schramm for offering encouragement, support, and light-hearted
ribbing.
James Kopet gets special mention as copyeditora challenging task with me as writer! I also want
to thank Bob Bilinski, Chris Lynch, Craig Pond, and Ron Schramm for reviewing the manuscript,
biting their tongues, and kindly correcting my horrendous abuses of the English language.
And finally, very special thanks to Lee Kaiser, whose creativity, support, and generosity helped
make it all possible.
Statement of Non-liability
I feel if a person makes a decision, he or she intelligently accepts responsibility for his or her own
actionsunfortunately this no longer appears to be true. Some people cannot accept responsibility,
instead suing, living in denial of any responsibility for their lives. Since its impossible to accept
responsibility for an irresponsible person, I have to say the following:
Building a car is dangerous; driving a car is dangerous; accidents happen. Any car may kill you
if you drive it long enough, including this one.
I am neither a professional automotive engineer nor a lawyer. Carroll Smith said it well in his
book, Engineer to Win:
If, while attempting to apply any of these ideas, procedures, or advice contained in this book, you
should come unstuckor your racer should breakit will be as a result of your own conscious
decision. I disclaim responsibility for your actionsand for your accident.
Preface
Building a car is the ultimate lesson in compromise.
Everything is connected to everything else.
How do you eat an elephant? One bite at a time.
This book presents one mans journey through the entire ordeal of building a car from beginning to
end.
My dream has always been to build a sports car and drive it to Laguna Seca Raceway, near San
Francisco, racing during the day and camping at the track at night, just like the old days. Ive always
admired the passion of early aviators, who flew from town to town, landing in fields in the evening
and sleeping under the wings of their planes, and of the amateur racers in Europe, who drove their
cars to events, raced, and drove home. Those days are gone, but with enough work, that passion and
adventure can be brought back.
What this book is
Because you have access to different parts than I do, this is not a cookbook, a hold-your-hand, step-
by-step process for building an exact copy of my Miniits something far more useful.
This book walks you through the process of designing a car based upon the parts you have, rather
than forcing you into using parts you dont have, dont want, or cant afford. This is a cookbook only
in the sense that it lays out a map of how to create your own recipe.
This book covers the research, design, and construction of a scratch-built, street and track, one-off
custom sports car, a mid-engine, tube-frame, composite-shell Mini. Starting with a simple wish list
and progressing through goal definition, chassis and suspension design, and construction, the entire
ordeal is covered, concluding with driving observations. In addition, the thinking behind the decisions
why something was done a certain wayis described. It shows the steps, mental and physical,
necessary to create a car from scratch and the philosophies of doing soand actually finishing it!
Also covered is what I would do differently next timesomething missing from many books!
This book was written to show that, yes, you too can build a car. It reminds me
of the scene in The Empire Strikes Back, where Luke says to Yoda, I dont
believe it, and Yoda replies, That is why you failed. If you dont think you can
build a car, then you have already failed, but if after reading this book you feel
that you canyou can indeed.
What this book is not
This book does not spend pages showing hand tools, explaining how to solder or weld, or how to
use wire cutters or screwdrivers. It does not teach how to drill holes, use a hacksaw, install rivets, or
paint. These topics are covered far better in books written by experts in these skills.
I wish I had a book like this when I had started so much so that I decided to give something back
to the car community.
Seeing people changing suspension components without understanding the
consequences or attaching suspension brackets anywhere convenient, without
understanding, was another big reason.
Ten years of learning and building is distilled into one book that can help you on your own
challenging, formidable, yet rewarding journey of designing and building your own sports car. If after
reading this book you succeed in building your dream, I will have succeeded. I think youll enjoy
watching over my shoulder as I figure it out, and regardless of what you decide, I hope this book
provides some entertainment!
The me-monster
While writing this manuscript, the word I kept showing upa lot. So much
so, that the book started resembling a self-serving, ego-inflated, pompous, self-
congratulatory tribute that wasnt the intent at all. It concerned me enough that I
tried rephrasing sentences to avoid the I word, but all that did was make the
sentences wordier.
A buddy said, Dont worry about it, you designed the car, built it, and drove
itjust use the word. It is the most efficient way to explain things, so I did. I
write this hoping that people wont read this book and think, He sure is full of
himself. The usage is to simply convey what was done and by whom.
Either way
This book will save you money. If you decide to build a car, it will show you what to do, and what
to avoid. And if, after reading this book, you decide, Gee, this guy is nuts, way too much work, this
book has saved you lots of time and money!
Kimini 2.2 Specifications
Kimini: Table 1
Weight: 1595 lb. dry
Weight distribution: 41F/59R
Height: 44
Length: 115
Width: 64
Wheelbase: 80 (stock Mini)
Front track: 57
Rear track: 54
CG height: 16
Front suspension:
A-arm, upper rocker arm, inboard Koni double-adjustable shocks,
cut-down
280ZX uprights
Rear suspension:
Trailing links + lateral links, Koni double-adjustable shocks, custom
uprights
Wheels : Taylor LiteSpeed, 13 x 7 front, 13 x 8 rear
Steering: Triumph Spitfire rack, stock
Tires: BF Goodrich R1, 185/60-13 front, 215/50-13 rear
Engine: Mid-mounted 1995 Honda Prelude VTEC, 2.2 liter
Custom header and exhaust
Custom intake
8 lb. Fidanza aluminum flywheel
Crank HP: 200
ECU: Stock Honda P13 (for now)
Transmission: Stock Honda 5-speed + Quaife limited-slip differential
Axles: Custom, Summers Brothers
Hp/wt: 8.0
Hp/ton: 250
Fuel tank: 10 gal. Fuel-Safe fuel cell
Radiator: Griffin Ford unit
Brakes: Front: Wilwood calipers + stock 10 Nissan 280ZX discs
Rear: Mazda RX-7 calipers + cut down front Honda Prelude discs
Brake master
cylinders:
Tilton 0.813
Seats: Cobra
Steering wheel: Momo
Instruments: Autopower + SPA tach/speedo
Lights: Stock Mk1 Mini + LED center-mount brake light
Windows: Stock Mk1 Mini windscreen, scratch-resistant Lexan for all others
060 mph: 3.9 sec.
Quarter-mile: 12.7 sec.
Top speed: 132 mph (est.)
Build Timeline
Table 1
Spring 1996: Lee Kaiser creates molds based on a Mk1 Mini.
Summer 1996: First (and only) carbon shell is pulled from mold.
Winter 1996: Decide to build a mid-engine sports car based upon this shell.
Ordered design books. Started reading and taking notes.
Spring 1997: Buy engine, start design.
Winter 1997: Design completed, start mock-up of chassis inside carbon shell.
Summer 1997: Build chassis table. Buy TIG welder.
Fall 1997: Place drivetrain and wood chassis mock-up on chassis table.
Winter 1997Spring
2000:
Finalize placement of all major components inside wood mock-up.
Finalize suspension pickup points.
Take full measurements off wood mock-up, and then turn it into
firewood
Summer 2000: Start steel chassis construction.
Fall 2001:
Basic chassis complete, start fabrication panels and drilling rivet
holes.
Spring 2001:
Cooling system + lines, fuel accumulator and pumps. Endless rivet
drilling
Winter 2001: Remove chassis from chassis table, sell table.
Spring 2002: Electrical work, start engine!
Summer 2002: Engine bulkhead fabrication, engine tray design.
Fall 2002: Transmission tunnel fabrication, foot-well paneling.
Initial front suspension upright fabrication.
Brake master cylinder mount.
Additional front bulkhead fabrication.
Shifter design and fabrication.
Gas pedal fabrication.
Winter 2002: Throttle cable install.
Mount radiator.
Radiator ductwork design and fabrication.
Brake and clutch line design.
Spring 2003: Brake and clutch lines in.
Seatbelt anchors.
Dash instrument placement.
Dash framework.
Finish center tunnel panels.
Wire engine computer tray.
Fabricate engine sensor mounting brackets.
Begin wiring.
Summer 2003: Wiring continues.
Wire dash, electric fan switch and relay.
Choose taillights.
Buy fuel-rail.
Build battery box.
Finish cooling system.
Get alternator working.
Rebuild brake calipers.
Install longer wheel studs.
Cut down Honda brake rotors.
Rear toe links, inboard suspension mounts.
Header design and fabrication.
Finish exhaust system.
Fall 2003: Fill cooling system, brakes, and clutch.
Make suspension bearing spacers.
Have injectors cleaned.
Kimini now sitting on her wheels.
New sparkplugs.
Wheel alignment.
Finish fuel system.
First drive, chassis only.
Fix fuel accumulator and MAP sensor.
Winter 2003: Add door bars.
Initial fit-up of carbon shell to chassis.
Endless composite work.
Fix rear suspension.
Spring 2004: Order new axles.
Design rear diffuser.
Front firewall/bulkhead.
Install shell mounts around windshield.
Order door parts.
Fire extinguisher mount.
Shell mounts at front end.
Finish front bulkhead.
Doors.
Summer 2004: The never-ending doors.
Endless bulkhead fabrication to seal the chassis to the shell.
Continue fitting shell to chassis and making Dzus mounts.
Fix leaking Tilton brake master cylinders.
More shell-chassis flanges.
Determine parting line of engine cover.
Endless composite work.
Front grill fabrication.
Work on engine cover.
More shell-chassis mounts.
Fall 2004: Fabricate front bumper.
Radiator inlet and exhaust ducting.
Correct top roll-cage tube.
Rearview mirror mount.
Electrical plug mounts front and back.
Dumb turn-signal switch.
Disassembly for paint!
Endless welding.
Powder-coat chassis.
Finish aluminum panels and suspension arms, send out for paint.
Winter 2004: Start final assembly.
Coolant pipes, fuel cell.
Bulkhead window.
Rivet floor.
Install engine bulkhead.
Install engine.
Engine tray.
Rivet in more panels.
Install front suspension.
Install rear suspension.
Spring 2005: Install pedals.
Shorten wire harness.
Order all Mk1 parts, lights, door rebuild parts, etc.
Continued electrical work.
Start engine again idle problem.
Fill and bleed clutch.
Fill and bleed brakes.
Install seats and seatbelts.
Corner-weigh chassis.
Finish shifter and suspension install.
Design and fabricate air filter box.
Doors againick.
Start composite work.
Summer 2005: Endless composite work.
Test drive, without shell.
Take passenger compartment to body shop.
Fall 2005: Shifter boot.
Brake proportioning valve installed.
Did I mention the doors again?
Nose composite work.
Permanently install painted passenger compartment on chassis.
Fabricate Lexan windows.
Send engine compartment cover to body shop.
Mount *$

!%@ doors.
Install Dzus fasteners on engine cover.
Prepare nose to go to body shop.
Mount windshield.
Mount taillights.
Wire engine cover for taillights.
Install Lexan windows in engine cover.
Winter 2005: Install door parts and Lexan windows.
Install doors.
Install headlights.
Insurance issues.
Wire nose for headlights and turn signals.
Mount Dzus fasteners on nose.
Add clear paint-protection film.
Mount radiator inlet grill.
Mount radiator exit grill.
Cut access ports in engine cover side windows.
Engine cover lower grill installed.
Mount rear Lexan window to engine cover.
Fabricate air-box cover.
Fuel vent hose, rubber fender rock guards.
Fix brake caliper leaks.
Found plugged up line in cooling system.
Done!!! Take car to first car show and winabsolutely nothing!
Introduction
Are you unhappy with overpriced, unexciting production cars? Does the idea of driving one of
600,000 identical cars leave you uninspired? Looking for that special car that brings back the passion
you remember from the old days but without the reliability problems? Remember driving cars that
had character? Unimpressed with the volatile kit-car industry? Ever think you could do a better job of
building one yourself? You can!
Building your own car from scratch means:
Painting it any color you want instead of only what the manufacturer offers.
Not accepting a companys idea of what constitutes a sports sedan.
Not having them choose the engine, steering wheel, and seat youll like.
Truly owning a one-of-a-kind vehicle.
Being able to say you are designer, builder, and driver.
Consider building your own car, exactly the way you want, for about the same price as a low-end
compact car. If you enjoy working with your hands, are handy with tools, and are persistent, you can
build the car youve always dreamed of.
I hope that this book serves as proof that anyone who really wants to can indeed build a car.
1: Why
I have always been fascinated with cars; my earliest memory was passing by the Del Mar,
California, fairgrounds and seeing strange cars, the helmeted drivers sitting upright in the wind,
sawing at the wheel, just barely missing the hay bales with their skinny tires as they flew past.
In hindsight, it must have been a vintage race, since that style of car predated
my birth by 20 years.
(These gray boxes provide additional commentary on a subject, often with the
benefit of 10 years of hindsight.)
Sadly, my father didnt stop that day, but I never forgot it. Later, I discovered I grew up within 100
miles of some of the most exciting racing to ever exist in North America. During that time, the CAN-
AM series and the Chevy Camaro and Ford Mustang factory race teams were fighting it out at
Riverside Raceway. I missed out on all of it, completely unaware of the history passing me by; only
years later did I learn Id lived through some of the most exciting racing everand never saw a
single race.
While I went to the local drag strip and watched the 70s muscle cars, they never really interested
me; I only became interested in cars after graduating from high school. My brother, Scott, bought a
1971 Datsun 510, and while we werent aware of it at the time, that sedan forever changed how we
looked at cars.
I soon got my own car, a 1971 Datsun 1200, and slowly began tinkering with it. (The one here is
my friend Steve Coes old car, but mine closely resembled it.) While underpowered, it was very
light, providing an excellent lesson in the importance of power to weight. Other cars followed, with
Scott getting a 1972 Datsun 240Z, followed by a Mazda RX-3 with a bridge-ported 13B, while I went
through several more 1200s.
Scott soon discovered the local autocross event in San Diego, and once we became involved in
that, it was all over. Gradually, I started modifying the Datsun, ending up with a 1500 cc twin-side-
draft A15 engine with full-race suspension. Every day I looked forward to diving into circular
freeway ramps, made all the more fun when some sports sedan followed me in at full tilt, yet the
sedan was never there at the exit of the turn, all thanks to the 1200s light weight and sticky tires.
Initially the car was used for autocrosses, and later, time trial events, now called High Performance
Driving Events (HPDE), time trials, or simply, trackdays. Once my brother and I drove in a
HPDE, it was very hard to go back to autocross!
While Im still interested in autocross, its hard (for me) to rationalize the
time. That is, nearly an entire day is consumed for approximately 180 seconds of
on-track time. I still autocross but stick to practice events, which provide more
track time. HPDE events give far more on-track time than autocrossfor a price.
But the price per lap is much less (and you dont have to work the course!).
On the other hand, autocross is invaluable for learning car control. At high speeds, the Datsun
would begin to rotate (spin) much slower than it did at slower (autocross) speeds, so it was much
easier to maintain control. Experience gained at the autocross was invaluable once I was on the road-
racing tracks, and it saved my butt many times.
Why didnt you ever road race?
I dont have the killer instinct. I also didnt want to spend hours fixing body
damage caused by people who forced the situation. They might be able to drop
their cars off Monday morning for repairs, but I had to fix mine myself.
The rampant cheating was also a big turnoff. I remember watching a bunch of
Datsun 1200s racing, and it was blatant who had the (illegal) close-ratio
transmissions, just by listening. Yeah, they could have been protested, but all that
would accomplish would be getting a $500 love tap during the next raceno
thanks. This was supposed to be fun, not something that would require hours of
bodywork after every race.
Instead I much preferred HPDE events, where anyone could show up driving
virtually anythingno rules, no protestsand just have a great time. I liked the
variety, with all the different types of cars. I liked the fact that I was racing the
clock, not having to decide whether to push someone off in order to win a $10
plastic trophy.
Modifications to the Datsun continued, with the predictable result that it ended up as a hardcore
road-race car I happened to drive on the street. The ultimate modificationnever fulfilledwas to
install a Mazda turbo-13B and transaxle in the backseat. This was before front-wheel drive (FWD)
drivetrains were widespread, and the existing ones had very little powerI would have to roll my
own. It meant buying a very expensive transaxle for the 13B that I simply couldnt afford. That
unfulfilled project never quite left my mind. The car was later sold when various parts I couldnt
replace started wearing out, and emissions tests were getting too difficult to pass.
College, marriage, and a full-time job followed, but I never forgot those days. While Im an
electrical engineer by training, I was (and am) very interested in all things mechanical. I kept up on
what the latest cars were, but since I couldnt afford a real sports car, it was a little depressing. The
only cars I could afford were too big and heavy, and these would get trounced at the autocross by
older, smaller, lighter cars. Seeing this confirmed that a light car was the secret, but since there were
no lightweight cars I could afford, it didnt look promising.
Modified cars
Most of us get started with cars by getting an old car, then spending many weekends making it run
as well as possible. Soon, we start adding things: intake, exhaust, wheels, tires, and suspension. Or,
we buy a new car and add the very same parts. The problem is, its still the same car. Sure its faster
and corners better, but it still has the original design limitations, be it drum brakes, leaf springs, or
simply being too heavy. With the Datsun, I could only go so far to correct these inherent limitations
before turning it into a big expensive mess. The absolute worst investment is buying an aftermarket
go fast part. Lets say someone buys a supercharger for his new car; the second it is purchased, the
value drops by half (just try selling a used new one). Next, he pays hundreds of dollars to have it
installed, because the engine compartments in new cars are so packed that no one wants to install it
himself. So hes lost money buying the supercharger, and its still in the box, losing more on
installation. The owner loses reliability, gas mileage goes down, and people wonder whats wrong
with the engine (because of the supercharger whine). Later, he tries to sell the car with the
supercharger still installed (because he doesnt want to pay someone to remove it) and loses
thousands more because buyers will instantly know the car has been abused. I knew a fellow who
bought a brand-new Camaro Z28 and added $15,000 in aftermarket parts. The car could do the
quarter-mile in the 11-second range and could corner pretty well, but it always had the one
shortcoming most difficult to fixweight. To add insult to injury, what did adding $15,000 of
performance parts do to the value of the car? It became less valuable than if it were never touched.
Of course we dont modify cars to invest in them; we do it because its fun. The point is, if youre
going to pour thousands into a car, make it count. Putting a supercharger on a 3400 lb. car will help,
but as the saying goes, its like putting lipstick on a pig. At some point, it would be better to just start
over. Now, you can carefully direct the money to go exactly where you want, rather than spending
money undoing something you paid money for as delivered from the factory. At least this way you end
up with a car with no built-in design limitations, most notably weight; its the big killer when it comes
to performance. Hot-rodders figured that out a long time ago.
About two years into the project a friend asked, Why dont you just buy a
third-generation Mazda RX-7?; you could have a real sports car right away. A
good question, and one I had to think about before answering:
Because I would find shortcomings in it: small brakes (for me), reputed
fragility (by some), poor paint quality, trim falling off, and turbo fires.
Because I would not leave it alone. I would put on large brakes, wheels, and
tires. I would probably up the boost, chip the computer, and add exhaust, a roll
cage, and belts.
Because by the time it was done, a fortune would be spent on a car, now worth
little, ironically because of the modifications. This is because of the real and
perceived wear on the car; anyone buying the car would know it had been raced
its hard to hide a roll cage. Its literally throwing away money to fix up one,
yet shortcuts cant be taken if it is to be reliable and safe for competition. If Im
going to have something thats not worth much when finished, it better be just the
car I want for competition. The RX-7 is not it, no matter what I did to itdue to
its weight.
Because building my car is a diversion from work, something I can see
progress at each step something to look forward to when its finished.
Because building my car is relaxing and rewardingno schedules.
Because I want to see if I can build something to compete against, and beat,
cars that cost five times as much. Im betting I can build something quicker and
faster for about $20,000.
Because I get tired of the compromises built into every mass-produced car,
such as too much weight, small brakes, and poor tires. Why pay for something
inadequate, then spend money replacing these same parts with the good stuff?
Because I gain skills I never would have acquired otherwise. After I am done,
maybe I could get a job building race cars (or not).
Because I always wanted to build a car starting from scratch.
Because I like working with my hands.
Because I wanted to, thats why.
Hot rods
I define a hot rod as a 30s to 60s American car with a large V8, side pipes, big rear tires, skinny
front tires, and lots of chrome. A hot rod gets a little closer to my idea of what I was looking for,
although with a minimal chassis and large engine, they can be very flexible and dont go around
corners very well. Of course, thats not what theyre for; it all comes down to what you want to do
with it. If cruising is your game, then youre all set, but if you plan on any sort of spirited driving that
involves corners, then you have to look further. The problem is, most hot rods are intended for
showing off, not handling. The extent of competition is usually infrequent runs at the stoplight grand
prix. The last thing you would want to do in a hot rod is to corner at any sort of speed. If any type of
competitionor even hard drivingis desired, 3 wide front tires, 500 hp, and drum brakes is not a
good combination. On the other hand, something like a Ford Cobra, Lotus Super 7, or GT-40 comes
closer to my definition of a sports car. Since the originals are worth a fortune, I became interested in
kit cars, which come close to the originals without the expense.
Kit cars
Here was a possible solution that would combine low weight and low cost; it could be just my kind
of car, so I started researching. There were kits that catered to everyone, and I found several meeting
my needs. It was exciting receiving brochures and carefully measuring my requirements against each
model. I dreamt that all I had to do was send them a check. Since they had done all the research and
engineering, all I had to do was screw it together.
The dream began to unravel after I started calling around. Several didnt answer at all; some had
answering machines; and others had only answering services. This was an ominous signhere I was
considering sending them a lot of money, and they werent even answering the phone. I could just
imagine how prompt their response would be after I bought a kit and later had problems. A sign of the
volatility of the kit-car market was when I called a manufacturer whose ad had just appeared; their
phone had already been disconnected!
On top of this, I found buying a kit definitely does not provide the no-brainer I just have to screw
it together since all the works done for me approach that Id hoped for. Over and over, I heard the
same story: new owners finding they had to re-engineer their kits because they didnt fit together as
advertised. That really bugged me, because I was paying good money to have this very issue dealt
withit should just fit together. If it doesnt, you may as well have built your own from scratch.
Hmmm.
By no means am I saying that all kit-car manufacturers are like this; some have been around for
years making top-quality kits. Its just that, in my case, I got a negative feeling from the kit-car
industry as a whole. Even when I called a manufacturers own suggested list of satisfied customers,
they often said great things about the car but bad things about the manufacturer. The most frequent
complaint was the quick promise to build anything desired in whatever time frame requested, only to
have extremely long delivery timessometimes only resolved by legal threats.
One of my pet peeves is the very poor quality Web sites some of the kit-car
makers put up, with vague technical facts, nebulous references to the designers
qualifications, and a lack of videos (its a really good marketing tool to show the
car being driven!). I think it comes down to misplaced enthusiasm on their part,
developing a car the builders think is going to make them millionaires, yet
completely forgetting about the very hard market theyre entering and the business
sense it requires.
There is another issue: how so many kit cars dont handle as well as they could, but for a surprising
reason. Some kits so accurately duplicate the suspension of the cars theyre copying that they sell you
obsolete underpinnings. The irony is that these cars are kits ; there is no mistaking them for the real
things. Why purposely design in a swing-axle or a flexible chassis just because the real one had one?
(Its because its cheap to build.) If you want a classic car look, okay, but why accept poor
suspension or chassis design? Its unfortunate to buy a new kit and find the replicated design flaws of
the 1950s or 60s. Its your money; dont you want the chassis and suspension to be the best possible
for the money spent?
Another point is that just because its a kit does not guarantee its inexpensive. Granted, the
manufacturer has to make a profit, but consider your budget. Many ads conveniently leave out the
expensive drivetrain. Many of the factory demo cars have drivetrains that are worth upwards of
$20,000. These are shown in glossy brochures, but it is not what you get. Its easy to be misled by ads
claiming you can be on the road for $10,000, but be realistic; will you be happy with rear drum
brakes, low horsepower, and stock Mustang wheels in your Cobra? Sure, the manufacturer will work
with you to give you exactly what you want, but dont expect the final cost to be anywhere near the
original low price. Most used Cobra kit cars sell for $35,000$50,000, which should be a hint to
those who think they can build one cheaper. Several of the GT-40 and Ultima kits are the nicest, most
complete kits Ive ever seen and are of very high quality. You get what you pay for, and the finished
cars cost upwards of $80,000. If you can afford one of these, you will be very happy with it; I could
not.
Much less expensive are the Lotus Super-7 clones, which are about as close as you can get to the
ultimate driving machine. Perhaps youve heard the expression a motorcycle with four wheels, and
it describes the car quite well. They are very small cars; two people in one will bump shoulders, and
at around 1100 lb., it doesnt take much to make them move! I wanted the best-handling car I could
get, so I finally decided on a Super-7 clone produced by the Rotus Company.
Why not a Cobra kit car?
In addition to the costs detailed above, they seem to be everywhere (yes,
vanity plays a part in this!). They can be a real handful, and I wanted a car that
handled well.
This was made clear once at a trackday event, when one passed me going like
hell down the back straight at Riverside. Then, unbelievably, I caught up to him
in the next turnin my Datsun. I got a first-hand view of how squirrelly they can
be: the guy was having a real hard time keeping it under control, while I was just
cruising along on his bumper.
He may have been a poor driver, or maybe the car had been set up poorly, but
either way it scared me. It was disappointing to seeas expensive as the Cobra
washow poorly it handled when really pushed. It was very surprising that my
dumpy little primer-gray Datsun got a faster lap time, but it was because the
Cobra driver was fighting the car. I wanted a car that I was driving, not one that
was taking me for a ride. This feeling was further reinforced at the Del Mar
Grand Prix, when I saw an original 427 Cobra lose control and go straight into
the end of the pit wallthe only time Ive ever seen someone killed right in front
of me.
While the Rotus Company (now defunct) made a fine Super-7 type car, even the customers they
referred me to said it was a great car but the company was very difficult to deal with. It was bad
enough that I didnt want to do business with them.
Being put off from dealing with them wasnt the end of it; the kit-car industry as a whole is littered
with the bodies of dead companies. Long on enthusiasm, yet short on business smarts and money,
many sell fewif anycars, run out of money, and soon close their doors. The existing customers
lose out: at best losing any possibility of every getting spares; at worst losing all their money.
Another very attractive car is the Ariel Atom (above), a car that makes the Super-7 look like a
luxury sedan. The performance and elegant design is most impressive, as is the $60,000 $80,000
price tag, which I simply couldnt afford. Of course, back when I was deciding what to do, the Atom
didnt exist! (Image credit arielatom.com.)
If I were considering a kit today, Id look closely at the Stalker (above), a Super-7 clone based
upon American donor parts. Since it uses parts from a common Chevy S-10 pickup truck, its
reasonably priced; with the optional 300 hp supercharged V6, it can give blinding performance yet
still be under $17,000. My brother, Scott, built one. (Image credit Brunton Auto.)
I also realized how kit-car manufacturers deal with meeting emissions requirementsthey dont.
Since the kit is shipped without a drivetrain, they dont have to deal with the legalitiesyou do.
Usually, you are told to install a pre-smog engine. In some states, this can be a very real problem that
absolutely must be researched before you buy the car. Imagine spending $10,000 on a kit and another
$10,000 to get it on the road, then discovering your state wont allow it on the road!
I saw this happen in person. A guy had just purchased a Cobra out-of-state, and
it had just failed California emissions testingboy, was he upset, but it was his
own fault. If hed done any research, he would have seen it coming. I didnt know
whether to feel sorry for him or to trout-slap him.
So I was left to digest dissatisfied customers, potential use of hard-to-find drivetrain parts,
emissions problems, slippery answers, and wishy-washy companies. It left me disappointed and
unwilling to part with my hard-earned money. I had just about given up on the idea of building a car at
all.
Building from scratch?
From time to time I had thought about building a car from scratch, but the task seemed so enormous,
especially the composite work. Composite work is a messy, smelly job potentially dangerousand
takes longer to do than any other part of the project. The dust gets everywhere, and if you have an
attached garage, it gets into the house. Building the shell would be an enormous undertaking and
wasnt something I thought I could do.
Beaten down
So having written off modifying an existing car, combined with my inability to afford a car I
wanted, the instability of the kit-car industry, expense of existing kits, and the daunting proposition of
building anything myself, the whole idea of building a car evaporated; I had pretty much given up. But
a dream can be persistent, and it never really went away.
And then
In 1995 I started a new job at an aerospace firm. There, I met people who were really car crazy.
Some came from the Nissan GTP race team; others were designing their own cars; one built a V8-
powered motorcycle; another, a sportbike-powered three-wheeler; and yet another was considering
putting a Supra engine in a Super-7. One day, I was talking to my new coworker, Lee Kaiser. Lee had
the same interest in small, lightweight sports cars as I, and wed been talking about cars for a few
weeks when I mentioned that the only kit car I liked was the Rotus; he looked at me oddly and said, I
designed that car. What?
Yes, years ago he worked for Rotus during the design phasewhat are the odds of that? He said he
was currently building a mold on a 1964 Mk1 Mini (and adding huge fender flares) as part of creating
his own ultimate autocross terror. So there it was, right in front of me, a solution to my biggest
concern about building a car: the composite shell.
Slowly, things started taking shape, and I realized that meeting these guys could be very expensive,
but very enlightening and educational. As I starting thinking the project through, I realized that while
the Super-7 was nearly perfect, the Mini also fit the requirements: small, light, and quick. The
decision was madeI would build a tube-frame Mini with a composite shell from the ground up. I
remember thinking, My wifes going to kill me.
But why a Mini?
First, its a coupe; I wanted a roll cage and feel open convertibles with roll
cages are unattractive. If there was going to be a roll cage, there may as well also
be a roof for some protection from the wind, sun, and rain. Second was chassis
stiffness; with the taller chassis of a coupe, rigidity would be much better.
The whole daunting proposition of building something myself, though, seemed so overwhelming, so
intimidating, and so expensive. Could I do it? I didnt know. Where would I start? Where would I
build it? I didnt have a proper workshop. There were lots of reasons not to build a carbut I
decided to do it anyway.
Back before I got serious about the project, Lee, some of his buddies, and I
made exactly one carbon-fiber body shell, which stayed in the molds for nearly a
year (which allowed the shell to cure very nicely). Later, I borrowed the shell
during initial planning, figuring Id pay him for it or rent the molds to make my
own shell.
Unfortunately, Lees plans changed, and he no longer had room for the shell.
One day he floored me by saying, You know, if you dont use this shell itll just
waste away, eventually getting thrown in the trash. Id really like to see it go to a
good home, to see it turned into something worthy of its intended purpose.
That was quite a moment, both for his selfless generosity and the realization
that I had taken on the responsibility of creating something worthy of this
generosity; I didnt want to let him down.
Little did I know what I was undertakingand so the journey began.
Later my buddy Sumant came up with a name for the car, Kimini. It
combines my wifes name, Kim, and Mini, and is pronounced Kim-many.
(Yes, I know it should be pronounced differently, but oh well.)
On the next page is my best buddy, Cooper. You can probably guess where he got his name; if hed
been a female, her name would have been, of course, Mini.
Since some pages end with empty space, Ill place some more of his pictures here and there. Here
he is with his favorite sweatshirt that he wears during the winter. Whenever he sees it, he becomes all
excited that he gets to wear it again.
2: The Journey Begins
Having made the decision to go through with building a car from scratch, I felt like Id jumped from
an airplane with no parachute and only a small bag of vague ideas. The reasoning was, Ill think of
something before I reach the ground. The toughest problem, the body shell, had been solved, but now
what? There were a million decisions to be made and I had no idea where to start. All I knew for sure
was everything had to fit inside the Mini shell.
Primary design issues
The first step was simply deciding requirements. If a list couldnt be composed, how could it be
built? Resisting my normal tendency to dive right in, I made a list of requirements for the little car-to-
beKimini.
Design goal #1: safety
Hurting or killing myself was one thing; injuring someone else due to my poor design was worse. I
knew friends and family would want a ride, so I owed it to them to build a really safe car; safety was
always kept in mind during design. Every design decision was affected to some degree by how I
thought the car would react in an accident. It was surprising to find that safety ended up driving most
of the design decisions instead of simple structural efficiency.
Safe doesnt just simply mean strong. If the car is infinitely strong, in an accident it will
expose the driver to infinite decelerationthe car wont have a scratch, but the driver will be dead.
Safe means the driver has a good chance of surviving the accidentnever mind the car. The
maximum deceleration a human can handle is around 10 G, so to survive a car crash means doing
whatever is necessary to keep the deceleration below that. To do this, the chassis and shell are
designed to absorb energy so the driver doesnt. In those nasty F1 or stock-car accidents, its normal
to see parts of the car flying off. It means the design is workingparts of the car absorbing energy are
being shed to keep the driver safe. Even the engine may break free, absorbing a tremendous share of
the potentially fatal force; getting the drivetrain away from the driver lessens the energy that may harm
him.
In the case of a home-built car, the passenger compartment within the roll cage is a non-negotiable
area, meaning it must not collapse. On the other hand, the chassis ahead and behind the passenger
compartment can be used as crush structures. I used thinner wall tubing in these areas so they will
either crush or even separate from the car. Do I know thats what they will do? No, but I at least
thought it through and did the best I could.
Whats a safe design?
It means considering how every part of the car will react in an accident.
Where will this tube go if the chassis gets crunched in here? Where will the
steering column go if I hit something solid? Where does the fuel cell end up?
How much force can that tube handle? Visualization can help with what might
happen in an accident, but only to a point.
The right way, of course, is to do crash testingsomething I couldnt afford to
do (it took forever to just build one car!). Short of crash testing, about all that
could be done was to try not designing anything that was obviously life
threatening. While I knew software existed to simulate a crash, it just wasnt
practical or affordable. I just did the best I could.
Design goal #2: top speed
Was I designing for speed? No, top speed was never an issue; it would go as fast as it would go.
F1 cars can go much faster than they do, so why dont they? Because they need downforce to hold the
car on the road when they corner, yet this downforce causes drag, which slows them down.
Nonetheless, the result is a faster average lap speed, which is what wins races. It surprises people to
learn that current F1 cars have a drag coefficient of around 1.0, which is about three times higher than
a production passenger car. Top speed is not the secret; its being able to fly through the corners that
makes a winner. Besides, how fast did I really want to drive a brick-shaped Mini?
The issue of top speed brings up the California traffic laws. If you are caught
driving faster than 100 mph, youre automatically arrested, your license is
revoked, and your car is confiscated. A very harsh punishment (thanks to street
racers), but it raises the question, why would I need a car that could go, say, 150
mph? I didnt.
Design goal #3: street-legal
I had no interest in spending years building a trailer queen that would sit around for months, only
occasionally being taken to the track. No, if it was going to take years to build, I wanted to be able to
drive it any time. This brings up the issue of emissions.
Emissions testingnot to be ignored
This is a huge, constantly changing topic; if you build your own car, please contact your local
government agency first. I will, however, cover the issue as it pertains to California in 2006.
When I started researching my project in 1995, the emissions issue was a pain in the neck, the
worst issue being the gas tank and fuel system. To be compliant in California, I had to install a gas
tank that had undergone DOT testing. This meant I could not use the ultimate in safety, a real fuel cell.
Worse, the gas tank is just part of the required equipment; all the vapor recovery equipment, charcoal
canister, and vent lines were required.
An interesting government requirement is that the factory Check Engine light
has to function in a home-built car. This is so the inspection station can read out
engine and emission-related codes. This isnt a bad idea, since youll need to
read the codes yourself until you get the car running correctly.
Hot-rodders found a way around all this, starting with an actual old car chassis, one that was
already smog exempt due to its age, and completely rebuilding it. If a brand-new V8 somehow found
its way under the hood, no one seemed to mind. Strictly speaking, this is illegal, yet it really comes
down to looking at the big picture versus following the letter of the law. The important thing is, the
new engine burns cleaner than the original due to electronic fuel injection, and isnt clean air the real
goal?
About the time the emission laws changed to exempt 25-year-old cars, I
wondered if I should have just bought an exempt car to work on. After thinking
about it, I realized I would have the same problems I had the last time I modified
a street car: the suspension design, flexible chassis, small brakes, and heavy
chassis would require so much workI just didnt even want to go there. I
wanted to build it my way.
So in 1995 it looked like the best idea was to buy an authentic Mini to use as a parts and identity
donor, and I purchased a basket-case 1961 Mini. The poor thing was perfect for my needs, gutted and
rusty beyond belief. Parts in boxes, a shell in sad shape, the engine in another box along with lots of
odds and ends, I was in hog heaven. After stripping off the glass, doors, and lights, everything else
was sold off, including the shell. The end result was that I had a bunch of Mini parts, and much more
importantly, the VIN tag.
In 2001, California changed the rules again with the passage of Senate Bill
100, or SB100. Now there were three ways the state could treat a component-
built car. You select one of the following ways to have the state treat the car:
1. Based upon the year of the drivetrain.
2. Based upon what the car appears to be.
3. Based upon a 1960 vehicle (with this option, why have the other two?).
This was a godsend, but there was a catch: each year only 500 certificates are
handed out, and in 2007 all 500 certificates were gone in two hours!
Because of this new law, I would have no problem driving the car on the
street. Be aware, though, that the rules change every yearso please check your
local situation before starting your project.
Design goal #4: insurable
Right from the start I guaranteed it would be insurable by adapting an existing cars identity. Is
this strictly legal? Depends who you ask, but well, okay, no it isnt. But Kimini replaced the
original car; there is still only one car with that original Mini VIN. Also, using a modern engine
ensures that the car puts out less smog than the original ever didthat has to be a good thing. Kimini
would be insured as a hot rod version of the original car; much like 1950s60s American cars are
converted into and insured by companies specializing in classic cars and hot rods.
Design goal #5: cost
At the start of the project, I guessed it would take about $15,000 to complete the car. This did not
include tools or the other million things that conspired to make it cost more. I didnt worry about the
exact total, because I figured the project would take so long that the cost per month would be pretty
low.
Lesson learned
The thinking was that I could buy the parts I neededas I needed them. I
thought this would spread the cost out to make the project affordable, kind of like
a monthly payment plan. This would have workedhad I been able to throttle
purchasesbut this did not work.
Perhaps the biggest lesson of the project is this: to design a car means having
all the specifications of all the parts, the weight, size, mounting-hole locations,
etc. I found out the hard way that it was impossible to know this data up-front. To
know what the engine weighed would require buying and weighing the engine. To
know how to mount the engine meant measuring the engine mounts. To know the
exact size of the seats and the mounting method required buying them first. To
know if various suspension parts would work meant going to the wrecking yard
and measuring them.
The point is, I had to buy most of the parts before I ever started building,
including the drivetrain, seats, fuel cell, etc. It ended up being far more expensive
up front than I expected.
The body shell preparation and the paint were a big hit to the budget, adding around $4,000 alone.
Of course, since the shell preparation came much later, it helped spread out the cost.
My buddy Ron told a story of when he was going to build his house, asking his
brother, an architect, what it would cost. His brother replied, How much can
you afford to spend? Add 25% to thator a little more.
Its nice to dream about building a car that accelerates like a dragster and corners like an F1 car,
but building one is a different story. I had to be brutally objective and honest with both my fabrication
skills and pocketbook. Sure, an aluminum DOHC 427 engine would be nice, but whos going to pay
for it? The engine would be very expensive, never mind the rest of the drivetrain. Six miles per gallon
gets old in a hurry if Im paying the bill. Even relatively ordinary 245/50-16 tires run $1,000 a set, so
the project had to be constrained. I would be the one paying for maintenance and insurance for the life
of the car.
This reminds me of a common question in the on-line forums; someone asks,
Do you guys think I should turbo my engine? The answer is invariable Yes!
Why? Why notthey arent paying the bill or fixing the engine when it blows up,
and they wont be there to help when it doesnt pass smog. No, when building a
car, the builders are responsible for dealing with the consequences of their
decisions.
Of course, since the car wouldnt be driven much, the running costs would certainly be lower than
those of a normal vehicle.
Design goal #6: reliable
Basing the car on easy-to-get parts was always kept in mind; being able to walk into any car parts
store was a very big plus. Imagine being 1,000 miles from home, attending some event, and having a
distributor cap crack. No problem if your donor car is a Honda, Toyota, Ford, or GM, but a
completely different story if your donor is a Cortina, Renault R5 Turbo, or Holden. I have nothing
against these cars; they just dont exist in Southern California. The point is, be aware of whats
popular [read cheap and easy to find] in your part of the world and design accordingly. For example,
as of 2006, the Mazda Miata is an excellent donor, but of course, in a few years it will be something
else.
I wanted the car to be reliableHonda reliable. I wanted to drive it to the track, race it, and
drive it home, all without doing anything other than adding gas. So many times at an autocross or time
trial, I see people messing with their cars. Whats the point of how fast it goes if it breaks all the
time? In fact, the real goal was to attend an event and never remove the engine cover.
I never did consider using an authentic Mini drivetrain; I just felt they were too
unreliable. While Ive never owned a real Mini, I am a member of a local Mini
club, so I hear all the stories. I just couldnt understand spending $4,000 for a
nice 1300 cc Mini engine, then another $5,000 for a five-speed transmission,
ending up with an unreliable and expensive power plant for all the effort.
The importance of avoiding hard-to-get parts is similar to that of avoiding unreliable components.
To me, parts that are hard to replace are a liability. In fact, theyre kind of like using unreliable parts
in the first place. That is, you drive down the road worrying about them: I hope the so-and-so
doesnt break; I have no idea where to get another. Doesnt sound like a way to enjoy a drive, does
it?
Design goal #7: practical
I had to be able to build it. Never did I consider a full composite chassis instead of metal, because
I thought it would be too hard to engineer and build. No, I wanted to build it the old-fashioned way,
with a welded-tube chassis and riveted panels. It did mean learning to weld, but I looked forward to
that. It also meant spending a lot of time on the project and less time with the wife, though with careful
management (working on it only one afternoon a week), it was doable.
So these are the goals I wanted to attain. My goals will be different than yours, and yours will be
different than your friends, so make your list with features you want. Just because I like things a
certain way doesnt mean anything if you dont like or want it. We builders are a very independent,
hard-headed bunch, wanting the car to be just the way we want. You will have your own personal list
of must-haves, so set the goals the way you want. This is supposed to be a fun, exciting project; no
point building something you dont likebuild what you want.
There is one more design goal, thoughone so large it needs its own chapter.
3: Handling
Right from the beginning, the primary goal was to design and build a car that would handle great.
To achieve this meant making decisions about weight, where its placed, suspension type, and
unsprung weight.
Weight
Weight, or rather lack of it, is everything in a sports car. I was willing to do without the usual
luxuries in an effort to minimize it. Every component of a car must be sped up, slowed down, and
flung around corners. Inefficient design results in a heavier-than-needed car, forever handicapping it.
Colin Chapman said it best: Adding power makes you faster on the straights. Subtracting weight
makes you faster everywhere.
Weight is not always a bad thing. Its interesting how heavy the world speed
record cars are. This is because they have plenty of room to accelerate for their
speed run and have no turns to worry about. But a sports car should go fast
around corners, too, and low weight and a low center of gravity are the secrets to
a fun, quick car.
Though Kimini is a sports car, it didnt have to be intolerable to drive. Going deaf, losing teeth,
biting my tongue off, or being pelted by gravel wasnt on the fun-list, so a balance had to be struck.
The goal would be to keep the passenger compartment as pleasant as possible; ventilation and
possibly a heater were the only items considered. Its an endless game of give-and-take
compromise!
What things really weigh
Heres a fairly complete component weight list for Kimini:
Chapter 3: Table 1
Seems like a fairly complete list, doesnt it? But the actual car (with driver, which is included
above) ended up 120 lb. heavier. Where did all the extra weight come from? Oh, a little here, a little
there, tiny indiscretions that conspire to quietly push the weight total higher without being noticed. Its
not the big parts that cause mischief, but the little stuffthings too small to bother weighingyet its
that little stuff that makes for a heavy car. Its tough to build a really light car. Granted, if Id made a
Super-7 type car, the doors, roof, parts of the cage, and some of the windows would go away. It
would still be fairly heavy, though, and where do you cut weight to meet the arbitrary goal of light
enough? What parts can be eliminated or have light parts substituted for them? It can be a very
difficult task to trim several hundred pounds off an already fairly light car! The lesson is, you get one
shot at doing it right; after that, its much, much harder to go back in and find the extra weight.
The next time someone says, Its just a pound; dont worry about it, thats
another pound that goes on the list. Only the final weight of the completed car
will show the ugly truth. I tried to do a good job keeping everything light, yet I
still got smacked!
How to design a light car
Some of weight reduction is easy. For example, we can do without the many options that come with
ordinary street cars. That means no heater, stereo, air conditioning, power steering, power brakes,
power mirrors, power windows, power heated seats, ABS, auto-aiming headlights, automatic
seatbelts, CD/DVD player, navigation system, air bags, buzzers, beepers, rear-view TV, etc., etc.,
etc.
I did add a vent blower; better would have been a squirrel-cage fan since they
are both quieter than an axial-flow fan and move more air.
A light car means:
using a four-cylinder engine instead of a V6 or V8.
using a 10 gal. fuel tank instead of a 2030 gal. unit.
using a 13 lb. Odyssey battery instead of a regular 50 lb. automotive unit.
using steel only where needed and aluminum or composite material everywhere else.
using Lexan windows, where you can get away with it.
using an aluminum engine flywheel (spinning weight is doubly bad, due to it consuming power).
using a lightweight header.
removing all unused accessory brackets from the engine.
replacing the stock alternator with a smaller, lighter, more compact unit.
removing all unneeded wiring from the harnesses.
using light brake calipers, wheels, and tires.
using largerdiameter, thin-wall tubing instead of small-diameter, thick-wall tubing.
using tubing only where needed, fully triangulating every bay.
There are many car designers (myself included) who are disappointed by what
their theoretical car weighs in real lifewhy is that?
Over and over I read the same comments on forums, Yeah, that part is a little
heavy, but its only a pound or so. This remark always bugs me; it is the very
reason why cars end up being too heavy. Its very much like being on a diet and
cheating here and there. The individual episodes are not that harmful, but its the
mindset, the day-to-day pattern of breaking the rules just a little. It conspires to
destroy the goal of always being on guard, of always watching for the opportunity
to make each part lighter. A phrase I read summed it up well, Find tenths,
meaning if you find all the little improvements it adds up to a big number.
I had delusions of grandeur thinking I could get the total car weight under 1000
lb. I had no idea what anything weighed, and that target turned out to be
completely unrealistic.
Weight distribution
Okay, so the car is as light as possible, but how the weight is distributed is just as important.
Polar Moment of Inertia (PMOI)
Heres a fun experiment. Next time youre grocery shopping and have something heavy in the cart,
try this. Put the load near the front of the cart, simulating a front-engine layout, drive around a few
corners, and feel how it handles, how reluctant it is to change direction. Now slide the weight to the
back to simulate a mid-engine layout, and drive it again. Notice how much easier it is to steer, how
easily it changes direction? This is a perfect example of how PMOI affects handling, how moving one
component completely changes how the car feels, yet retaining the identical overall vehicle weight.
As a related example, consider two objects with the same weight: a bowling ball and a barbell.
With the same weight, both cars will accelerate and brake exactly the same. But what is not the
same is how they react to changing direction. Try rotating the bowling ball suddenlywith its
compact shape it will rotate fairly easily. Now do the same with the barbell, hanging it by one arm
and giving it a twist; note how it is much more difficult to rotate. Why? Because the weight is
distributed further away from its center of gravity, its polar moment of inertia is very different.
The bowling ball represents a mid-engine car because all the major elements (engine, driver, fuel)
are near the CG. The barbell represents a traditional front-engine/rear-drive car. The front-engine
car, with its drivetrain way out ahead of its center of gravity, and the gas tank and heavy rear axle
way out back, may very well have a 50/50 weight distribution. Its ability to corner quickly its yaw
ratewill be much lower, though. The mid-engine car will dart into and out of corners faster, all for
doing nothing more than giving some thought as to where the major parts go. By tucking the major
components in close to the CG, the car is more willing to change directionsthe definition of what a
sports car is. In my mind, for a car to handle as well as it can means placing all heavy components as
near the center of the car as possible. Keeping PMOI low is perhaps as important as overall weight,
or locating the CG position.
Engine placement
The heaviest component is the engine/ transmission, and if it can be moved close to the CG, the car
will be able to change direction quicker (due to the lower PMOI). The compact nature of todays
FWD engine assemblies, the huge choice of power plants, and the impending demise of four-cylinder
front-engine/rear-drive setups make FWD engines a natural choice. Besides, placing the drivetrain in
the mid-mounted position, where God intended engines to be, seemed so natural.
It was really astounding how just the placement of the engine [behind the
driver] gave comfort to the drivers. The whole feeling of being at the end of the
stringwas gone. Phil Hill
Even better is that FWD drivetrains are typically oriented east-west, so they take up very little
room fore-to-aft. A mid-engine layout also helps avoid many of the common problems with front-
engine sports cars, such as tiny foot wells, having your feet cooked by the exhaust headers, making
room for a spinning driveshaft, the exhaust interfering with the steering shaft, and a lack of traction
due to little weight on the rear wheels.
The decision was an easy one, and with virtually no downside, a FWD drivetrain in a mid-engine
configuration was chosen. I must add that Ive always had a soft spot for the old Renault R5 Turbo, a
small, unassuming car with big fenders and a mid-mounted engine. Its no coincidence that my car
vaguely resembles one! (Image credit Sport Compact Car.)
Whats front engine, mid-engine, and rear-engine?
The Miata and the Super-7 are examples of cars with front-engine layouts; the
Ford GT and Lotus Elise have mid-engine layouts; and finally, an old VW bug
and Porsche 911 are examples of rear-engine cars.
Lately, some manufacturers claim their front-engine cars are actually mid-
engine, because the engine is located behind the front axle. While its a matter of
semantics, I stick with the old-school definitionI say it is front-engine.
CG location, vertical
Keep the CG low. Having one pound four feet high has the same effect as two pounds two feet high;
both exert the same tipping force on the car. Think of it as torque, which is whats really going on
when the car goes around a corner. All the components are working together to make the car lean in a
turn.
In many cases, you dont have much of a choice where things go, but you can still make small
improvements: the battery for example. Will it be mounted standing upright? Why? If the unit is sealed
(and it should be for cleanliness and safety), it can be mounted in any location or orientation, so do
so. Laying a heavy battery on its side helps every little bit helps. After placing the major
components like the engine, driver, and fuel tank, moving smaller parts around didnt have a large
effect. In fact, it was surprising that moving the fuel cell several feet only moved the CG about an
inch! It does add up, though, and its always a worthy goal to keep things low, pushed in the direction
you want the CG placed.
An interesting side note is that, for the same shape, the CG of a block of lead
and a block of Styrofoam is in the exact same spot! How the material is
distributed is what places the CG, not what the material is composed of. This
may be obvious, but I found it surprising.
It is easy to make small improvements; an example is how some drivers mount fire extinguishers
proudly on the passenger-side windshield pillar (which they cant reach). While its up high to
impress his friends, it shows that the driver knows nothing about the effect: it is actually raising the
CG and making the car less of a real sports car.
Mount everything as low as possible. Many times, this doesnt cost anything more than a little
thought, and little things add up. Adding an item above the existing CG will raise it. Even instrument
gauges are a place for improvement, because theyre located relatively high. Perhaps they can be
lowered, or consider using smaller, lighter gauges, or dont use as many.
Consider the not-so-small consequence of seat placement. Second only to the engine, the weight of
the driver and seat has a significant impact on the CG, and this means placing it as low as possible; I
placed mine right on the floor.
On your scale drawing, draw a line from your eyes, over the rim of the steering wheel, to the
leading edge of the hood. Lower the seat as much as possible until the steering wheel, leading edge of
the car, and your eyes coincide. If the seat cant be lowered enough, consider reclining it, which also
moves the CG back and down. My seat and steering wheel are such that the driver views the top of
the wheel right at the front edge of the hood. In other words, the steering wheel does not block my
view important for avoiding, oh, autocross cones!
Per the spreadsheet, the CG ended up at 16 above ground.
CG location, fore/aft
Many people say a 50/50 weight distribution is best for a sports carI disagree. Reading up on
sports car and race car designs led me to believe that 40% front/60% rear is a better goal.
Chapter 3: Table 2
They all have weight distribution near 40/60, and all are the best in the world for performance.
This weight distribution has a twofold advantage: during acceleration, we want weight on the rear
wheels for maximum traction, and 60% weight on the rear tires helps this; it also decreases (but
doesnt necessarily eliminate) the need for a limited-slip differential (LSD).
FWD unloads the driven wheels during acceleration, exactly the opposite of
whats needed.
Under braking, weight transfers to the front. Having 60% on the rear tires helps keep the back end
of the car planted on the ground. This rear weight bias allows setting brake bias (the percentage
braking force fore-to-aft) further toward the rear. Now all four tires are helping slow the car instead
of just the front ones like in most typical front-engine cars.
My brother got a ride in a Porsche 911 at Laguna Seca; he couldnt believe the
incredible braking. Its due to the rear-weight bias; the engine holds the back of
the car down so the rear tires can aid braking.
After my car was done, the first track event showed that the front and back
wheels were nearly the same temperature. This meant the rear tires were doing
nearly the same braking work as the frontsa good thing.
In the special case of going around a turn at constant speed, yes, but how often are we in this
situation? Coming into this fictitious turn, we are on the brakes. Under braking, we want equal load on
all the tires for maximum braking ability, and a rearward weight bias helps accomplish this. As we
come out of the turn, we step on the gas. For maximum traction, we want more weight on the driven
wheels, and a rear weight bias accomplishes this, too.
One big downfall of the 50/50 theory is that its common to have larger tires
on the rear. In that same corner, our perfect 50/50 car will not have evenly
loaded tiresthe fronts having higher loading because theyre skinnierso
well see understeer from this perfectly balanced car.
A car balanced at 50/50 often requires anti-roll bars. Why, if its perfectly
balanced, would it need bars? Its because the roll-couples are not the same
front-to-back. The point is, 50/50 weight distribution is not a magic formula for
balanced handling.
As for what manufacturers claim is best, I think it is marketing BS. That is, with a typical front-
engine car, in a practical sense, it is impossible for manufacturers to get weight distribution to 40/60,
so its in their best interest to then claim that 50/50 is best. Whats ironic is that when this same
manufacturer comes out with a mid-engine car, all of a sudden the 50/50 argument disappears.
Will going from 50/50 to 40/60 make a night-and-day difference? It depends who you ask; I felt it
would made a big difference (and it did).
The CG ended up at 41% front/59% rear, with me in the car and half a tank of
gas.
Component placement: a difference of opinion
I placed the front calipers behind the front-axle centerline and the rear calipers ahead of the rear-
axle centerline. By moving calipers toward the CG of the car, I reduced the PMOI. A friend said I had
compromised accessibility for a negligible improvement. After thinking it over, I disagree.
The reason our project cars end up heavier than we want, with the weight distribution or PMOI
wrong, is a lack of vigilance. Unless we adapt a mindset to always be on guard, to always watch for
any opportunity to improve things, were going to be disappointed. As said elsewhere, achieving
weight goals is very much like being on a diet, always tempted to be lazy and cheat. Thinking, it
wont hurt, just this once, or itll have no measurable harm is a great way to miss the goals.
I was also told that I was wasting time fussing with the weight and placement of small parts, too. I
then asked, if one pound is unimportant, what about a 1 oz. component? What if there are 100 of them,
then does it matter? Taken as a whole, yes, but when choosing the 1 oz. or the 1.5 oz. component,
most people wont care, because its such a small difference. Thinking this way is the first step to
disappointment when the assembled car is weighed.
In the case of the calipers, they have to be there anyway, so why not put them where they minimize
their impact to the design goals? Moving them toward the center of the car costs nothing, and it
decreases PMOI. In my car, that is 40 lb. that has been moved one foot closer to the CGwhy
wouldnt I make this improvement? And regarding accessibility, as long as the brake caliper bleeders
are at the top of the assembly, bleeding the brakes is just as easy.
The best place for calipers, from a weight standpoint, is at the bottom of the
rotor, to lower the CG. Unfortunately, its then hard or impossible to bleed the
brakes, plus they interfere with the lower ball-joint location.
Suspension types
The eternal argument is whether to use a live-axle, de Dion, or IRSan endlessly debated topic I
am unlikely to settle. To me, it came down to two qualities: unsprung weight and camber
compensation.
Unsprung weight
Unsprung weight is the weight of all components that directly follow the road
surface: tires, wheels, uprights, brakes, plus other components, depending upon
the suspension type. It is the weight of all components not supported by the
suspension springs.
Sprung weight
Sprung weight is the weight of all components supported by the suspension
springs: chassis, shell, engine, transmission, fuel tank, driver, etc.
IRS typically has the least amount of unsprung weight. A live-axle (a pickup truck axle, for
example) has the worst unsprung weight, with the entire weight of the axle tube, axles, differential,
brakes, wheels, shocks, and tires moving up and down over the road surface. A de Dion is
somewhere in between, using a chassis-mounted differential like the IRS, but with the wheel hubs
connected together like a live-axle.
Unsprung weight
Drive an empty pickup truck down a bumpy road for a good demonstration of a poor sprung-to-
unsprung weight ratio. Now fill that same truck with rocks and drive down the same road, noting the
improved ride quality. The improved ride is due to increasing the sprung weight. The same
improvement can be made by decreasing the unsprung weight. Low unsprung weight provides two
desirable qualities: a smooth ride and the ability of the tires to follow an uneven road surface.
While people argue endlessly about which suspension is the best, everyone
agrees that having low unsprung weight is a good thing.
If the unsprung portion is a large fraction of the total, it means when the car hits
a bump, the large unsprung component of the force will feed a large jolt into the
smaller (by comparison) sprung weight above, unsettling the car. How much of a
jolt is simply a proportion of the total:
K = an unknown constant
F = the force fed into the chassis
A luxury car achieves its smooth ride by being heavy (raising the sprung
weight), lowering the unsprung weight, or both. Both result in a smaller force (F)
feeding into the chassis for the same bump.
Example:
Car X, with a live-axle, hits a two-inch pothole, causing the chassis to jump
upward some amount. Car Y, with IRS and half the unsprung weight, will jump
less than half as much, giving a smoother ride. More importantly, it keeps the
tires on the ground. Having the tires follow an uneven road surface is extremely
importanthow can the car accelerate, brake, and corner if the tires arent on the
ground?
Shock and spring rates affect the feel but cannot overcome the high unsprung
weight. Having a heavy live-axle means stiffer shocks and springs are needed to
keep the 250 lb. of unsprung weight from bouncing over bumps and hitting the
chassis. Those stiff springs and shocks will result in a very rough ride.
How to lower unsprung weight
Remove as much weight as possible from the wheel assemblies riding on the road. This can be
accomplished by:
1. using independent suspension.
2. using smaller tires.
3. using aluminum or magnesium wheels.
4. using lightweight uprights (which probably means fabricating or buying them).
5. using lightweight aluminum calipers (usually aftermarket) which can save 20 lb.!
6. using carbon/ceramic brake rotors (if you have $6,000 for a set).
7. moving brake calipers and rotors inboard (possible only with IRS and de Dion).
Even though I had independent rear suspension, I could not employ inboard
brakes. My FWD drivetrain does not have room around the axle output ports for
brake rotors and calipersa shame. Inboard brakes can allow (custom) uprights
to have some really nice specs. Scrub and kingpin inclination can be nearly
eliminated, but, oh well, next time.
Inboard versus outboard suspension
Inboard:
Lowers unsprung weight (good)
Gets shocks out of the wind (extremely important on 200+ mph cars) (good)
Can have any shock ratio (good)
Is usually a little heavier and more complicated (bad)
Takes up space within the chassis (usually bad)
Moves weight closer to the CG (good)
Decreases PMOI (good)
Outboard:
Frees up space within the chassis (good)
Raises unsprung weight (bad)
Simpler, lighter, cheaper (all good)
Hard to get a high shock ratio (bad)
Increases PMOI (bad)
Is not as sexy (that depends)
I chose front rocker-arm (inboard) suspension. Building a car again, I would
probably use ordinary outboard suspension, because its easier and frees up
space around the drivers feet. I used it because I got sucked into the it looks
cool and everyones doing it thing. Then again, if I build something completely
different, inboard shocks may be a necessityit all depends on the requirements.
Packaging usually ends up driving the decision of what to use.
Camber compensation
IRS can compensate for chassis roll by keeping the tires upright in turns; the de Dion and live-axle
dont need to, theoretically. Since I didnt have 500 hp spinning the tires halfway down the track, I
wasnt concerned about the tires being exactly upright during acceleration. Of course, it didnt get me
off the hook, because there could still be a problem with traction under braking, when the tires also
need to be upright. Stiff springs can helpto a point.
Hoosier, a manufacturer of road-racing slicks, specifies a crazy -3.5 static
camber to handle tire rollvirtually impossible to do on a live-axle setup. Yes,
a live-axle can have some static negative camber, introduced by bending the
housing, but its neither simple nor cheap to do, and it cannot introduce enough
camber to suit many race tires.
A de Dion axle can have static camber to counteract tire roll in cornering, but it may hurt
acceleration and braking, since the tires are no longer square to the road. Of course, IRS and live-axle
can have the very same problem! To muddy the waters further, which suspension you choose may
depend upon how wide the tires are. How important is it that they stay flat on the ground under
cornering or braking and acceleration? It all depends whats most important to you as designer.
Its the proverbial three corners of a triangle: cornering, braking, acceleration.
Choosing your suspension to excel at any one of these comes at the expense of the
other two. Most designs are somewhere in the middle.
In choosing a suspension type, it comes down to the definition of whats better, which I counter
with, Better for what? The common argument is that, for racing on a smooth track, there is little
difference. What rubs me the wrong way is this billiard-smooth racetrack everyone talks about. Ive
never driven on one, but whatever.
Colin Chapman developed the Lotus-7 for one purpose: to win races on race
tracks, not to drive on bumpy public streets. The live-axles he used back then
weighed half of whats being used now; hed probably roll over in his grave
knowing that truck axles are now being used in Super-7s!
The reasoning goes, de Dion suspension is better than IRS because its simpler and therefore easier
to get right. Its an uncertain way to win a race, though, counting on the IRS suspension designer to
have screwed up.
Kind of like dating the less-pretty sister, because you assumed the pretty one is
difficultmaybe, or maybe shes worth the effort.
The biggest point is that most scratch-built cars are street-going, driven far more on the street than
on the track. The real world is full of speed bumps, dips, driveways, potholes, and cracks, and
unsprung weight alone determines ride quality on real-world streets, regardless of suspension type.
Its very hard to make IRS work well, but very easy to have it work badly.
If independent rear suspension is too hard to design, what about the front
suspension? No one complains about that, yet it, too, is independent suspension,
just like the rear. You have to go through the same design process to confirm
there are no problems with the forces involved and no toe control issues, and you
still have to make A-arms, just like at the back.
There are two more important differences between a live-axle and IRS. With IRS, the roll center
(RC) can be placed anywhere we want. With a live-axle, unless a more complicated lateral axle
locator is used (Mumford or WOB link
1
), the RC will always be at the center of the axle, usually
higher than what is desired. The RC of the live-axle will move up and down with the axle locating
device. As we will see later, having an RC that moves around relative to the CG of the car is an
unwanted feature. Adding a Mumford or WOB link lets us place the RC where we want but makes the
simple live-axle installation less so, thereby negating some of its simplicityits main advantage.
Yet another issue with the live-axle is traction-related. Because of the live-
axles construction, engine torque is trying to rotate the entire axle the opposite
direction that its spinning the driveshaft. This translates to the axle trying to lift
one wheel off the ground and drive the other into the road. It means that one tire
will have perhaps 200 lb. more load than the other tire. While a limited-slip
differential will force the tires to turn the same speed, it doesnt get around the
fact that one tire is doing more work. Drag racers are familiar with this issue,
which is why they invented jack screws.
And this all means?
So after this much-opinionated commentary, does one suspension design stand out as superior to all
others? Noeach has its good and bad points. The IRS has the most potential, but it is the most
complex and likely the most expensive. The de Dion has lower unsprung weight and is more complex
to design than the live-axle, but it does not provide dynamic camber compensation.
Another perception of the de Dion is that it does not have toe-in and toe-out
issues like the IRS is perceived to havewrong! If the trailing links are not
parallel to the ground at normal ride height, the rear wheels will either toe out or
toe in under bump and droop. Do not assume that just because de Dion is not IRS
that its somehow impervious to poor design!
The live-axle has only one thing going for it, simplicity, but this is a big advantage. It comes down
to your particular set of requirements.
When researching this subject, to make up your own mind, be aware of the
human element. Most people seem to feel whatever they have must be the best
choice, treating any other approach as inferior. Its quite unlikely that youll meet
car owners insulting their own suspensions! Most owners of Sevens like the live-
axle, which strangely happens to be whats in their carsand to be fair, it is the
easiest to package.
The real way to settle all this is to compare three identical carsidentical
except for rear suspension: one with IRS, one with de Dion, and another with a
live-axleon the track and on the street. How many of us have done that? The
next time a driver tells you that his suspension is superior, ask him if hes driven
identical cars with the other suspension types for comparison; if hes honest,
hell say No. Have I compared these three fictitious cars? No.
The drift myth: a rant
This isnt exactly handling-related, but it is perceived as such.
Ever since the sport of drifting a car around a track arrived, there has arisen a
perception that drifting is the fastest way around a track. The thinking goes
something like this: Because the brakes arent used coming into turns, no speed is
lost, so the lap time will be faster. This view, no doubt, is put forth by people
who have never driven a car on track. They see videos of sideways cars, think it
is cool, and adapt it into their way of thinking. After all, the sideways cars on the
video do look like race cars.
Using some of their own logic, however, tells a different story. If the drift
car travels 60 mph on a two-mile road course, the drifter should get a two-minute
lap time, regardless of how many turns, because he doesnt have to slow down
(per this logic). Of course, that doesnt happen, because instead of using the
brakes, the driver bleeds off speed by throwing the car sideways. A really good
drifter can keep the car sideways with a ridiculously low forward speed, so it
certainly has nothing to do with lap time.
Drifting is an exhibition sport, like skateboarding or figure skating. It may be
fun to watch, but reaching the conclusion that its the fastest way around a track is
flat out wrong. Whoever thinks this has been watching too much Japanese anime
for example, Initial-D.
Shown here is Cal Poly San Luis Obispos 2006 FSAE entry, the only car with a carbon live-axle
tube, carbon wheels, and a single rear brake. It lifted the inside rear tire clean off the ground during
skid-pad competitionyet did extremely well.
Taken to an extreme, any design can be made to work very well. Of course, this is a track-only car
that doesnt have to deal with real-world streets. There was also no concern about uneven tire wear
due to the short events. Note the tiny mono-shock located top-center above the two pushrods pretty
wild.
My brothers Super-Stalker uses an S-10 (pickup truck) live-axle that weighs
about 150 lb.without wheels or tires. Very heavy, and very counterintuitive to
the lightweight nature of a Super-7. Im very curious as to how well it handles
bumps250 lb. unsprung weight at the back of a Seven cant be good. When you
have a live-axle that weighs nearly as much as the car its supporting, just who is
springing whom? I guess I need to wait before passing judgment, but no matter
how simple it is, it cant overcome its weight. Aluminum live-axle housings and
differential housings are available from stock-car parts suppliers, but they are
very expensive.
In my case, it was an all-or-nothing decision; the goal was to not look back someday and think, I
should have built a better rear suspension. So, I decided right from the start to incorporate the best
suspension I could, difficult or not, and that was independent rear suspension. Would the increased
complexity be worth it? I didnt know, but at least I wouldnt wonder, years later, What if I had
or, I should have. Besides, having a smooth ride would make everyday driving much more
pleasant, and I felt that was a worthy goal, since thats where 99% of the driving happens anyway.
Dont you want to impress your passengers with the ride, rather than have them biting their tongues
when you hit a bump?
Conclusion
So based on the above, the goals for Kimini were further refined:
Mid-engine
As light as possible (within reason)
A low CG
Low unsprung weight
40/60 weight distribution
Independent suspension all around, inboard front suspension
All major components located toward the CG to minimize PMOI
So after all this, no, I will not say there is one superior suspension design to use. Every suspension
is a compromise. Springs, shocks, and anti-roll bars also affect matters. Its nearly impossible to say
absolutely whats best because every car design is a different kettle of fish: a mix of rules, goals,
timeframe, budget, track condition, and dont forget designer and driver skill. Whatever advantage
one suspension type has over another may very well be overridden by other factors.
For a Seven-type car, the utter simplicity of the live-axle is hard to beat. For the street, though,
real-world bumps will make the ride very rough, so a de Dion is better. And finally, it is up to you to
decide if the additional complexity of the IRS is worth itit was for me. I chose IRS because I felt it
got closer to that elusive goal of perfection. Does that mean Ill automatically beat cars using other
rear suspension designs? Of course not. A car is the sum of many decisions and factors; choosing one
suspension type over another doesnt guarantee success or failureat least in non-professional
racing.
Component Weight (lb.)
Complete drivetrain (including exhaust and axles) 450
Driver 150
Chassis 250
Composite shell 100
All paneling 40
Seats 30
Fuel cell, half full 60
Radiator, full 20
Doors, two 60
Windshield and windows 30
Front unsprung weight (wheels, tires, suspension) 20
Rear unsprung weight 125
Battery 15
Cooling fan 7
Steering shaft and wheel 5
Wire harness and computer 10
Lights 7
Exhaust 10
All hardware, rod-ends, bolts, rivets, etc. 50
Gauges 7
Plumbing 20
Paint 20
Steering rack 10
Pedal cluster 10
Accusump, full 20
Shifter assembly 10
All the little stuff Im forgetting 150
Rough total 1680 lb.
Porsche 911 40/60
McLaren F1 41/59
Lotus Esprit 42/58
Lamborghini Gallardo 42/58
Ford GT 43/57
4: Setting Requirements
Now that I had decided upon the basic goals, a list of requirements was needed. I came up with a
list of questions to help focus on and clarify what the project would consist of. You may spend weeks
coming up with answers, but it is time well spent. Spending years and thousands of dollars on
something you end up not liking isnt any fun, so we better point ourselves in the right direction before
diving in.
The Quiz
1. What will it look like?
2. How many seats will it have?
3. Where is the body shell coming from?
a. Buy it?
b. Make it?
4. What do you want to do with it?
a. Cruising?
b. Autocrossing?
c. Drag racing?
d. Road racing or time trials?
e. Weekend driving in the mountains?
5. What suspension will it have?
6. What amenities must it have?
7. Where will it be stored?
8. Where will it be built?
9. What engine will it have?
10. Will it use a single donor vehicle?
11. Performance goals:
a. Handling
b. Top speed
c. Braking
d. Quarter-mile time
e. Weight
12. What size wheels and tires?
13. How long are you prepared to work on the project?
14. Where will the time come from to build it?
15. What is the budget?
16. How much are you willing to spend on insurance every year?
17. How much are you willing to spend for maintenance each year?
18. What skills do you bring to the project?
Taking my own medicine, here are my answers:
1. What will it look like?
I liked the looks of both the Lotus Super-7 and the Mini. They were the lightest cars I knew about in
1995, though in hindsight there were others Id consider now. The Lotus 11 and Lotus 23 are beautiful
cars, but so are the Ginetta G12 and the Ariel Atom. But since Lee was already working on a Mini
shell, that pretty much dictated the answer to this one.
2. How many seats will it have?
While a single center seat is cool, I wanted to give rides. Three seats were tempting, since I could
sit in the centerMcLaren F1-stylebut in a Mini, three seats wont fit. A four-seater was simply
out of the question, so it would have two seats.
3. Where is the body shell coming from?
My buddy Lee. If I were doing the project again, without his assistance, Id buy a composite shell
from a kit-car company and go from there. Well okay, I might also dive in and make my own from
scratch, not knowing how much work it was at the time.
4. What do you want to do with it?
Cruising
Spirited driving in the mountains
Autocrossing
Drag racing
Road racing and/or time trials
Yes to all the above, which meant it had to be street-legal.
5. What suspension will it have?
Full independent suspension all around, with rocker-arm front suspension.
6. What amenities must it have?
Fresh air vent, fan, maybe a heater later, and nothing else, for weight reasons. Besides, it would
probably be loud enough that a stereo would be impossible to hear anyway.
7. Where will it be stored?
In a normal-size garage; a Mini fit that goal admirably.
8. Where will it be built?
In one half of a normal two-car garagenot my dream shop, but if I waited for that, it would never
happen.
9. What engine will it have?
Something reliable, something with enough power right out of the box to be good enough without
working on it. Something affordable, and something that would fit!
10. Will it use a single donor vehicle?
I used various parts from various cars. It would have been much easier and faster to use a single
donor car, since it immediately guarantees that everything will fit together; a Miata is a great example.
Be careful choosing the particular donor, though; if its a really low price,
theres probably a good reason. If possible, try to start the engine; itll give you a
great feeling right from the start. Getting a car that was crashed can either be a
great deal or a big mistake. If the front end is crushed in so far that the engine has
gotten hit directly, its a bad sign. A cracked or bent crank pulley can be an omen
of internal damage, and an engine pushed back may also have transmission and
even rear-axle problems.
Another issue is donor cars coming out of the Gulf states. Be wary of these,
because its possible that they have been submergedin salt waterin a
hurricane. Salt water is extremely corrosive to metal and especially electrical
wiring and connectors. If a fabulous deal appears from Florida, Mississippi,
Texas, etc., give it special attention before buying it.
Regardless of condition, make sure to get documentation. An evasive seller is
likely a sign of troublethe car may be stolen. Many states require traceability
on all major components to make sure your pride-and-joy isnt composed of
stolen parts. A good idea is to have the DMV or police run the VIN as soon as
possible. Dont wait until the car is done to try and come up with paperwork!
11. Performance goals
a. Handling
This was the most important goal, next to safety. The car would have independent suspension all
around, solid rod-ends to minimize compliance, wheel geometry aimed at maximum cornering (even
at the expense of tire life), and high-quality double-adjustable shocks.
Because of how light the car was, I was told to expect even very sticky tires to
last a surprisingly long time and to use about 1525 psi tire pressure!
b. Top speed
Never an issue; it would be whatever it ended up being.
Due to tire size, top speed is about 132 mph not a problem. It will probably
be reached at the end of the straight at Willow Springs raceway, which is how it
is supposed to be.
c. Braking
It would have disc brakes all around, large enough that they would never overheat, regardless of
how I drove the car.
If one car weighs a lot more than another, it will still stop in nearly the same
distanceno, I didnt believe it either. While there are some assumptions, it
really is (more or less) true. A 6000 lb. SUV doesnt take three times as far to
stop as a 2000 lb. sports car.
Of course, a lightweight car can accelerate faster and corner harder, and its
brakes wont overheat as quickly. But for a one-time stop, dont underestimate a
heavy cars stopping distance. A thread on this is at http://www.Honda-
tech.com/zerothread?id=1513032.
d. Quarter-mile time
Somewhere, years ago, I got this funny number in my head: 12.7 seconds. I dont know where it
came from, but I thought, If I ever build a car, it would be great if it could achieve that (and it did,
and then some).
e. Weight
The third-most important goal, closely intertwined with handling. The rather arbitrary goal was
1000 lb., though it was flexible upon safety and reliability.
It ended up being a lot more, but the McLaren F1 missed its 1000 kg goal, too.
Does this place me in the same class of designer as Gordon Murray? Um, no.
The question of a weight budget comes updo I have one? I suppose I could claim I did, but no.
Instead, as each part was chosen, I developed the mindset of always being on guard for weight, or the
lack of it. Only this way could I complete the car and not feel too bad for its final weight.
12. What size wheels and tires?
They must be street-legal, preferably in a common size, and not hard to find or too expensive.
Another benefit of such a light car is that even very soft tires will last a long time. Above all else,
they need to be reasonably priced, since they are the primary consumable. Having a $1,200 tire bill
was not a goal! This big topic is covered later in its own chapter.
Ever wonder, with the billions of worn-out tires out there, why our streets
arent knee-deep in rubber dust? Its because there are microbes that eat rubber
really!
13. How long are you prepared to work on the project?
It took far longer than expected to finish the car, mostly because I only worked on it Sunday
afternoons. Actually, thats not quite right; I spent every moment planning what I would accomplish
the next Sunday. If you put in a few hours after dinner every day, itll go much faster.
I was prepared to work on ituntil it was done. Figuring I spent seven hours each Sunday
afternoon, that averages to one hour a day. Since the project took nearly 10 years, thats somewhere
north of 3000 hours. This total does not include the many hours of research done during the rest of the
week, which probably adds up to another 3000 hours. This isnt meant to scare peoplethats just
what it took to do it. If you use a single donor car itll be much less. I was stubborn and wanted what I
wanted, so it didnt matter how long it took. (Okay, I confess thinking it would take about five years.)
14. Where will the time come from to build it?
I havent watched TV in years, so much of the time ends up being free, rather than having been
wasted away watching dumb shows. Television is an anesthetic to creativity.
As I write this, a group of Hollywood writers is on strikewriters for reality
shows. Think about the irony of that.
15. What is the budget?
I guessed it would cost about $15,000, yet it ended up being about $25,000, including tools. Spread
over nearly 10 years, thats about $7/day. Not a lot; if you take your lunch to work, skip Starbucks,
and eliminate the cable TV bill, you get the car for free! Thats great, though I learned a big lesson
about this later, that most parts were required up-front. While I halfway on purpose didnt keep
receipts, heres a rough idea of what it took:
Chapter 4: Table 1
This, no doubt, leaves out a bunch of stuff Im forgetting and doesnt include tools. It gets expensive
if you do whatever you want. On the other hand, if you use as much as possible from one donor car,
itll save a lot of money.
*Notice the low cost of steel compared to the total, yet a common complaint
from builders is how expensive steel is. I wonder if they have a grasp of what
they are building! While my car is fairly expensive, $300 out of $10,000 is
nothing. Youll spend more on one shock absorber or a set of research books than
on steel.
16. How much are you willing to spend on insurance every year?
The plan was to register it as a 1961 Mini hot rod. That meant insurance would run around
$200$350 a year in California (insuring it as a kit car with a hot-rod insurer).
17. How much are you willing to spend for maintenance each year?
Not much; in fact, about the same as a regular car. I wanted to drive it, not work on it. Other than
consumables (tires, brakes, oil, and gas), I wanted it as reliable as a Honda or Toyota.
18. What skills do you bring to the project?
I knew electronics and had rebuilt several car engines. Id done woodworking projects, yard
projects, a telescope, a fishpond, and lots of electronics projects over the years, many requiring
metalwork (for the chassis). So I was reasonably handy with tools.
But I had never welded anything, didnt own a drill press, lathe, hydraulic press, or various other
tools probably needed. Yet, those could be bought, and a welding class could be taken, so it wouldnt
hold up things.
So now what?
So there are my answers to give an idea of my mindset going into the project. Next its time to start
fulfilling each goal.
This book is organized in the order to go about designing and building a car. This is the Big Secret,
to proceed in the following orderthe Ten Magic Steps of creating a car:
Tires
Drivetrain
Uprights
Wheels
Roll center
Suspension
Springs
Shocks
Steering
Anti-roll bars
A couple of years after developing this list, I bought Staniforths excellent
book, Competition Car Suspension, and was stunned to see a very similar list. I
was briefly hesitant about including my list here, concerned that it might look as
if I was passing off his work off as my own. But since I had figured it out for
myself, years earlier (the hard way), I felt I earned the right for it to be here. The
order of this list is one of the most valuable lessons that I learned.
So with this list we now set off to design our car.
Engine/transmission $3,500
Bodywork, paint $4,000
Chassis, steel $ 300*
Bent roll-cage tubing $ 200
Wheels $1,200
Tires $ 400
Shocks $2,200
Fuel cell $ 800
Front uprights $ 50
Rear calipers $ 40
Steering rack $ 40
Steering wheel $ 140
Quick disconnect $ 80
Gauges $ 500
Seats $1,500
Metal paneling $ 200
Custom axles $ 350
A/N bolts, etc. $ 200
Rivets $ 400
Windows $ 250
Electrical $ 300
Accusump $ 170
Very rough total $17,000
5: Psychology
This chapter is here to prepare for what is coming and for what comes after. Building a car truly
enlightened me to the words compromise, persistence, loneliness, and satisfaction.
Sometimes, it was the realization that I couldnt afford the best parts, didnt have the mill that
would have been perfect, or couldnt decide where to place roll centers. I had to play an endless
game of give-and-take in order to get the thing built. If I didnt, Id risk the dreaded design
paralysis: the inability to make any decision whatsoever because of numerous conflicting issues.
This comes from the belief that, when the car is complete, one or more of the following will happen:
Pushing it out of the garage for the very first time reveals that, in your excitement, you neglected
to add brake fluid. This causes your car to coast down the driveway, across the street, and into
the neighbors yard. This is the same neighbor who is already upset about the loud banging
sounds and blinding blue-white light coming from your garage late at night.
Pushing it out of the garage for the very first time reveals that you have the steering rack installed
such that left and right turns are reversed. (Dont laugh, this can happen.)
After finally finishing, a friend casually says, You know, for all that work and expense, why
didnt you just buy a used Viper? and you hesitate, wondering, Why indeed?
After agonizing for months on design issues, you make the hard choices and build the car, only to
have people second-guess every decision, and you realize their ideas are superior.
After putting it into gear for the very first time, revving it up, and popping the clutch, you realize
your custom shift linkage causes reverse to be where first gear should have been.
After working out all the calculations and factoring in your 10x safety multiplier for the
suspension, the very first pothole you hit breaks off a wheel.
At the first driving event you enter, a stock Honda Civic, driven by a kid with a shaved head
wearing baggy pants and a backward baseball hat, squarely beats you.
The above remark was written years before the car was finished. At the
second practice autocross, I was squarely beaten by a young lady driving a
modified Honda CRXkicked my butt fair and square.
You take your car to The Big Event. You notice some guy standing there looking at your car and
realize it is Carroll Smith (rest in peace). He comes over and casually asks, Is this yours? You
answer, Yes, sir. He asks, Why did you do it this way, what were you thinking, and this, this
here, its all wrong, and that over there is in single-shear. Did you read my books? Didnt you
understand anything I wrote?
The irony is that this could have happened, because I volunteered to help at the
2006 FSAE West competition in Fontana, California. Carroll Smith was always
a judge at the event, and had I driven the car, I realized there was a fair chance he
would have seen KiminiI wonder what he would have said. Alas it wasnt to
be; he passed away a year before Kimini was completed.
While these are somewhat humorous scenarios, they arent quite all in jest; they really were my
mindset. While e-mails came in saying great job, privately I was always haunted by the feeling that
there was always a simpler, cheaper, lighter solution. Yet, I couldnt just sit there, paralyzed by fear
of screwing up, and not make any decision at all. I had to force myself to get on with it.
Often, my wife would come into the garage to ask me a question and find me
leaning against the bench, just staring at the car. She probably wondered why I
wasnt working on it.
The truth is, I was using the 3D-simulator in my head, moving parts of the
car this way and that: How about if I put this here? If that moves over there, then
I could put this here. So many times I worked through a placement problem
without ever moving a part. Once it was figured out, only then would I start
cutting. I ended up saving a lot of time and money by not wasting effort and
material on something that wouldnt work.
If you arent good at spatial relationships or three-dimensional orientation, you
will be by the time the projects done!
Importance of deciding
Lets just get this out in the open: yes, there is always someone smarter out there, someone who
knows how to make the part youre working on lighter, stronger, cheaperbetter. With that out in the
open, we can move on. A part can always be made better if you think about it long enough or ask
enough people. You will always find how it could have been made better in hindsightand yet you
have to move on. Is the part good enough? Thats the real question: whats good enough versus
whats best?
Partway through construction, a buddy said something I havent forgotten:
Youre walking a fine line between getting it done and doing it right.
Hes right. I knew there were better solutions than those I chose, but so what?
While people might shoot holes in its design, at least it was completed. Until they
design and build a car themselves, I tend to take any negative comments with
some skepticism, suspecting that if they built a car, they, too, would realize it
isnt quite so simple.
Preparing yourself
You need to be very, very persistent to complete a project like this.
Some days it will be lonely, occasionally frustrating, and sometimes, seemingly undoable. Many
times during the design, or the build, I became frustratednever quite enough to consider quitting, but
enough that it stopped being fun a few times.
For example, the electrical work was no fun at all; months went by and yet the car looked exactly
the same. Progress was being made, but it just wasnt visible, making it rather frustrating. Maybe it
says something about me, that I like to see progress each week. Theres that aspect of it, then theres
the sheer size of the project. Heres what I wrote in one of the build diaries:
I had one of my anxiety attacks. Occasionally, and never on purpose, I make the mistake of
raising my head from the small part Im making, and wham, the enormity of the project hits home.
When this happens, its very hard to keep working; the project is just so big, so huge, it seems like
itll never get done. The shifter took three weeks, yet it seemed like three months. I can see the gas
pedal taking several weeks easily; its just a stupid gas pedal, yet this little partjust like every
other little partcan kill me if it goes wrong. It has to be made right.
Maybe its time to start the engine again. But before that happens Id like the cooling system to
be complete, which means adding radiator-mounting brackets. But I need to know where the front
extreme of the car is. Its always something.
Ah yes, I remember those days.
Try to find someone who is also building a car or at least someone, or some group, on the Web
who can offer support when the going gets tough. If nothing else, make friends with the guys at the
local parts store.
My parents built the house that we kids grew up in, and my dad said that while
he was building it, he was dedicated to The Project. He pretty much dropped off
everyones social list until after the house was complete. The car is the same
way, because distractions are huge showstoppers. I became very possessive of
my weekends, unwilling to give up even one hour unless it was really necessary.
My buddy Cecil was (is?) designing and building an awesome three-wheeler,
all the while complaining that he never had enough time to work on it. Then, he
bought an old Porsche 914, and later, an ancient Honda Mini-Trail. Ugh!I was
pretty annoyed with him but still hold out hope that hed someday get rid of the
distractions and get back to work on his real project.
That said, since theres no deadline, you are free to decide what you do with
your time. The car isnt as critical as building a house. If you want to take a
break, its of no concern. But unless you apply a steady pressure against the
project, pushing it slowly along, it is all too easy to let up and let it grind to a
halt. Once stopped, it is much harder to get it moving again! Kind of like pushing
a railroad car, Id imagine.
Stay focused! A lack of focus is the number one killer of projects. Concentrate on building the
car, not:
learning CAD to model the frame (unless you already know CAD or dont mind the added time
or are planning to go into production).
building a fuel injection system from scratch to run it.
saving money by making your own fiberglass parts. (If someone makes it, buy it.)
becoming addicted to Internet car forums instead of working on the car.
creating your own suspension analysis software.
using an oversize engine that doesnt fit, throwing the design goals into the weeds.
choosing an engine because it was sitting in the backyard or choosing one that is too big, is too
small, you cant find parts for, or would cost a fortune to rebuild.
buying a broken engine or transmission, requiring a rebuild.
designing and building your own welder (I actually saw this on-line).
adding a turbo or supercharger before the cars on the road.
building it without spouses support.
Time-sucking side projects are the death of finishing a project of this magnitude.
That said, Im fully aware of the requirements of parenthood. I say the car
comes second; kids grow up so fast that if you miss out on the experience, what
was the point? On your deathbed, what accomplishments in life will you look
back on, raising kids or building a car that long since has rusted away or
crashed? Remember whats important in lifethe big picture.
Set small goals
Do not set a schedule; schedules are for work. Remember, this is supposed to be fun; you dont
want to work all night because you scheduled something to be done by the next day.
I once got in an argument with a car builder when he announced he would be done in one year. I
couldnt believe itand I told him so not because I thought it was impossible, but because he was
forcing the issue. I tried explaining that applying a deadline absolutely doomed the project to
disappointment. I was taken aback when he argued that he is a manager, and deadlines are a
requirement for getting anything done. Maybe for him, but as a hobby, a schedule would ruin the fun
for me. Bottom line: it has to be fun, or dont bother. His self-imposed deadline did, in fact, come and
go, due to personal events getting in the way. Does that mean he failed? Its up to him to decide that,
as it is up to all of us when we build our own cars. I never had a deadline, so I was never
disappointed by missing any schedule.
I can just imagine having to work until 4 a.m. midweek in order to stay on
target. How well will that sit with the family and the boss at work? Just what
was so critical that it absolutely had to be done that day? Nothingits just a
hobby.
Perhaps youve seen the PBS TV show Plane Crazy (definitely worth watching). The host sets out
to design and build a plane from scratch in 30 days. What happened was that the ridiculous goal and
stress caused him to burn out. The show was very good at showing how people deal with hopeless
deadlines. Keep your eyes open for the show in case it reruns.
I wasnt impervious to burnout myself; heres a diary entry of one of my moments during a
particularly tough session:
Im burnt out; there, I said it.
Its a combination of many things, but sanding is certainly the biggest culprit. As before, I made
the mistake of thinking about how much more work theres yet to do. Last time this happened, it
was the enormity of all the different tasks to complete; this time I dont even have that diversity to
look forward to. No, this time its just endless sanding, from here to eternity.
Having finished the first round of filling and rough sanding on the passenger compartment, I
realize it needs to be gone over many more times. Then theres the engine cover and the nose,
where I get to start all over again. Oh, and I cant forget the doorsugh. My concern is that Ill
spend a lot of time getting the shell to what I think is good enough, only to find that, once I take
it to a shop, theyll say it needs a lot more work, at great expense. I dont want to waste a bunch of
time on bodywork if it isnt good enough. Or, perhaps, its that I just hate doing bodywork.
Clean parts
This might seem like an odd place for this, but itll go far in making the project pleasant to work
on. Putting greasy, dirty brake calipers on your brand-new car will, subconsciously at least, make you
not like the car as much. Just a little time with a brush, soap and water, or at least some brake cleaner,
will get the worst of the filth off the parts. Simply spray painting suspension uprights and brake
calipers does wonders to improve the appearance of the car. It also allows working on the car
without getting your hands caked with grease and dirt.
Incentives
At work, I have a poster that I got back when the project first started. It showed some GT-5 race
cars on track. The interesting thing was that the car in the lead was a lowered Mini with big fenders.
Trailing it in second place was a Datsun 1200. In effect, the poster showed that the car I was creating
was replacing the car Id had. Pretty cool, I thought. The poster was a constant reminder of what I
wanted the car to become. While it sounds corny, it was a constant aid to visualize the end result.
As added incentive, I felt like the ghosts of Chapman, Smith, Costin, and all the
other great designers were silently watching over my shoulder; I didnt want to
disappoint them.
Summation
A project of this magnitude requires faith. Not necessarily in the biblical sense (thought that helps,
too) but rather, faith in yourself. It really is a step of faith, too, opening a never-before opened door,
not entirely sure whats behind it, yet having faith that nothing evil, nasty, or impossible lies beyond,
and stepping through. You must have faith that, with enough time, anything can be overcome.
Sometimes I thought of the project as a steep mountain road. When it got really steep, Id downshift
a couple of gears, using time as energy to grind through whatever obstacle was in the way. The cold
driving knowledge was that it would be completed, so whatever immediate impediment that was
holding things up would get solved.
There was never any question it would get finished; you just have to believeand if you dont, its
like what Yoda said, That is why you failed.
My wife and I recently helped our son move. Standing in his old place, looking
at the huge amount of stuff that had to go, was a little overwhelming. And then it
hit methis was a lot like building a car. Looking at all that work, it was hard to
know where to start. Do we start with the big stuff, the small stuff, the heavy, the
light? It was easy to get overwhelmed and just freeze up.
And then, just like with the car project, I realized it doesnt really matterjust
start. Just pick something up, put it in the truck, and get on with it. Standing
around, forever trying to decide what to move, wastes a tremendous amount of
time.
Yes, building a car does require a lot of planning, but at least something is
happening, instead of you being frozen, wondering where to start. Just start.
6: Preparation
This refers not only to books, but also catalogs, vendors, tools, the workshop, the skills you bring
to the project, and people you know.
Assembling a library
Though the car took 10 years to design and build, it wasnt all cutting and welding. After deciding
to use the shell, I spent a lot of time reading and rereading design booksabout a year. Reading,
taking notes, reading some more, changing the notes, changing them again, round and round. This is
what it took to make sure it came out right the first time, and to be safe.
Building a library is the foundation of the design; the consequences of winging it are too severe
expensive at best, fatal at worst. It can easily cost the same to build a car right or wrong, so take
the time and use the material the smart way, the right way. That material is the stuff between our ears,
so the research library is not an area to save money. Buying a dozen research books will cost perhaps
$400, yet consider the alternative of not buying them. You save a little money, but potentially build a
$25,000 car thats dangerous. Considering that $400 is less than 2% of the total, it makes sense to buy
all the books you need; the ones I read are listed in Appendix A. Also check out the books on Aircraft
Spruces Web site, specifically those put out by the Experimental Aircraft Association (EAA). Even
though their books are for aircraft construction, many of the design and construction practices are
applicable to sports car construction.
Other sources are aftermarket catalogs, listed in Appendix B, and automobile-related Web sites, in
Appendix C. Also drop by your newsstand each week and pick up the Auto Trader publication,
Heavy Equipment Trader ; its not too early to start looking for tools.
Regardless what books you buy, theyre of no use if theyre not readseveral times. The way to
learn something is by immersion, reading everything. If you dont understand something, fine, just
keep reading; eventually it will start making sense. A sign that someone is in the design phase is to
find design books near the computer, the living room, and of course the Master Research Facility,
the bathroom. You will get used to hearing, Are you still in there?
For a year, I did nothing but read and take notes; thats what it took to learn this stuffat least for
me. You absolutely must understand what youre getting yourself into. Remember that your life is at
stake; if you dont understand something, ask someone or reread the section in question again, but
dont skip it. Often reading another book on the same topic can give new perspective on a difficult
issue.
To each his own
Would you build a house without first consulting an architect? Design a boat
without researching the topic? Construct an airplane without extensive reading?
How about building a car from scratch with no suspension design just start
building? Some builders scoff at the vast history of hard-learned knowledge in an
apparent attempt to be different.
There are a million ways to build a car, and most work pretty well. On the
other hand, some construction ideas can create potential dangerous situations:
suspension pivot points placed as convenient, instead of providing proper wheel
control; suspension links attached to weak, cantilevered brackets; or rear toe-
control links attached with tiny, single-shear mounts. In the case of composite
construction, some builders conclude that because they can do fiberglass layups,
thats all thats required to build a composite monocoqueno.
Going around a corner at any sort of speed will show that the Goddess of
Suspension is a very harsh mistressintolerant of ignorance and absolute in her
judgment of poor design. A poorly designed rear suspension, one causing large
positive camber in bump, will cause the car to spin in a turn, under braking, and
maybe even under acceleration. All of this can be avoided if the builder simply
places a protractor on top of a wheel, moves the tire through its suspension
travel, and asks himself, Does this make sense? or Will this work? or Does
this design keep the tire flat on the road? If he does this, he will understand why
it is unsafe.
Suspension design all comes down to wheel control. If its there, the Goddess
of Suspension will let us go on our way, but if not, she will dish out the harshest
of consequences. The moral of the story is: read, learn, and apply. Suspension
design is well understood; the knowledge is out thereuse it. It all boils down to
putting suspension brackets here instead of there. The cost of doing it right is the
same as doing it wrong, with the difference being knowledge.
Good craftsmanship means that a bracket is pretty or that a composite layup
has no bubbles or wrinkles.
Good engineering means that a bracket is of the proper material thickness and
bend radius and is in the proper location. Good engineering means that a
composite layup is of the proper thickness, weave, and orientation; is used with
the correct epoxy; and receives the proper cure.
Do not confuse good craftsmanship with good engineeringthey are not
the same thing.
Designing something differentjust for the sake of being differentcarries
great responsibility. The flaw is that its often based on ego instead of sound
engineering. Suspension and composite design is very well understood; its
already known what works. To shun this knowledge, in an effort to think outside
the box, carries consequences. Since proven methods are being ignored, where
does this leave this design? Suspension or composite monocoque design isnt an
area in which to be randomly trying new designs, without extensive design and
testing, before taking it out on public roads.
Bottom line: research first or dont even start the projector any project for
that matter. A poorly built car can kill the driver, passenger, and other people if it
goes wrong. There is nothing new here; read how to do it right and learn from
those who came before. Building a car isnt ground-breaking; there is no need to
figure it out by trial and error. If something truly new is to be built, do extensive
research and testing before anything is built. Going off into uncharted territory is
fine as long as youve mapped it out first!
I highly recommend creating a spreadsheet! This helped enormously during design, listing all major
components and their weights and locations. It allowed calculating the following chassis design
parameters before the car was ever built:
Total vehicle weight
Power/weight ratio
CG longitudinal location
CG height
Front-to-rear weight distribution
Front and rear unsprung weight
Front and rear natural suspension frequency
Wheel rates
Shock/spring compression
Spring free length
Required roll-bar resistance
How much the car leans in a corner
How much the front end dives under heavy braking
How much the car will squat under acceleration
Creating this was invaluable. I worked out some of the equations and the rest came from Milliken
2
.
Unfortunately it does not fit well into this book, would be hard to read, and even if it were easy to
read, would not be helpful without the underlying equations. I have posted the spreadsheet on my Web
site at a non-public address, just for you. Supplying it this way also means you dont have to type in
all the equations. Go to Appendix D for a description and download instructions. The spreadsheet is
referenced many times all through the design and build.
Coworkers and friends
Another important source of data is people you know and interact with. Depending on what you do
for a living, you may find a treasure chest of information right under your nose. It took a while for me
to realize how fortunate I was to work where I did. Electrical and mechanical engineers,
aerodynamicists, composite technicians, machinists from former race teamsthey were all there.
While I did all the research myself, it was very helpful having other people to confirm my thoughts
and to bounce ideas off of. I found it amusing that three other coworkers were also building cars. As
we discussed our projects, I realized why we build our own cars: were stubborn! Everyone has a
very strong opinion of what constitutes the perfect car, and I found it intriguing how creativity is so
personal. We each approach design issues differently, yet we all ended up with something that
answered the same question: what is the car that I want to build?
Importance of time for research
You absolutely must take all the time necessary to know your material. If you cut corners, there is a
risk of making expensive or dangerous design decisions. I realize the desire is strong to pick up the
tools and start cutting, but resist! If you go through the entire design step-by-step, by the time you start
fabrication you will know exactly where youre going. Building the car becomes the icing on the cake,
the reward for having done the design right. Its where you finally see the design coming to fruition. It
wont be some half thought-out, potentially dangerous mess, but instead, a well-rehearsed
performance. If nothing else, it minimizes having to do things a second time because they werent
thought through beforehand.
A surprising number of builders buy steel tubing as their first purchase.
Depending upon the motivation, this may not be a good idea.
Steel should be the last thing purchased; buying it now means itll just sit and
rust. Some builders counter that it serves as motivation to keep working on the
project. They also point out that, with steel prices rising, its a good investment. I
look at it as expensive material thatll be in the waymost of us dont have room
for a bundle of 820 ft. long tubing. Depending how fast steel prices are rising,
its probably wiser to just leave the money in the bank until the materials really
needed.
I look at buying steel like buying cement for building a house. Yes, 50 bags of
cement can serve as motivation, but unless you already have plans in hand, the
bags are going to just sit there. Motivation? Maybe, or maybe its an unattractive
pile thats in the way, a reminder of a project that isnt moving (and hardening in
the bags due to moisture absorption).
My recommendation is to not be lured into buying stuff up front; theres
plenty of time for shopping binges later. If you absolutely have to buy something
now, buy what you need now, like the engine, or wheels and tires. That serves the
important task of providing weights and dimensions for these important
components. These serve as much more of a motivator than the steel doessteel
can wait.
(I see a surprisingly common complaint about steelthe cost. Steel is about
the least expensive component of the whole project! If the steel seems
expensive, theres going to be a lot of complaining about every other part, too!)
The workshop
A lofty title perhaps; most of us dont have a 2000 sq. ft. shop with heat, air conditioning, and
three-phase power, so we make due with what we have. This typically means getting the job done in a
suburban two-car garage. To make things more interesting, often we have to share the space with our
significant other. I didnt have the heart to ask my wife to park her car outside for what turned into a
seven-year construction phase. Early on, I decided I would make due with half the garage. The small
garage worked out well, in large part because of how small the Mini is! It also forced me to keep it
clean.
Efficiency of space
Regardless how much room you start with, you always seem to need more; the best you can do is
maximize whatever you have right from the start. Some ideas are:
Clean out the garage and have a garage sale, and stuff you cant sell, give away. You need a
workshop, not a storage shed for lawnmowers, clippers, fertilizer bags, bug spray, and paint
cans. You can use the income from the garage sale to buy a storage shed to put these yard
supplies where they belong, in the backyard! I read of an interesting way of dealing with garage
storage: store things in boxes and write the date on them. If, in six months, the boxes are
unopened, throw them out or give them away. You might want to do the throwing out when your
spouse is not home.
If you have an air compressor, consider placing it outside the garage, freeing up valuable floor
space as well as sparing you from having to listen to the thing. Build a small unobtrusive
enclosure to shelter it from the weather.
Build workbenches at a convenient height and with enough space below to accommodate parts
storage. Think ahead about the size of parts and where they are going to gobefore you find
your shelves are too small. If the compressor is to go under the bench, will it fit? How about the
vacuum cleaner? Does the drill press mount on the bench or floor? Is there space so the drill
press belt-drive cover can be lifted, or does it hit the shelves above?
Most garage floors are sloped toward the door for drainage. If you build all your benches
exactly 36 tall, you wont have a level work surface. Watching your 12 mm socket roll off the
bench, out the door, and down the driveway will remind you of that often.
Consider making paper dolls of all the major objects in the garage, drawing them to scale so
you can move them around the floor layout until they fit.
Make a grinding center. I read about this in an old Experimental Aircraft Association (EAA)
publication. Basically its a wheeled stand where you mount all your power grinding tools,
buffing wheel, deburring wheel, sanding disc, and belt sander.
Store your metal tubing up off the floor, preferably on a shelf with one end visible, so you can
check what sizes you have in stock. Another place is under the chassis table, but up off the floor.
I found storing it up off the floor kept it from rusting as quickly, as well as keeping me from
tripping over it.
If you have a traditional peaked roof garage, clean out the rafters so you can store large (but
light) bulky items. Those spring-loaded foldable attic ladders are very convenient for access.
If youre working on your side of the garage, with a car in the other bay, move it outside
during periods of cutting or grinding. Grinding throws thousands of little metal bits everywhere,
and what gets on the car rusts, eating its way into the paint. This is not a good way to make
points with the spouse. (Guess how I know.)
Place the welder near the garage door, so the welding cables can reach objects to be welded
both inside and outside. Same goes for the air compressor. Having the welder near the door is
doubly good because it makes it easy to get the replacement gas bottles to and from the welder.
Put something down on the cold, hard concrete flooranything. While most people use
expensive coatings or expensive floor tile, I used thin indoor/outdoor carpet scraps. My knees
thanked me many times. If it gets dirty, roll it up, throw it away, and get another piece.
Install lots of lights. I have six fluorescent fixtures, three above the work benches and three over
the work area. More is better.
Tools
The car will not get built without tools; the problem is deciding what you absolutely must have to
make it happen. Its a balancing act between being able to do everything yourself and having someone
else do it, and how quickly you want it done. If you buy a lathe, mill, band saw, shear, brake, and
welder, you will certainly be able to do everything, but it will cost you. The trick is balancing the
cost of acquiring the equipment versus what you would have paid someone to do it for youand it
can drive you nuts. It also doesnt help that it brings you face-to-face with the typical guy thing of
wanting to be totally self-sufficient and doing everything yourselfvery expensive if you build only
one car. Having 10 bushings made at $40 each can either be viewed as I should have done it myself
on my own lathe or as, $400 is a fraction of the cost of a lathe, and I get them right away. Often,
the decision to purchase a tool gets confusing because of the contradictory needs, resulting in the
following pseudo-logic:
Time costs money.
You will use the tool a certain number of times and may or may not use it after the car is
finished.
You may be able to rent or borrow the tool.
You could buy it new or used, and then sell it afterward.
You could take the part to a machine shop and have them do it, especially if you need to have the
operation done only once.
Having a tool saves time but costs money. If youre in a hurry, you need the right tool, right now,
but getting it fast may cost more money.
I spent months debating whether to get a lathe, and all the contradictory comments above certainly
applied to me. I would rationalize one way, then the other, flip-flopping every other week. The key
was to buy a used machine so I wouldnt have to take the depreciation. (Of course, if you never plan
to sell your tools, you may want new ones anyway.) I guessed that I had about $2,000 of lathe work to
do, so that became my rough budget for a lathe. (It ended up being much more than that.) I figured
since I had to spend the money anyway, I might as well end up with a free lathe when it was all
done, instead of paying someone else to do the work. (Be aware that if someone else does the work,
he will need detailed drawings.) Eventually, I found a used 12 x 36 import lathe for $1,200. Of
course, by the time I added accessories, it cost quite a bit more.
If you buy a used lathe, a lesson learned: make sure to try and get the tooling, too. Those little
pieces will nickel and dime you for hundreds of dollars.
Some items like screwdrivers and wrenches are certainly necessary, but power equipment is
tougher to decide on. People have built cars with little or no power equipment at all (watch the
movie, The Worlds Fastest Indian). Power grinding equipment such as a 1220 disc sander is
very handy for deburring tubing but isnt absolutely required; it just makes the job go faster than using
a file. You have to choose a reasonable balance between buying car parts, buying tools, what your
time is worth, and the budget.
The tools I found most helpful were:
Lathe, with tooling: $500$A lot
1220 sanding disc: $150$1,200
Grinding and deburring wheel (very useful!): $200
Drill press: $400
Vertical/horizontal band saw (very handy; used every day): $140$200
12-ton press: $100
Tubing notcher: $40$200 (well worth it)
A good vise (surprisingly hard to find): $50$500
A digital protractor (angle finder). Used this all the time; also good for home projects.
Rulers, lots of rulers (okay, engineering scales). You cant have too many6, 18, 24, 36,
48, 60, and 72 were all used. My favorite is a flexible 24 stainless scale, with fractions on
one side and decimal on the other. I used it every day and highly recommend it.
A wall-mounted drill chart, extremely valuable every time you need to do any drilling or
threading.
An air compressorthe bigger the better. And air tools!
An auto-rewinding hose reel for the compressor; I mounted it to the ceiling to keep it out of the
way.
Lesson learned
Buy a set of reamersit was one of the best tools I bought. It is very
important to have a close tolerance fit on suspension pivot bolts. Sliding a
high-quality bolt through a properly sized and reamed hole gives a new
understanding of the word precision. By using high-quality AN bolts and
rod-ends, and a reamer, theres virtually zero play in the suspension pivots,
and thats before the bolt is tightened.
A Whitney hand punchvery, very useful for making clean holes in sheet metal, especially for
rivets.
Clecos, for holding panels being riveted. I bought about 30. Make sure to choose the size
appropriate for the chosen rivet diameter.
A good multimeter. Youll use this many times for diagnosing wiring issues.
Two fire extinguishers, one near the garage door and one near the door to the house.
Air compressors, along with vacuum cleaners, have some of the worst
advertising BS out there. For example, my 6 hp compressor runs on 240 V at
about 10 A. But at 240 V, one hp is about 3 A, so why isnt this 6 hp
compressor using 18 A? Campbell-Hausfeld couldnt answer this question,
either, and they lost a class-action lawsuit because of it.
The only way to know what youre really getting is to instead compare air
delivery rates, usually measured in cubic feet per minute (CFM) at 90 psi. Oh,
and dont use an oilless type unless you want to go deaf (or can insulate it inside
a boxoutside the garage).
Shown below is what happens when you use an ordinary socket in an impact wrench; note the crack
at the left end. I wont do that again.
The case of the imported tool
The reason we find ourselves considering an import tool is cost; we dont all have the money to
buy the best. All we really want is a decent tool for the job, and sometimes this means that a less-
than-perfect tool is good enough. Certain things just dont warrant spending much money on, like an
engine hoist or hydraulic press. How can we beat spending $90 for a brand-new imported 12-ton
hydraulic press? Just buying the steel, parts, and paint, not to mention the time, didnt justify doing it
myself. An import tool, especially one that doesnt have many moving parts, may be adequate. There
is a big difference between buying a lathe and a press, since one has many precision parts while the
other is just a bunch of steel. By the way, many domestic brand names, while implying they are
domestic (read superior quality), are actually built in Taiwan or China, right along with openly
offshore productsthe difference being the price paid for the domestic nametag.
I hear mostly negative comments from professional machinists about anything made in Asia, like
that wouldnt last a week here. I agreewhen used in a real machine shop. Yet, if this same tool is
used in my garage, it may last for years simply because I use it far less than a real shop would. If it
does the job for me, it makes sense to buy it, cheaply made or not. I dont think its fair to dismiss
places like Harbor Freight, Rutland, or Grizzly. Its easy for this machinist to be ruined by using
top-of-the-line equipment and say an import lathe is no good. Its much harder for that same machinist
to answer the question, Okay smart guy, what lathe would you buy if your budget was $1,200? The
lathes I saw in this price range were typically old (or very old) domestic units. While one day I plan
to completely restore an old domestic lathe to like-new condition, I needed something that could be
used as is. I didnt need a lathe restoration project getting in the way of building a car.
Welder
Most likely, the car will be welded together, but who will do the work? You? A friend? A shop? It
may be presumptuous to expect a friend to do all the weldingmany hours will be required. Also,
understand the reluctance due to all the lawsuits these days. Either way, the welding issue must be
resolved before you proceed further with the design. There are many good welding books available,
several of which belong in your research library. Other than some comments below, welding
technique is not covered. This is best left to the books listed in Appendix A.
Before buying any equipment, see if you can enroll in a Regional Occupational Program (ROP)
welding class. This program is a real treasure for educating yourself in many hands-on trades. In
California, at least, the program is fully funded by the state, which means the classes are free. This is
one place I dont mind my tax dollars going; the classes are excellent.
The welding class taught gas, arc welding, and TIG. I talked to the weldors where I work, asking
their opinions on what to buy. Almost everyone recommended the same type of welder: a TIG unit
(now properly referred to as GTAW). If you can purchase only one welder to build your car, this is
the one to get. With either of the two units described below, you can weld steel, stainless, and
aluminum and can even stick-weld (handy for building trailers). The other welding rig you might
consider (in addition to the TIG) is a traditional gas welder. These are very handy to have around for
many non-welding tasks such as cutting thick metal, heat-treating, heating bearing cups, brazing, and
removing those bastard-tight bolts.
In 1996, there were two entry-level TIG welders, the Lincoln Square Wave 175 and the Miller
Econo-TIG. Of course, in addition to buying the welder, the following parts are needed: a gas bottle,
a regulator, a helmet, welding rod, and gloves.
I purchased the Miller Econo-TIG, and while it was adequate for most everything, it had a
surprising shortcoming (to me). The unit is capable of 150 A at the high end, which is fine, but the
problem was at the low end, where it would only go down to about 20 A. At the end of a weld, as I
tried to taper off the current, it turned off all of a sudden, leaving a crater. The other problem was that
I couldnt weld thin material; I found steel less than about 0.05 inches thick to be a big problem, and
thin aluminum was even more difficult. To be fair, Miller does put a sticker on the unit showing
minimum material thickness. The Lincoln does a little better, with a 12 A minimum current, which
obviously allows welding thinner material and is better at crater-fill.
A neighbor was surprised that I covered my arms when welding. He had a
MIG welder and had no idea that high levels of ultraviolet light come out of the
arc (yikes!). I was amazed he didnt know. I wonder how he thought he was
getting those nasty sunburns.
One of my friends said I should get a unit that goes down to 5 A. Unfortunately, going down to 5 A
required the purchase of a machine that also goes up (along with the price) to about 250 A. When
moving away from the entry-level unitsinto real weldersthe price goes way up. Real machines
start at about $1,900 for just the welder. Total package cost is about $3,500. If you can, get a used
machine, which I eventually found, trading up for a Miller Syncrowave 250. It is the single most
expensive tool I own, but I have no regrets.
There are now solid-state welders that are much, much lighter and are coming
down in price, so keep your eyes open for deals. Paying the extra money for a
portable unit, though (in my case), didnt make sense. Also, get the water-cooled
torch and tank; the torch gets really hot if you dont! The water cooler, by the
way, is (or was) much larger than the tiny welders, so I didnt see much point of
paying for the small portable unit with a huge cooler.
Assuming youve read your welding books and taken a class, you have to decide what type of
welding method you will use. While I used a TIG unit, people have used brazing, MIG, and straight
gas torch welding. All the methods work; you just have to pick one and move on.
The nature of metal
Once you have the welder, you should practice as much as possible before starting construction.
The reason, as I found, was not taught in welding class. Welding class teaches how to join two pieces
of material together. What it does not teach is that metal is alive; if you build any of the tools and
fixtures suggested, you will see for yourself! When metal is welded, it heats, expands, and then melts.
When it cools, it tends to have a thicker cross section than before it was heated. The result is that it
contracts, pulling like crazy in the direction of the weld, which can drive you nuts if you dont plan
for it.
For example, if you place two metal plates at 90 and weld the inside corner, you will see the two
pieces move toward each other as they cool, pulling inward to perhaps 80. Either you have to preset
the angle to 100 or place it in a fixture to force it to stay where you want. The fixture method will not
work unless the fixture is much stronger than the material being welded, because the shrinkage force
is enormous. Tacking all sides of the weld firstbefore performing the full weldhelps a lot and
doesnt require a fixture. I was surprised when told to consider myself fortunate if the suspension
pickup points ended up within 1/8 of where they were supposed to be. I was glad to hear it, because
had I not known, I would have been very unhappy! Its another reason to use adjustable suspension
joints.
Lesson learned
When I fabricated the curved tube around the edge of the instrument panel, I cut
slots in the tube, bent it so the cut edges touched, and then welded itnot good.
Each time I welded a cut, it pulled the metal inward, tightening the radius of the
curve. I welded them all, and when it was done, it didnt resemble what I wanted
at all, and it required a lot of rework.
Whats nice is to have an expert welder friend who can help you, at least for advice. Hes been
through it all before and can be a big help by making suggestions or lending a hand. Someone who has
built a race-car chassis can be very helpful, too, possessing knowledge as to how tubes interact when
welded. But dont expect this friend to weld it for you, due to the liability and huge amount of time
required.
Lesson learned
I got a great tip from Alan Brickey. He said when Im welding high-precision
bearing cups (where bearings will be pressed in) to make some brass plugs first.
These are pressed into the recess before welding. During welding, the brass
prevents the cup from distorting due to welding heat, and it really works. Not
doing this would likely mean the bearing cup distorts so much that the bearing
couldnt be pressed in.
As an alternative to purchasing tools, you can make several of them yourself. Several of the tools
you will need are specific to your project anyway, so why not make them yourself? You will save
money, but more importantly, you will gain experience handling and welding steel. It is better to learn
how metal reacts while building tools than when performing critical welds on the chassis.
Weld or braze?
There are advantages to both, so see which suits your situation best.
Brazing
An inexpensive gas torch setup can build the entire car.
You can take a gas torch to the tracknot possible with an electric welder.
You wont need an electrician to install an outlet.
The metal doesnt get as hot as with welding, so the metal distorts less.
It can be stronger than welding because of the larger contact area.
It is also used for cutting thick metal, heat-treating, etc. Its good to have a gas setup regardless
what type of welder is used to build the car.
TIG Welding
A TIG welder can be used for welding everything: steel, stainless, and aluminum.
It is very precise and cleanno smoke or sparks.
It makes pretty welds.
It is much more expensive than a gas welding setup.
As said elsewhere, I suggest that no matter how you choose to assemble your chassis, take a
welding class.
Lesson learned
I found out the hard way about those magnetic helper magnets. Oh, the
magnets stick to the steel tubes just fine; the problem is when you strike an arc to
weld the tubesthe arc follows the magnetic field! It was pretty wild the first
time I tried it, the arc wanting to bow outward away from the metal, then curve
around and back down. The magnets spent the rest of the project unused, sitting
on the shelf.
Tools you can build
As a way to gain experience fabricating and to save some money, here are some simple metal
projects. These will really help you see what youre getting yourself into, but more importantly, you
will produce some extremely helpful tools for yourself that you will need anyway. Yes, I know these
are verging on being dreaded side projects, but unless you already have metalworking experience, its
time well spent, plus you end up with something useful instead of having to buy it.
Dumb helper
A dumb helper is a three-legged stand with a roller on top. The roller can be adjusted to any height
and is used to support the far end of tubing while you machine the other end. Its a simple tool, but it
pays for itself because you wont have to ask your spouse to keep coming out into the freezing garage
to hold the end of something. These stands are intended for supporting lumber off the end of a table
saw, but they are perfect for holding tubing or sheet metal. Yes, you can buy one from your local
builders supply for about $25, but you should build one yourself.
Chassis helper
Resembling a giant vernier height gage, it allows for setting parts at a precise height. I was always
propping stuff up with coffee cans, boxes, or wood blocks, which always managed to fall over right
when I needed them to stay put. The one I made uses a 12 diameter, thick steel base. Mounted to
this is a 1 x 2 four-foot tube that has a 48 ruler riveted to the side of it. An adjustable foot can be
set and clamped at any height, so I can set the far end of something such as a chassis tube right where I
want it without having to yell from the garage, Honey, can you hold this again? Shell appreciate
not bumping into things after being blinded by the welding arc.
Welding cart
In my case I didnt bother, since the welder came with one. I only use it to roll out the welder in
order to change gas bottles. When welding, I run the cables to where Im welding; theres no need to
drag the 300 lb. machine around. With a six-foot heavy-duty cord, how far can it move anyway?
Engine/transmission caddy
This is a low-wheeled stand for the engine and transaxle. As the engine and transaxle combination
weighed 450 lb., they needed to be moved around easily, yet be able to be pushed under the counter
when not in use. It was designed to support the drivetrain at the height at which it would be installed
in the car, which conveniently allowed rolling it out from under the bench for mock-up purposes. The
stand supports the drivetrain but not by using the engine mounts. This allows the drivetrain to be
rolled into place while keeping the engine mounts free to be bolted into the chassis. That way the
drivetrain could be removed or installed without any help, which can be very convenient during late-
night repair sessions.
Grinding center
A grinding center is a wheeled stand to which a 12 sanding disc, belt sander (which I hardly ever
used), chop saw (which I didnt use much either), grinding wheel, and deburring wheel are mounted.
Valuable workbench space is freed up, and metal grit doesnt get thrown all over the work area. It can
be rolled around to where youre working to avoid wasting time walking back and forth. A plastic
storage box below holds all the special wrenches, sanding discs, buffing wheels, and Allen wrenches.
The small-diameter, light gray wheel in the middle of the picture is the magic 3M deburring wheel,
the best $50 I ever spent. That single wheel deburred every metal component in the entire car.
Band saw stand
This is a stand for the very popular, sub-$200, import 4 x 6 horizontal/vertical band saw,
replacing the flimsy legs that come with the saw. It even had a chip tray and coolant pump. Instead of
buying an expensive coolant tank and pump, consider buying a small fishpond pump from a home
improvement store. Since many coolant mixes are water soluble, these pumps work fine. One note:
dont buy a magnetic-impeller type pump. The magnet will only serve to attract all the steel bits
suspended in the coolant and clog up the pump. A link to the plans for this stand is at
www.homemetalshopclub.org/projects/sawstand/sawstand.html.
Engine hoist
This is literally a heavier project that the others, but it teaches how thicker metal behaves. Go to a
local store where they sell hoists and make a sketch with dimensions (when the salesman isnt
looking). Building one of these isnt hard but requires thought if you want it to break down for
storage. An engine hoist isnt just for engines. I used it to move the lathe, to lift the chassis table, and
even to move the entire tube chassis. (I only suggest this project; its very hard to justify building one
when you can buy one for less than $100.)
Chassis table
One way or another, the chassis must be built upon a surface, which can range from the garage floor
to a heavy-duty ground-steel chassis plate. As tempting as it was to just get on with it, I knew I
would regret skipping this step. Whatever money and time that would be saved by not building a
chassis table would be rewarded with an out-of-square or warped chassis. What I hope is already
apparent is how unpleasant building something on the floor is. If you havent experienced the
frustration, I hope you can take my word for itbuild a table.
After finishing the car, I built a traileron the floor. My knees hurt, my back
hurt, and I couldnt get under or around all the joints to weld them. Using my foot
to activate the TIG welder control, while on my knees, was very awkward.
Never again! No matter what I build next, itll be elevated, even if its just a
couple of sawhorses.
Regardless of how fancy the table is, simply getting the work surface up to a convenient height pays
you back many times in the quality of the welds, the accuracy of the finished product, and your ability
to work in comfort without a sore back. No matter how the table is constructed, its better than
bending over a tube on your knees or trying to weld upside down. One way or another, you need a
table. The question is how nice you want it to be.
My buddy Cecil built what must be considered the first-class chassis table. It used 6 x 8 I-beams
welded in a rectangle, which was then ground flat, supporting 2 ground square tube crossbeamsall
ground to within 0.010 of each other. With a heavy coat of paint except for the ground surfaces, it is
truly a beautiful table. The advantage of this type of table construction is that everything can be held
down with C-clamps, anywhere on the surface, since the crossbeams could be moved.
Several alternatives to this type of table are:
Buy ready-made maple workbench tops and add heavy-duty legs.
Build a poor mans chassis plate by using two sheets of plywood separated by 12 ribs,
forming an extremely stiff surface.
The problems with wood tables are several: they catch fire(!), the tubing youre welding isnt
automatically grounded, and humidity affects them quite a bit. The other problem is that you have a
continuous surface with no access from below. To clamp the work down requires some sort of
fixture. A C-clamp cant be used like on a metal table, so youll have to drill large holes wherever
needed or hold it down with wood braces, screwed down on each side of the tubing. The wood table
that is supposed to be much cheaper may end up taking longer to build and cost more. It all depends
on how cheaply the wood can be obtained.
I was very fortunate to find a large 5 x 7.5 steel table made of 3 tubing and a 3/16 steel top it
was going to be thrown away! It had the disadvantage of having a solid surface, which meant I could
not come up from below to work from the inside. But since it was free I couldnt turn it down. Two
4 I-beams (H-beams to you Brits!) were placed on the table, and cross-tubes placed on top of those.
This was so C-clamps (G-clamps across the pond!) could be placed anywhere necessary. I welded on
legs with caster wheels and leveling pads to create a work surface 20 from the floor. The height of
20 was chosen so that:
while seated, I could weld at a comfortable level.
while standing, I could perform welding on the top tubes without having to climb up on the table.
reaching or leaning to the center of the table could be easily achieved.
The casters were nice because the table could be spun around for easy access. The leveling pads
were required because of the sloping floor most garages have. Its very important to have a level
work surface; otherwise the whole car will be tilted. Youll forget that when welding on a vertical
tube using a T-square as a guide, the tube will end up with several degrees of lean! Using a digital
level to check if things are 90 to the bottom tubes will introduce errors if the table is tilted (unless a
carpenters angle is used).
The most valuable tool
Without a good idea of what you are doing, having the best tools wont help. People have built cars
without even a hand drill and succeeded, yet without knowledge, tools are useless. The skills you
bring to the project are more important than any tool. Take classes, talk to people, and go to races and
hang out in the pits. Subscribe to and read car magazines. Read everything.
There is one more tool that was very useful: cardboardyes, cardboard,
like the backing on a pad of paper. Its perfect for making mock-ups and
templates; once cut to shape its easy to trace onto metal. Collect as much as you
can before starting the project, because its used everywhere, and every day. I
found mine in my companys repro center; they throw out big sheets of the stuff.
Lesson learned
When I started the project, I figured I could purchase parts in the order I
needed them. Not buying anything I didnt need right away would keep costs low.
It didnt work out that way. As I started in on the design I realized I couldnt get
far without knowing the exact dimensions and weights of the major components
now. Since I thought I could finance it as I went along, this was quite a shock. It
meant everything had to be purchased before the thing ever got off the drawing
board; this was a real kick in the pants. It slowed everything down because I had
to save up first to buy everything. To add insult to injury, after I bought
everything, all I had to show for it was a pile of parts in the garage and no plan to
put them together! Unfortunately, this is what it took to make sure it went together
right the first time.
7: Tires
BRUCE MCLAREN: Wheres my race car?
JOHN COOPER: In that rack of tubes, son.
I could never confirm if this is an actual quote, but it fits our situation!
So fasten your seatbelthere we go. We will design the car in this order:
1. Tires
2. Drivetrain
3. Uprights
4. Wheels
5. Roll center
6. Suspension
7. Springs
8. Shocks
9. Steering
10. Anti-roll bars
Tire choice is the first, and most important, decision of the entire project; everything is related to
what tires are chosen. The physical size of the tire is most critical for suspension design, followed by
expected grip. Once tires are selected and the suspension designed, the tires should not be changed. If
they are, the suspension design must be revised to make any corrections required by the tire size
change. This is because virtually every aspect of the car can trace its roots back to the tires. Changing
the tires can mean the entire suspension has to be redesigned, so its important to use care in choosing
a tire, and to stick with your choice.
What looks right is right
What tires fit the cars proportions? For a Mini, only small tires, 1013, look right. Any of the
huge present-day tires are way too big and would look ridiculous.
Tire clearance
Given the Minis shell, the chosen tires must have room for suspension travel and steering.
Affordable
Tires are the most expensive consumable, and I wanted to keep costs under control. The ridiculous
price of big tires was a huge incentive to stay with small tires. It all depends on your budget; mine
wasnt that big, and I didnt want a huge expense with each visit to the tires shop.
I owned a 1995 Chevrolet Camaro Z28 (the most unreliable car I ever owned,
but I digress). In one autocross practiceabout twelve laps half the tire tread
was worn off. At the time, that worked out to about $500 in tire wear, or $45 per
lap. It was a good example of how manufacturers make big, fat, heavy cars
handle. They cover all of their sins by using huge sticky tires, and the owner is
stuck with paying for them, forever.
ABS and traction control made sure the tires werent wearing due to spinning
the tires; all the wear was due to the weight. It was the first and last time I
autocrossed that car. Kimini has been to three practice events, two trackdays, and
several thousand miles of aggressive street drivingthe tires are still doing
well. Weight is the single biggest tire wear issue in a sports car.
My brother, Scott, dealt with the same issue when choosing tires for his Super-
Stalker. He went with smaller tires than he would have liked because he was
unwilling to spend a lot of money on tires that might last only a year. Its all about
compromise.
Polar Moment of Inertia (PMOI)
Less is better. Try spinning an 1824 wheel and tire. Its hard to start it spinning, hard to stop it
from spinning, and hard to make it change direction. Now take a 13 tire that weighs the same and do
the same experimentit will be literally twice as easy to spin. This isnt imaginary; going to smaller-
diameter tires frees up real horsepower, and it isnt a trivial amount. I expect 510 whp (wheel
horsepower) to be freed up because of reduced PMOI. The 13 wheels and tires also weigh much
less than 18s, so the gain is doubly important. And finally, theyre much cheaperdont forget the
never-ending expense! The gyroscopic effect of large diameter wheels and tires also conspire to make
it hard for the car to change direction quickly.
Oil pan clearance and ride height
I didnt want to have to drive around speed bumps or risk the car getting high-centered, so the car
couldnt sit lower than about 4.0. This value was found scientifically (using a ruler to check several
local speed bumps!). Since the bottom of the oil pan and transmission would be the lowest point of
the drivetrain, they would dictate how high above ground the engine would sit. This in turn fixed how
high above ground the axles exited the transmission. At normal ride height the axles should be fairly
straight (to minimize heat and wear), which sets the radius of the tires and dictates the tire diameter.
Some people cut down the oil sump, making it shallower to increase ground
clearance. This is fine as long as it fits the goals of the car. If its a cruiser, great,
but if the car will see track duty, it is a big mistake. A shallow pan means the
depth of oil in the pan is less, so when the car corners, its more likely that the oil
pickup tube will become uncovered, causing the oil pump to suck air. No engine
will last long that way.
Brake disc size
The idea was to use the largest that would fit inside the chosen wheels, preferably with the same
brakes that came from the donor. Given the lightweight nature of the car, and not having 500 hp, I felt
that stock production-sized brakes (made for a car twice as heavy as mine) would be fine.
Effective gear ratios
If different diameter tires are used (compared to the donor drivetrain), it changes the effective final
gear ratio. Using smaller wheels and tires means the car will have a lower top speed. I had no desire
to go 150 mph in a Mini, so if I lost 20 mph off the donors top speed, that was okay.
Due to this, my top speed was reduced from around 145 mph (in the donor) to
about 130 mph.
DOT-approved tires
Since the car would be driven on the street, I would have to use DOT-approved tires, limiting the
selection to only a few manufacturers, which isnt so bad. Since 13 tires are used mostly on race
cars, even the street tires are very stickya good thing. The theory was that since the car would be
driven perhaps once a week and either trailered or driven to distant events, even sticky tires should
last a long time. Tire wear should be much less for the same performance as a larger, heavier car.
And the answer
BF Goodrich Comp T/A R1s, 185/60-13 fronts, and 215/50-13 rears. They looked right, fit the
wheel wells nicely, gave sufficient wheel travel in bump and droop, and had minimal rotational
inertia. I didnt bother researching the fastest tires; there would be plenty of time later to tune the car,
and while R1s were the best at the time, some other brand would be better later on.
The tires are 21.5 in diameter, with a 10.75 radius. This is important because it dictates the
height above ground that the axles exit the transmission. (Now you start to see the interconnectivity of
everything!) With the transmission outputs roughly fixed, the final location of the bottom of the oil pan
(presumably the lowest point on the car) can be determined.
Lesson learned
If the selected tire size cant be found, the cars not going anywhere! In my
case of 13 tires, I didnt (and dont) have much of a choice, due to the Mini
dictating the size.
My saving grace is that 13 tires remain popular with small race cars.
Companies like Hoosier, Kumho, Hankook, and Yokohama continue to make very
sticky race tires and barely legal DOT-legal street tiresperfect for Kimini.
Unfortunately, tire sizes seem to only be getting bigger and bigger every year.
If I were doing it over (on a different car), I wouldnt go smaller than 15 tires,
or even 16, and would try to find very light wheels to make up for the higher
weight. PMOI would be higher, but what choice is there?
As a hint of things to come, BF Goodrich ceased production of their 13 R1
tires soon after I bought themfigures.
Later on, as the original tires wore out, I was forced to confront the slim
selection of 13 tires. If my car were a real race car, it wouldnt be an issue,
since I could just run slicks or DOT tires. If it were only a road car, it would
be tougher, but still possible, to find 13s. But Im trying to do both at once,
putting me in a tough middle ground. Super-sticky 13 tires wont last on the
street, and hard compound tires (if I can find them) wont do well at the track. So
I may be forced to have two sets, which brings up the issue of size. Toyo (RA-1),
Kuhmo (V700 and V710), Yokohama (A032R), Hoosier (R6), Hankook, and
Azenis make very sticky 13 and 15 tires. Most of these are very sticky street
tires, but the 13 selection is thin and sure to get worse. So what tires are used
depends what the cars designed for.
Eventually, as 13 tire supplies dry up, Ill be forced to go to 15 tires, at least
for the street. The 13 wheels would be retained for the track, with really sticky
tiresif not slicks. Even the lowest profile 15 street tires available will make
the car sit higher, but since its for the street, its not a big problem.
8: Drivetrain
Drivetrain weight is a large contributor to how much force is applied to the suspension. Choosing
the drivetrain now, before designing the suspension, lets us know what type of axles connect the
transmission to the wheels. In my case, I chose to use a FWD drivetrain, relocated to the back of the
car, so I knew what axles would be used. This in turn determined what outboard wheel bearings were
needed in the rear uprights, but Im getting ahead of myself. We need to select the drivetrain before
going further.
Cost
It had to be affordable, and for a variety of reasons I put the ceiling at $3,500 (1996).
Thats way too much
I didnt want to rebuild or otherwise mess with the engine. While I could have
bought a blown-up or worn-out engine, Id be spending more money on a side
project rebuilding it. Side projects are bad news; the only way to finish a project
is to keep at it and to stay focused, not spinning off into innumerable time-sucking
distractions. I was designing and building a car, not rebuilding an engine. There
would be time for building a monster engine later on.
Another reason was that back in 1996, there were very few inexpensive high-
power FWD drivetrains. The few that were out there were relatively new and
expensive. If I wanted to play I had to pay.
Availability
The engine had to be somewhere close enough to pick up, or at least ship inexpensively, and its a
really good idea to see it run first. I bought mine sight unseen, a risk that more or less worked out.
Power
After building the Datsun 1200, I knew what it was like to start with a small engine, one that wasnt
quite enough (insert your definition of what that means). I was designing and building a car, not an
engine, so the goal was to find a drivetrain that, right out of the box, provided sufficient excitement.
A common forum topic is comparing a 150 hp 1.0 L sportbike engine and a 150 hp 2.0 L car
engine. The car proponents say the car engine, with its higher torque, will accelerate fasterno.
Lets say the above one-liter sportbike engine makes 80 ft. lb. of torque, while the car engine
makes 160 ft. lb. The bike engine revs to 12,000 rpm while the car engine revs to 6000 rpm; the key is
the gear reduction in the bikes transmission. The bike engine makes half as much torque but spins
twice as fast, so after its geared down 2:1, it makes the same torque at the same rpm as the car
engine. And thats exactly what the horsepower equation takes into account. Bottom line
horsepower is horsepower: hp = torque * rpm / 5252
Of course, the torque curve of the car engine will be wider, so less shifting is needed, which may
or may not translate to faster acceleration. But the bike transmission can be shifted much faster than
the car, so its not easy to say which combination will be fastest. The bike engine and transmission
also weigh several hundred pounds less which can be a big percentage of the cars weight.
Just to throw in another variable, the bike engine will run at a much higher rpm in nearly all
conditions. Whether you consider that to be noise or music is up to you.
Another requirement of my engine was to have enough power so the rest of the car could be
concentrated onenough power for, as someone put it, shameless smoky burnouts.
How did we ever come to equate tire smoke with fast? Smoking tires means
there is a lack of traction, and lack of traction means the car is not accelerating as
fast as it can. Take a pickup truck, rev up the engine, and pop the clutch. Okay,
the wheels are spinning; does that mean it is fast? No. I like racing someone who
smokes the tires off the line; why would I want to race someone with traction?
Ive often wondered why no one makes a drift car out of a pickup truck; its a no-
brainer, and it would be cheap. I think its because it would embarrass cars
costing far more. Drifting takes great skill, but dont confuse it with being fast.
Reliability
An unreliable engine would make the car no fun at all, and building up a small engine, to make it
acceptable, was not a goal. Doing so would make it less reliable, more expensive, and a dreaded
side project. I remember when my brother and I used to run HPDEs where many cars, after they came
in off track, immediately opened the hoods. No, as said before, the goal was Honda reliability.
Some modern engines can be difficult to use in a car project, and its only
going to get worse. Late-model cars have more electronics than even 10 years
ago.
Case in point: the 2003 Audi TT; it has a really nice engine that I might
consider using. The problem is how invasive the ECU is. For example, the
ignition key has a chip in it that the ECU reads before the car can start. If this
drivetrain is put in a project car, how could it be started? It could be a real pain
incorporating the factory ignition switch. Worse, the Audi (and others) now use
the CAN bus, a serial network connecting almost everything in the car. Even
things like taillights are on the network. What this means is that you may have to
use an aftermarket ECU to run the engine. While thats not an entirely bad thing, it
adds time and expense, both for the ECU itself and the associated tuning,
something many people fail to include in budgets.
Volkswagens latest ECUs apparently calculate how much torque the engine
should be making (the engines are drive-by-wire, a real can of worms for a
homebuilt car). If the owner increases turbo boost, the ECU will be none too
happy, dialing out throttle since it thinks the engine is misbehaving. Once again,
an aftermarket ECU will be needed. So, if a newer engine is being considered,
add into the budget the potential expense of an aftermarket ECU and the tuning of
it, which could easily total $1,000 $3,000 if a lot of dyno time is needed.
Regardless what drivetrain is considered, try to find out how invasive the ECU
is, to see if it will work with you or against you. Fortunately, I found a 1995
engine, the last year before OBD2 was implemented.
Availability of a Limited-Slip Differential (LSD)
Regardless of drivetrain, an LSD was required. Why? In my opinion, it is right up there with tires
and suspension modifications for the effect it has on a twisty course. Building up a car, only to spin
one tire coming out of corners, would be a real disappointment. On the other hand, the new Lotus
Elise does not have an LSDthe factory says it doesnt need one due to its rear weight bias, and
talking to an owner who ran his on track confirmed this. But other owners say that an LSD is needed
for autocross (due to the sharper corners and increased weight transfer), so it depends on the
applicationwhat youre building and what its forbut regardless, its good insurance to make sure
an LSD is available, just in case. For me, a Quaife LSD was a must-have.
Why a Quaife? Why not a clutch-type LSD?
Personal preference. I drove the Datsun for years using a clutch-type LSD, and
it worked perfectly. I just feel that the Quaife is more reliable, works better, and
will be more consistent over the years, but thats just me. (It also has a lifetime
warranty.)
It has to fit
It all depends what you are building! Starting out with a Mini shell narrowed down the choices; a
big engine simply wouldnt fit. There are a couple of V6 drivetrains that might fit, but they wouldnt
fit the lightweight character of the car, violating my self-imposed design goals of weight, handling,
cost, and tire usage. No, the car needed to stay true to its character: a small, lightweight fighter.
Cramming a huge engine in the back, along with everything that goes with itbig wheels, big tires,
big brakes, a stronger, heavier chassis, a big heavy battery, and a bigger fuel cellwould make things
get out of hand. Adding in all this stuff would make the weight skyrocket, and the center of gravity
would move away from its target location.
The chosen engine had to fit behind the front seats in the Mini; this perhaps more than anything else
dictated the choice of a four-cylinder. Once the chassis was sketched out, only about 19 existed
between the seat bulkhead and the rear-axle centerline, ruling out large engines.
I was amazed to read that a builder welded together his Locost chassis, then
tried fitting a DOHC V6 into it, and he was completely surprised when it
wouldnt fit! I was dumbfoundedthats what a tape measure is for. Dont
assume a given chassis will fit any engine on Earth (is this not obvious?).
Measure everything first!
Using a four-cylinder engine has many advantages, but it also has secondary benefits. With a light
engine, the chassis can be lighter, the battery smaller, the starter cables thinner, the axles lighter, the
brakes smaller, the wheels and tires smaller, lighter, and cheaper, the gas tank smaller, etc. It all adds
up; a light engine saves far more weight than just the difference in engine weight alone.
Since the Mini is such a small car, the width of the drivetrain was also limited. With the wheel
locations already fixed by the shell, it was clear that a wide drivetrain wouldnt work. Well,
technically, a large drivetrain might fit, but it would need really short axles. This is a real issue,
because as the suspension moves up and down, the A-arms move through an arc, as do the axles. As
the suspension compresses, the axles move inward toward the transmission. As the suspension
droops, the axles move outward. Modern CV (constant velocity) joints have plunge assemblies to
allow for thisto a point.
I had to shorten each axle by about four inches, which increased the angular
motion of the axles quite a bit. It required extensive measurement to ensure the
axles would neither bottom out, nor pull out, of the CV cup. It worked out okay,
but it was close.
So where does all of this lead? It means the chosen engine had to be affordable and reliable, and it
had to fit, have enough power, and be new enough that it didnt have to be rebuilt. Just drop it in
and gothat was the goal. But it still didnt answer the question of which engine. A large engine
would certainly bias the weight so far to the rear that it wouldnt handle well. A really heavy engine
in the back risks not being able to turn under accelerationtons of understeer. Worse, if you really
stomp on it with a big engine in back, you may lift the front tires. Thatll impress people but makes the
car useless for anything but drag racing. If the tires are in the air, how can you steer?
I thought of using a VW VR-6 drivetrain for a while, but I suspected that it was too heavy. Too bad,
because it was quite attractive, combining the small package of a four-cylinder with the power,
torque, and sound of a V6. Another big plus was that the exhaust is on the backside of the head.
I learned later that a complete 2.8 L VR6 drivetrain weighs about 530 lb.,
which was indeed too heavy.
This revealed another problem: the complete lack of information concerning drivetrain weight and
size. Without size and weight data, I had little basis to make design decisions. Logically, the larger
engines would weigh more than four-cylinders, but it was like designing in a vacuum. Not knowing
the weight meant my design was not very intelligent. It meant the drivetrain had to be purchased
before knowing weightsthe project could get very expensive buying parts by trial and error.
Finding out that a part wouldnt work, after purchasing it, was to be avoided.
This is not quite as much of an issue as it was back in 1996. Back then, there
was no information anywhere to give me an idea what various FWD drivetrains
weighed. At least now there are various Web forums where you can ask, and you
might get a good answer. On eBay, one sneaky way to find out is to ask sellers
what the shipping weight is.
There is a shareware program called CarTest2000a free drag-racing simulation program
containing many production cars. The beauty of it is that you can easily try what ifs without
spending a cent. Want to see how fast your car will be with double the horsepower? No problem, just
change it. Next, try cutting the weight of the car in half to see what happens. Now move the drivetrain
to the mid-engine position and watch how traction changes. Experimenting illustrates how low weight
makes a car just as fast as adding expensive horsepower, and low weight also improves cornering
and brakingadding horsepower does not. CarTest can also give an idea how fast the car youre
designing will bepretty cool.
In 1996, after much research, it came down to these candidates:
Toyota MR-2, 1988 supercharged, or the later turbo
Integra GSR
VW VR-6
Mitsubishi turbo
Nissan SE-R
Nissan Maxima V6
Honda Prelude VTEC
As was mentioned before, with no source of weight data, even from wrecking yards (who ship
them and should know), I had to guess.
The list started getting shorter when availability of the Quaife LSD was factored in. Several of the
drivetrains already had factory viscous limited-slips, but I read some people had trouble with them
overheating on-track. Any money saved by buying the drivetrain with this LSD would later be wasted
replacing it. Thus the SE-R and Maxima eliminated themselves; the SE-R was also underpowered,
and the Maxima was too heavy.
Because autocrossing was high on the list of requirements, turbo lag would be a problem. Coming
out of corners requires immediate power, so having an engine off-boost was not desirable; talking to
several national autocross champions seemed to confirm this. This removed the MR-2 and Mitsubishi.
The old supercharged Toyota MR-2, while cheap and light, was a risk because of its age, and only
having 140 hp (after a rebuild) put it in the same situation as the SE-R. The VW VR-6, while
attractive, had no LSD (though it does now) and had the weight handicap of an iron block.
In order to reach my 180200 hp target, some of the contenders could cost a lot to build up. It
seemed like a waste; why not just start with a larger engine to begin with? I had already had enough of
high-strung little engines with no torque. I wanted something with some grunt, something that right out
of the box would be plenty fast. On the other hand, weight was a big issue, and it was very frustrating
having to decide based on guesses.
After much debate, a 1995 190 hp Honda Prelude VTEC H22A1 won out. It was fairly light (in my
fantasy world I somehow decided it weighed 350 lb.), and it had good power and Honda reliability.
At the time, they were hard to find, and when I finally found one, I had trouble with the vendor.
(Expensive lessonget all the parts up front.)
When the big box finally arrived, boy was it heavy! Reality slapped me with a 475 lb. lump
much heavier than what I was hoping for. Some of the weight was due to internal balance shafts, so at
least it would be smooth! While I knew the Integra drivetrain was lighter, I was never able to
determine by how much; regardless, the torque and power of the larger engine was a goal. Having
already chosen the tires, running the numbers showed that top speed would be about 130 mph with the
7300 rpm engine red-linenot a problem.
Choosing the Honda drivetrain paid off big-time, due to the huge aftermarket
and on-line forums. This was a big help when things werent going well,
especially with the ECU. (A big thanks go to the guys at Honda-tech.com.)
After all unneeded accessories and brackets were removed, the factory header replaced with
lighter stainless steel, an aluminum flywheel installed, and the wire harness cleansed of unneeded
wires, the final installed weight was reduced to about 450 lb.
A FWD engine in the back; wont the car run backward?
A very frequent question! There are two ways to look at it:
1. Pick up a FWD engine from the front of a car, move the engine six feet
straight back and set it back down.
2. Pick up a FWD engine from the front of a car with an engine hoist; push the
car six feet forward and set the engine back down.
Think of it from the engines perspective; its still turning the wheels to move
the car forwardall I did was slide the engine back. Forward to the engine is
still forward to the car.
The H22? The Honda K20 would be much better.
Back in 1996 there was no K20; the H22 was the engine to have at the time. If
I were starting over today (2005) I would probably use the K20. The K20 has
some very desirable features; most notably, the engine is reversedthat is, the
intake is on the forward side and the exhaust is on the backside (very important
for heat reasons). The engine is on the passenger side of the car, nicely balancing
the drivers weight; its smaller and has more power. It also rotates the right
way, meaning if a dry-sump were needed, it would be much easier to source. Of
course, in another few years there will be a superior engine. And as of 2006,
K20 engines are still expensive. (Said another way, no ones volunteered to fund
a drivetrain swap!)
Why not a motorcycle engine?
I seriously considered this back at the beginning of the design, but there were
no sportbike engines with decent power. In hindsight, given the more powerful
sportbikes of today, the following issues must be confronted:
Frantic behavior. Yeah, a Hayabusa has 170 hp or so, and an R1 about
the same, but they are frantic engines. Do you mind cruising at 60 mph on the
freewayat 6000 rpm?
Reliability. I cant believe that asking a high-strung sportbike drivetrain to
push three times the weight it was designed for is going to be good for it.
No reverse. Quaife has a perfect solution: a motorcycle-to-IRS axle
conversion unit with both a limited-slip and a reverse gear. The price, not
including axles or CV joints, is $3,200.
You do get a sweet sequential gearbox and a high-revving engine, but taking
all the above into account, I didnt think it was a good choice for me. On the other
hand, if Kimini had weighed 700 lb., then it would be worth considering, but that
weight goal wasnt going to happen. Having said this, after I got a ride in a
Megabusa, a Hayabusa sportbike-powered Super-7, I felt that if the car is light
enough (<1000 lb.), it is worth consideringas long as you like frantic. In this
context, frantic is like hanging out with your eight-year-old nephew who keeps
drinking sugar-laden, caffeine-infused drinks. The nervous, bouncing-off-the-wall
antics can get a bit tiring.
9: More Power
This chapter is here because of consequences that may not become apparent until the car is driven.
In other words, this chapter will give you a glimpse of what might happen (to you) after the car is
done.
After the car is conceived, after it has been built, after the first tentative drives, and after its finally
driven hard on the track, there may come a time that you wish for just a little more poweror
maybe a lot more. This was not meor so I thought. The situation arose after I had driven Kimini on
track for the first time. It handled so well that my passenger, Max Nealon, observed, You know, it
does so well in the corners, you could use more power. Oh how I wish Max hadnt said that!
A more typical reason a builder finds himself in this situation is if the finished car is heavier than
expected. The expected power-to-weight ratio will be wrong (always less than desired), leaving a
vague longing for more.
I picked the Honda Prelude H22A1 engine because it had lots of power and torque. In 1996, it was
the engine to have. It could be dropped right in and I wouldnt have to worry about the car being fast
enough when it was completed. This way I could concentrate on building the car and not worry about
adding go-fast parts to the engine. But the draw to the dark side, of more power, can be strong. Its a
very subjective issue that is not at all logicalmentally or financially!
Just what is enough power? Kimini can do a sub-four-second 060 and a mid-12-second quarter-
mile with the stock engine; why would I want more? At track events, some pretty powerful machinery
shows up. It isnt unusual to see 500 hp Subaru STis or Mitsubishi Evos. In addition, a real
troublemaker is the Corvette Z06thatll never do.
I chose a horsepower goal to match or beat the power-to-weight ratio of the 2007 Corvette Z06. It
meant that Kimini needed more than 270 hp at the crank, or greater than 230 hp at the wheels. Very
fortunately for me, a coworker was none other than Paulus Lee, a very well-respected Honda racing
engine builder. I had several long conversations with him regarding my horsepower goals.
In either case, the Honda H22 block immediately posed a problem. Both the cylinder walls of the
aluminum block and the pistons are coated so that they are compatible with each other. Whether the
stronger engine is normally aspirated or forced induction, it will require different pistons to either
raise or lower compression, and there are no compatible aftermarket pistons. Recently though, Mahle
started making pistons specifically for the H22A (a large cost savings since the block wouldnt need
sleeving).
There are also compatible Honda-manufactured S pistons that raise
compression from 10:1 to 10.6:1, but that isnt enough for an all-out NA engine.
An H23 crank can be swapped in, too, but it wont last long in a high-revving
engine.
The cylinder wall issue isnt true of all Honda engines, so some lend themselves better than others
to modification, and if I were doing it over again (knowing that I would be modifying the engine), I
would consider more candidates.
If Id had a bit more forethought, I may not have chosen the H22 drivetrain. Dont get me wrong;
its completely reliable and capable of moving the car very quickly. But if I had known that I would
become weak, tempted by needing more power, I may not have chosen the H22, which isnt the
easiest engine to modify. Ironically, the problem is that Honda already extracts the easily obtained
power. As delivered from the factory, it provides 190 hp of very reliable power, but trying to
increase that power becomes expensive and difficult. It all depends where you want to draw the line.
With the H22 already in the car, I didnt want to swap it for something else; it was more work than
I wanted to deal with. So even though the H22 wasnt the easiest to modify, I pushed forward to see
what it could do. It did have the big advantage of having good torque already. To do even better, the
two avenues of gaining power were examined: normally aspirated (NA) and forced induction.
Normally Aspirated
Achieving 240 whp is a tall order. Every modification available must be used to reach that target:
increased rpm, displacement, compression, and improved breathing. Its easy to list off what to do but
it is very expensive to implement. The biggest challenge is to make a reliable engine that can rev
higher than stock. That alone turns out to be the most expensive target to reach.
Chapter 9: Table 1
Many people may disagree with this large expense. Granted, it is a big, scary number, but this is
what it takes to make reliable power. Some people will say that some of these modifications arent
necessary, and perhaps they arentfor a street car. But this engine will see many hours of use on
road-racing tracks, so it has to be very reliable. Because of the high rpm, it affects every moving part
in the engine; virtually everything has to be replaced with higher quality parts. Unfortunately, this
drives the cost up very quickly.
Reliability costs money, and like the saying goes, Fast, cheap, reliable: choose two.
Forced induction
Forced induction, through supercharging or turbocharging, requires some of the same
modifications, with one big exception: since the engine doesnt need to rev any higher than the factory
red-line, the valvetrain and head can be left stockthis saves a small fortune. The pistons still have
to be replaced, both because lower compression is needed and because the H22 has infamously weak
ring-lands; any boost over 7 psi will crack the pistons if any detonation occurs. Another benefit is that
the stock forged steel crank is strong enough to handle up to about 400 hp, so that can be left alone,
too. Lastly, once a turbo engine is set up, additional power can be obtained by turning up the boost. In
the NA engine, we went to great expense to exploit every gain in power, and 240250 hp is about the
best we can do without spending $10,000 and up.
Chapter 9: Table 2
Additional expenses, in both cases, may include injectors, a larger fuel pump, axles, a clutch,
plumbing, etc. Since these expenses are the same for both approaches, they are left out of the tables
above. But regardless of whats done to the engine, tuning is not an option that can be left out!
Note how the expense of the turbo engine is significantly less than the normally aspirated engine.
This is entirely due to the turbo engine remaining below the factory rpm limit. Whatever stresses the
turbo engine has to deal with, the higher engine speed of the normally aspirated version is at least as
bad, and I suspect worse, for engine life.
Lag
Lag, not a problem in a supercharged application, can be a real issue in a turbocharged setup. The
term refers to the rubber-band feel of the gas pedal; when more power is asked for, there is a delay
before it is produced. It is caused by two things: the time needed to spin up the turbocharger and time
to compress the air springthe column of air between the turbocharger and engine. The way to
minimize lag is to use a small turbo, and to minimize the length of compressor-side plumbing, employ
a short exhaust manifold to the turbo, and have a low-restriction exhaust downstream.
Lag on a road course
While lag is not so important in drag racing, it is a big deal in the road-racing department.
Nowhere is throttle response more important than mid-turn on a race track. The tires are on the very
edge of adhesion, balanced between cornering and accelerative forces; the driver is typically steering
with the gas pedal as much as with the steering wheel. If the engine doesnt do as he wantsright
nowit can cause the car have an overall slower lap time, even though it has more power.
Another turbo-related issue is heat, the enemy of all turbocharger installations.
Someone wrote me to say that a dyno showed his car put out 300 whp, yet right
after he came in off track, it was only 220 whp. The point is, unless the additional
heat can be controlled, the actual power output may be much less on track than
expected.
In hindsight
I dont mean to be critical of the Honda engine, and Im happy with my original drivetrain choice. I
am, however, a little annoyed that I find myself on that slippery slope known as, Its never enough.
The drivetrain you choose ultimately depends on what you can afford and how much power you
want nowand later. If there is any suspicion that more power might be in the cars future, evaluate
whether the drivetrain can handle the increase, before buying it. Some engines take to it like a duck to
water, while others dont. For example, the four-cylinder turbo Dodge SR4-T drivetrain provides 260
hp right out of the box. With just a couple of tweaks, 300 hp is easy. On the other hand, a Honda
S2000 drivetrain is already well optimized from the factory. Extracting any additional power is very
difficult (and expensive), due to the tremendous job that Honda has already done.
The point is, think carefully about what you want now and also about what you might need in the
future. Choose wisely.
There is another way
Instead of spending thousands on engine improvementsresulting in a less-
reliable power plantconsider learning to drive! Have you ever had your car at
the track and let the local ace take it out for a few laps? The driver will likely
beat your lap time by several seconds. Where did that extra speed come from?
Its the same engine and the same horsepower, but he drove it as though it has
another 50 hp. Theres a lesson there.
Adding more power will make the car faster, but improving your driving
skills will make the lap time drop far more than adding expensive, unreliable
power.
Data logger
Dennis Palatov, of dpcars.net, has driven over 5000 laps at Portland International Raceway. He
said that the single most useful component he ever added to his cars was an onboard data logger. The
unit above is from race-technology.com (image credit Race Technology). At around $1,000 (2007),
its far from cheap, but as Dennis explained, it did more for lowering his lap times than 100
additional horsepower did.
It records not only G-forces in every direction, but also engine parameters. With its GPS receiver,
it logs and stores the exact shape of the track, so you can see exactly where you are on each lap. The
effects of trying different braking points or turn-in points, or a different apex, will be graphically
plotted to show whether it helped or hurt track time, and by exactly how much.
Of course, having both the skill and more power is the best, but power alone is often a crutch for
avoiding the admission that our driving skills arent what they could be. Im going to take some
driving courses, regardless of what Ive done to the engine.
As this book goes to print, I decided to hold off on building a second engine.
For now, Im going to concentrate on improving my driving skills.
Cylinder liners $1,000
Pistons $ 500
Rods $ 500
Stroker crank $1,000
Port/polish head $1,500
Valves $ 500
Valvetrain $ 500
Cams $ 500
Cam gears $ 200
Good header $ 800
Upgraded intake $ 500
Upgraded oil pump $ 200
Total $7,700
Cylinder liners $1,000
Pistons $500
Rods $500
Turbocharger $1,000
Intercooler $300
Turbo manifold $300
Plumbing $300
Total $4,000
10: Upright Design
Little is accomplished by drawing the perfect suspension and then finding that only custom uprights
fit the designunless that was the plan. While A-arms are fairly easy to fabricate, it is a lot more
work to create the uprights (especially if they require tight-tolerance bearing bores). Most of us
would rather use donor uprights, which means we form our design around existing parts.
I took a long time picking uprights. The Honda drivetrain came from a FWD car, so I considered
using the factory front uprights for simplicity, or that was the idea. Since 13 wheels would be used,
though, the choices were really limited.
Front uprights
Front upright definition is important (more so than rear) because you feel the car through the front
suspension. Camber, caster, scrub radius, and kingpin inclination all combine to determine the feel of
the car. It is desirable to end up with zero scrubto make the car easy to steer, especially since we
arent going to use any girly-man power steering. However, with the engine in the back and little
weight on the front tires, steering was expected to be pretty light. (Referencing the spreadsheet
showed roughly 300 lb. on each front tire, far less than the original Honda.)
Driving the car later showed that steering was very light. I am very happy with
it and have no desire or need for power steering.
On factory uprights, outboard lower ball joints are located all over the place, yet the cars all
handle pretty well. The point is, there are many ways of designing suspension, so dont spend a ton of
time worrying about where to put things exactly. Since the majority of the cornering force is fed into
the chassis through the lower ball joint, the force involved deserves some thought. Looking at it from
a mechanical point of view, it seemed that the lower ball joint should be as low as possible for best
mechanical advantage, which is what I did.
This caused some difficulty with my desired flat under-tray. With the lower
ball joint so low, the lower A-arms pass through the plane of the floor. In other
words, there could be no chassis tubes below the lower A-arms because they
would hit. If downforce-producing tunnels are desired, the lower suspension arm
should be as high as possible to stay above the ductwork.
The upper ball-joint attachment point has much less load applied to it. Its mounting point is
determined more by the desired wheel control than strength.
Kingpin inclination
Kingpin inclination (KPI) should be as close to zero as possibleimpossible in a practical sense,
so we use whats availableand production-car parts are typically about 612. We want little-to-
no kingpin offset (also called scrub radius) to make the car easy to steer, which depends on the
uprights, the wheels, and the kingpin inclination aboveI ended up with 1.5. Full custom front
uprights could improve this, but in a mid-engine configuration, there isnt much weight on the front
suspension. Said another way, the less weight thats on the front axle, the less important scrub radius
is, at least for a street-driven car.
BMW has/had an interesting front suspension design, where instead of a
traditional A-arm, they use two separate links. Instead of an A it looks more like
an H with the sides tipped in, but not touching. This has the intriguing feature of
having a virtual length longer than the actual links. The virtual intersection of the
two joints is the point that the suspension rotates around. This allows placing the
upright virtual pivot point further outboard than the physical points actually are,
in order to obtain zero kingpin offsetpretty cool. I didnt use it because I
wanted to keep things simple. It is mentioned here as an opportunity to think
outside the box.
Two very important components that drive upright (and wheel) selection are brake calipers and
rotors. Both have to fit inside the wheel and bolt to the upright. At the front, I had no desire to make
my own uprights, so I looked around the wrecking yards to check stock. It became apparent Id have
to do one of the following:
1. Use the uprights from the car supplying the drivetrain.
This has the appeal of simplicitythere they are, and they are guaranteed to fit. Unfortunately,
uprights from the factory are designed to work in a car weighing far more than what were building
meaning theyre needlessly heavy. Another way is to cut down the enormously tall S-shaped Honda
upright and weld on an adapter. I considered doing this, but my fabricator buddy, Alan Brickey,
expressed reservations about the unknown type of metal used in the upright. His concern was about
the quality of the resulting weld, and his concern became my concern, so I decided against it.
I would consider a bolt-on or brazed-on adapter bracket next time, allowing a
cut-down production part to be used. The trick is keeping the weight down. Then,
of course, we get to mess with the brake caliper mounting points. Doing it over
again I might do this, or just use Miata uprights. Or then again, fabricate my own
using bolt-on spindles.
2. Find factory uprights that are close enough.
The problem with using junkyard parts to build a car is the complete lack of data. Nowhere is there
a database listing all uprights on all cars, with weights and dimensions; I was stuck figuring it out
myself. Unfortunately, most of the yards (around here) that let you wander around only contain older
cars. Late-model wrecks are hidden away at boutique wrecking yards, where you tell them what
you want and they get it for you, for a price.
3. Modify an existing upright.
About 20 years ago, Id seen a Nissan 280Z race car where they had cut down the front
MacPherson struts and used them in a double A-arm arrangement; that idea was filed away.
4. Use an existing upright with an adapter.
Some struts bolt onto the upright, which can make it a good candidate for our needs. Replacing the
strut with a short bolt-on adapter allows placement of the upper pivot in a much better location.
5. Go full custom and make them from scratch.
Sometimes this is what it takes to get what you need.
I decided to modify existing MacPherson struts from a 1979 Nissan 280ZX. The Nissan parts were
plentiful and cheap, and brake upgrades were available. The first task was to measure the size of the
brake rotor and caliper to find the required space inside the wheel. After measuring the brake caliper
radius, determining how to minimize the kingpin offset, and choosing the bolt pattern, I could now
select the front wheels (covered in the following chapter).
The top of the strut was cut down so an upper pivot could be added to make it suitable for upper
and lower A-arm use. The strut tube was cut, leaving only about 3 remaining, necessary to miss the
inside of the wheel. The upper pivot is a tie-rod off a -ton Chevy truck, bolting into a tapered seat
that bolts into the top of the hollow tube.
The Nissan 280ZX upright worked well, with a handy hole at the bottom for locating a lower
extension. I bolted the strut to a wheel and measured how much room there was for it (dont forget the
space required for the rod-ends and bolts!). The extension moved the lower pivot as low as possible.
At the bottom, the extension is threaded for a 5/8 bolt, to bolt through a spherical bearing in the
lower A-arm, into the upright. A new steering arm was welded on later and the stock brakes and
rotors were retained.
Why 5/8? Thats way too big
The reason is bearing life. Look at the specs of spherical bearings and rod-
ends; it will show they are amazingly strong. Technically, I could have used
bearings, but they just dont last. Oversize parts were chosen for long life and to
give a warm fuzzy feeling that they would never break. I didnt want to be
replacing bearings; I wanted to be driving the car! All suspension pivots are
rod-ends, except for the outboard rear upright pivots, which are Honda Accord
ball joints.
Brake pad availability for the Nissan is a bit thin. If I were doing it over, Id
consider other uprights like the Miata. Of course, back in 1995, the Miata was
still rather new, with none in the wrecking yardsthey werent on the radar. In
hindsight I was a bit shortsighted, not realizing the Miata would soon be very
cheap and common. (The brakes were something of a pain, and there is a separate
chapter devoted to them.)
The lower pivot location is fixed; time now for the upper pivot. The reason the Nissan uprights
work well is the strut tube; it can be cut down and a bearing attached at any point. It was placed to
just miss the inside surface of the wheel.
One perk of the Nissan struts is that the wheel bolt pattern matches the Honda
bolt pattern on the Prelude hubs, which I had already decided to use.
Rear uprights
Satisfying the goal of independent rear suspension required the selection of rear uprights too (as
opposed to using a live-axle). It would be nice to use the uprights from the drivetrain donor since the
axles and wheel bearings would just work, saving a ton of work. Caster, scrub radius, and kingpin
inclination arent issues at the rear, which makes things easier. As at the front, we want the lower ball
joint as low as possible. Unfortunately, the Honda uprights just didnt work out; they were too big and
heavy, and the pickup points werent where I wanted. I ended up making my own from scratch (a very
big job), designing them to use stock Honda wheel bearings and ball joints. This ensured that the
Honda CV joints and axles would fit.
The rear uprights were a lot of work to make, both in design and fabrication. Unfortunately, there
arent too many alternatives if there are no production parts that fit the need. Even if theyre designed
in CAD, they still have to be fabricated, either by you or an expensive CNC machine.
Recently I read how a builder modified his stock hubs; he cut the upper and
lower sections off a production FWD hub, then put the entire assembly in the
lathe, turning it down to remove all metal except the large bearing cup. He then
welded his own extensions onto it. I felt pretty dumb for not thinking of it myself
but I have a good excuse. Alan Brickey said not to trust that the cast-metal factory
part could be safely welded, so I took his word for it and made my own.
It came down to making a bearing housing in the lathe, then welding on extensions out to where the
upper and lower ball joints would be, another tube for the steering link, and a plate for the brake
caliper attachment points; easy to say but a lot of work.
Shown here is the hub, brake disc, caliper, and wheel bearing cup. This mock-up ensured the
extensions off the bearing cup would clear everything. The cardboard method worked well; after
trimming to fit, it was transferred to steel, tacked in place, and rechecked. When attaching the brake
calipers, I fed air to the caliper to force it into its braking position, and then the brackets were tacked
in place. This was the only way to guarantee that the brake calipers were aligned with the brake disc.
This is how a lot of the car was built, using the actual parts as templates and building out from them,
which only works for a one-off car. It wasnt productive to put the parts into CAD and model the
whole mess just to find where the connecting brackets should go. I could figure that out very quickly
with the parts right in front of mewithout involving a computer. The upper and lower pivot points
were placed as high and low as possible while still missing the wheel. I welded square 1.25 x
0.125 wall square tubing to the bearing cup, and to those I welded tapered mounts for the Honda
Accord ball joints.
The bearing cup pretty much designed itself, meaning it had to fit the bearing and have a reasonable
wall thickness. Thankfully, I had the original Honda uprights to copy.
I learned a mechanical engineering rule of thumb for bearings. For an
interference fit, bearing cups should be 0.001 undersized per inch of bearing
diameter; a 3 bearing needs a 2.997 diameter bore.
I used the stock Honda brake discs (from the front of the donor car) to simplify attaching them to the
axle CVs. The stock Prelude discs were 11 in diameter, so they were cut down to 10 on the lathe.
This raised the question of what to use for rear calipers. Because the discs were now smaller, the
stock Honda calipers wouldnt fitnot a bad thing, since the Honda calipers weigh a massive 13 lb.
each. Wandering around the wrecking yard, I found that 1986 Mazda RX-7 front brake calipers were
perfect; they fit on 10 brake discs of the same thickness as the Prelude, and at 8 lb., they were much
lighterproblem solved.
I had to grind the rear Mazda calipers slightly to make them fit; there was not
even room for stick-on wheel weights. I cut things a bit close; when even small
pebbles get between the caliper and wheel, they leave gouges and make a nasty
noise.
Since we are discussing the rear suspension, on the topic of axles, I was afraid
to do anything too non-standard. It never occurred to me to use anything but
Honda parts for the brake rotor, spindle, bearing, and CV joints. After it was all
done, I realized I hadnt been very open-minded. What woke me up was when I
had the axles custom made.
I had to specify the splines at each end, plus the diameter and length, and
realized that the axle could serve as the bridge between dissimilar part
brands. For example, using a Honda drivetrain with Mazda Miata uprights would
have been easyhad I thought of it. It might have meant not having to make my
own uprights at all. Then again, I didnt know at the time if Miata rear uprights
would work (there were no Miatas in the local wrecking yards in 1995). The
point is, avoid thinking that all parts have to be from only one manufacturer.
Often there are many more solutions than you might think.
I did something I hardly ever see, attaching the shocks directly to the rear uprights instead of the
lower A-arm. This removes all bending loads from the lower A-arm and tension from the lower ball
joint.
Lesson learned
In hindsight, Id be less picky. While the suspension parts I picked worked
very well, I did make a lot of work for myself! Just how close to perfect must it
be? Getting 80% of perfect is easy, but that last 20% gets very expensive, in both
time and money.
All the work centered on the fabrication of the bearing cup. Avoiding that step
would have sped things up considerably, since the rest of the upright is just steel
plate and some tubing. Since the car was completed, Ive seen two fabricated
rear upright designs that I feel are much simpler than mine. The secret is to use a
bolt-on spindle. Above-right is a fabricated spindle using a 240SX bolt-on rear
spindle (image credit John James). At right, and below, is a rear-upright using the
Chevy Cavalier front bolt-on spindle. (image credit Alan Bertwistle). Another
possible choice is a rear-axle bolt-on spindle from a late-model Honda Civic.
Im very happy with how my rear uprights turned out, but next time Id look for
a way to avoid fabricating the bearing cupa precision component that took
many, many hours of labor. A bolt-on spindle works well, and it doesnt matter if
the drivetrain and uprights are of different brands.
11: Wheels
Wheels have far-reaching implications on brake caliper radius, fender clearance, turn radius, scrub
offset, ride height, bolt pattern, and on and on. This point in the project was perhaps the most difficult
because the many different factors converged into a tricky puzzle. The uprights, brake calipers, Mini
shell, desired suspension type, and tire data all have to play nice together. It is very much a concert,
working with all the related parts, looking for the best balance. The trickiest part was finding wheels
that had enough room for the brake calipers; the 13 wheels made it real tight. It would be a waste of
time to design everything perfectly and expect to order a wheel with arbitrary specifications. Instead,
it is a case of having to use what is available and making it work.
Bolt pattern, offset, inside clearance, width, weight, cost, and looks are the necessary decisions.
Bolt pattern
Since the Honda hubs would be used at the back, I wanted to match the bolt pattern at the front, and
fortunately the Nissan front uprights had the same bolt pattern.
Lesson learned
Since the front and rear wheels are different sizes, a rear wheel would never
go on the front, so why did I make the bolt pattern the same? In hindsight, it was
for no good reason at all. I forgot that the reason its normally done is so a spare
tire fits both the front and rear hubs, but since I dont have a spare, it isnt an
issue. Everyone else does it that way, and I wasnt thinking. There was no real
reason to do so other than aesthetics and the desire for uniformity.
Offset
At the front, we want as little scrub offset as possible to make the car easy to steer. The Nissan
uprights needed a lot of wheel offset to minimize the scrub radius. This made finding a wheel just that
much more difficult.
Inside clearance
This was the toughest specification to meet. I had made it more difficult since Id chosen 13 tires,
but so be it. With 10 brake discs, it left very little room for the brake calipers.
Width
For maximum cornering it is important to use the widest wheels possible. The wheel provides
maximum support in a turn, when the tire is trying to roll off the rim.
The front wheels are so wide that I have to be very careful of curbs; the edge
of the wheel will hit the curb before the tire does. When laying the tire down on
its side, the tire doesnt touch the floor.
Weight
Buy as light a wheel as you can afford. This weight is doubly important; not only is it unsprung
weight, but also a rotating mass. Of course, this must be balanced against the perils of real-world
streets, with potholes, curbs, driveways, and speed bumps. Some really light wheels are good only
for smooth racetracks. On the other hand, cutting things too fine in the quest for lower weight can
cause wheel failure at the track, tooduring an unexpected off.
Cost and looks
This is entirely up to you! At the front, the constraints were:
The wheel offset and uprights set scrub, which we want to minimize.
The wheel had to fit the Nissan upright.
The wheel had to clear the brake calipers, which was the toughest goal to meet.
At the rear there were fewer restrictions, though brake clearance was still an issue. One problem
during the wheel search was finding technical information about the wheels; many manufacturers
didnt provide inside-wheel clearance figures. I needed this to determine if the wheel would hit the
calipers.
Finally I located a wheel; these Taylor Racing wheels (at bottom-right) met all of my
specifications. Because they are now out of business, if a wheel is ever damaged, Ill probably have
to buy a set of four new ones. The chosen sizes were 7 x 13 front and 8 x 13 rear. These were as
wide as I could go for the chosen tire widths.
Doing it over again Id use one of these two below, Panasport or Revolution.
As a fallback position, fully custom wheels are always availablefor a price.
Revolutions corporate situation is unclear at the moment, so if you consider their
wheels, confirm theyre on firm ground first.
Regardless of the wheel, make sure to get matching lug nuts. I didnt realize lug
nuts have different tapers: either 60 or 90. This cost me $20 buying lug nuts
with the wrong taper. The one on the left has a 60 taper, and on the right is 90.
You can see how the one on the right wasnt seating evenly. The good news is
that the correct part on the left is much lighter. Yes, smaller, lighter lug nuts do
help reduce rotational weight and lighten the car a tiny bitbarely measurable,
maybe, but every bit helps. Its true: having small nuts is a good thing.
12: Roll centers
With the wheels, tires, and uprights defined, next is determining where the uprights go, relative to
the chassis. In the case of the Mini, it meant carefully measuring the shell, especially around the wheel
wells, to determine the appropriate position. To be sure, the shell was positioned over the wheels to
make sure they didnt look goofy and werent tucked too far in or sticking out too far. These numbers
were carefully recorded because the entire suspension is designed inward from these points. No
matter what type of suspension, uprights, or chassis, the wheels have to be in just the right position.
First, the upright was bolted to the wheel and the assembly placed on the floor. The height of the
pivot points was carefully measured and recorded (several times). The wheel was then propped up
vertically and the same points recorded again. These numbers, along with knowing where the tires
were positioned in the wheel wells, produced the outboard suspension pivot points. Armed with
these, we are now ready to choose the roll centers, but first we need a tool to find and test them. This
can be done with graph paper, pencil, and ruler, CAD, or the String Computer.
Suspension design tools
There are several ways to design the suspension. In order of increasing expense:
1. Guess.
2. Make many drawings using graph paper.
3. Use Staniforths String Computer or Smiths paper doll model.
4. Use suspension design software.
I dont recommend option 1!
It was pointed out that, while free, option 1 could be the most expensive
expensive in terms of time and money, the consequences of making very
poor decisions and having to redo it all!
I used option 2 and consumed a lot of paper, but it was simple, cheap, and a good way to double-
check numbers found using CAD. So whats a String Computer? Staniforth
3
came up with the name;
its a two-dimensional working model of a proposed suspension using movable wooden links on a
hardboard base. Strings are used to project where the links cross in order to find the roll center; when
the suspension moves up and down, the strings show what the roll center is doing. (Smiths paper-
doll method is identical.)
If you decide on the String Computer method, heres a tip to simplify its use:
Instead of using strings, buy two small laser pointers and attach one to each A-
arm. Now, when moving the suspension to new positions, just see where the laser
beams cross; itll make it easier and faster to find the roll center.
Option 4, suspension design software, is faster and far more accurate, yet it is impossible to make
use of the accuracy. The chassis will warp due to welding heat, and it is hard to get closer than
perhaps 1/8 accuracy. Thats why the suspension is adjustable!
Consider that a typical torsional stiffness of a Super-7 chassis is maybe 3000
lb./deg. A deflection of one degree is one inch, over a distance of 57. For
convenience, lets say that our 7-chassis has 28.5 between suspension mounts
(half of 57). It means that if 3000 ft. lb. of torque is applied to the front
suspension mounts, the chassis will twist (rotate) relative to the rear suspension
mounts by two degrees. But now, consider applying only 300 lb. Even this
relatively low force will still displace the suspension pickup point by 0.20!
This is the very reason why its so important to design a stiff chassis. Otherwise,
having suspension pivot points calculated to three places past the decimal point
is totally pointless.
It is nearly impossible to keep the suspension mounts exactly where wanted
during fabrication, regardless of how many digits of precision the CAD program
specifies. If it says to put a suspension point at 27.576, that precision is useless
the chassis moves around, both due to welding distortion and torsional loading
during cornering. All we can do is try to make the chassis as stiff as possible, but
dont think that because we used CAD, were guaranteed to have a superior
chassisit simply isnt true.
On my budget, suspension design software was too expensive, around $350 for the Mitchell
WinGeo Suspension Design software. Other programs include SusProg3D by
www.bevenyoung.com and Suspension Analyzer by www.performancetrends.com. I found a free
suspension design program, too, but have not used it, so I cannot vouch for its accuracy or stability.
Since its free, though, it could be worth checking out:
http://www.locostusa.com/yahoo/Wishbone_setup.exe.
If youre making just one car, I recommend using the String Computer or the
cardboard models Smith suggests. If youre going to make several cars, using real
fixtures and having parts made with CNC, use CAD. I borrowed a buddys copy
of the Mitchell softwarean early, very buggy, nearly unusable DOS version. I
also used the String Computer for validationit would have been fine to use it
on its own.
In the case of the perfect suspension, there is no scrub, no roll-center movement, and perfect
camber compensation over full wheel travel. The reality is that while we can achieve one of these
goals, it comes at the price of making the other two worse. Choosing suspension pickup points can
drive a person batty, and I took it too seriously. There are many different suspensions that work pretty
well, so a design doesnt have to be perfect.
Remember that saying? Fast, cheap, celiable pick two.
In the case of suspension design its more like: Tires upright in turns, tires
upright in bump, no scrubpick one. Designing suspension to excel at one
comes at the price of the others, so every designer compromises, picking a point
somewhere in between the three.
Balsa FEA
Although the suspension design tool has been chosen, it still doesnt answer the question of where
to put the tubeswhere to place them in order to create a stiff chassis. Enter the Balsa FEA model.
FEA is an acronym for Finite Element Analysis. Its a process, almost always done by computer,
that calculates the stiffness of a three-dimensional structure, such as a tube chassis.
In general, FEA software is expensive (but getting cheaper every day). One way to estimate the
stiffness of a proposed chassis is to instead build a physical model. Building the model isnt a good
way to estimate chassis stiffness itself, but it is very helpful for finding out how to make a proposed
chassis stiffer.
Basically, a balsa wood model of the proposed chassis is made; 1/8 scale is convenient, since 1
tubing can use 1/8 square balsa strips, a common size.
First, all major components are measured and replicated in balsa: the engine, transmission, driver,
and gas tank; suspension parts are not needed at this point. Next, the proposed chassis is constructed
around these non-negotiable components. With the basic chassis completed, this is where the fun
starts. Hold the chassis in two hands, the front in your left hand, and the back in your right. Now,
gently twist the chassisnot like a dish towel, but just a slight twist. This is what will happen in real
life to the actual chassis when it hits a bump with the front tire.
The trick is to watch how the balsa model twists to see what areas may need an additional tube
added, or perhaps whether a sheet should be applied to an open bay, to improve the chassis. A
slightly better qualitative method is to anchor the rear of the chassis to the table, to free up a hand.
Then, twist the chassis as before, with the left hand, but have calipers in your right hand so you can
measure the changes in the various bays. This, perhaps better than any other method, graphically
shows whats moving, and in which direction. Adding diagonals exactly where they do the most good
is the goal.
There is an excellent, in-depth, step-by-step article on how to strengthen a
chassis. The author ran an Finite Element Analysis (FEA) on the Locost chassis
and shows the benefits and costs of each added tube. He ended up more than
doubling the torsional strength with no gain in weight:
http://www.locost7.info/files/chassis/kitcaranalysis_V2.doc.
Regardless of the type of car you are designing, the same techniques can
greatly improve a potential design.
It may even be possible to eliminate tubes that arent doing their job or add full panels in select
bays to greatly increase torsional strength.
Using balsa sheets and strips, available from your local hobby store, coupled with a hot-glue gun,
you can very quickly determine the potential merit of a chassis. In fact, I consider this method better
than FEA (at least for a one-off) because its more tangible. Also, in one evening, you can have good
results and be on your way. Using the full-blown FEA will take far longer.
FEA may be more precise, but the precision isnt that useful in our situation. Far more helpful is
being able to quickly find potential improvements that can be made with little or no weight gain.
While it may be old-school, its effective, fast, and inexpensive.
The roll center is the point about which the car rotates when it leans. In the figure above, the roll
center is at the centerline of the car because the left and right suspensions are identical. In a turn,
though, the suspension on one side compresses while the other droops, and often the roll center
doesnt stay at its static position, moving up, down, or sideways. It is best when the suspension acts in
a predictable manner, instead of having the car feel odd in a turn. When the roll center moves, the
way the car leans will feel different. Subconsciously, the driver will counteract, but it is one more
thing that he has to contend withand shouldnt have to.
It is important to design the suspension so the RC does not move relative
to the CG. When the car hits a bump, the car will bounce over it, with the CG
moving with the chassis. The goal is to have the RC move with the CG. Only this
way does the cornering force acting on the CG have a consistent, predictable
force and a consistent wrench length (on the next page). If the RC stays fixed
relative to the ground, it means that the wrench changes length as the chassis
moves up and downnot whats needed for predictable handling.
Of course, nothing overrules keeping the tires flat on the road. If it comes down to a choice
between keeping the RC motionless and keeping tires upright, the tires must win. Contradictory? Yes,
but so is suspension design!
There is a common issue that builders unwittingly face. Imagine lifting the entire suspension off a
Miata and putting it on a Seven. Everything is moved over: the uprights, suspension arms, wheels,
tireseverything. Also, the inboard suspension pivot points are copied exactly. But because the
Seven is narrower, the left and right suspension assemblies are closer together than they were on the
Miata. Guess what? The RC is no longer where it was in the Miata! When the left and right
suspensions are moved toward each other, the roll center gets lower by the same proportion. That is,
if the Seven is 25% narrower, then the Miatas roll center will be 25% lower. This isnt a big deal, if
anticipated and understood.
Cornering force makes the car lean, and that force is applied at the CG, rotating the car around the
RC. If the RC is high, the distance between the RC and CG is small, resulting in a short wrench.
Since it is short, just like when using a short wrench to loosen a bolt, its hard to achieve much
torque. Its the same here; the cornering force trying to rotate the short wrench has a hard time, so the
car doesnt lean much. It seems to be a good thing, but in a turn, the car is also trying to pole-vault the
RC up and over the outside tire. The higher the RC, the more upward force there is (see Smith
4
). It
also makes tuning the chassis difficult; since the chassis doesnt roll, how can handling be tuned? If
the car doesnt lean, stiffening or loosening the anti-roll bar or springs will have no effect!
If the RC is low, it results in a long wrench, so the car leans more, a bad thingbut also a good
thing, since the suspension is now easy to tune. Its enough to drive people nuts. The RC can be
located anywhere, but typically is above ground level but no higher than the centerline of the tire (a
live-axle, for example).
In addition to placing the RC where desired, it is important that it does not move. Staniforth goes
into great detail as to why he feels it is so important, and I agree with him (unless it comes at the
expense of keeping the tires square to the ground). If the RC moves around, the car will handle oddly,
and the driver does not need yet another variable making his job more difficult.
Note that having a fixed RC, relative to the CG, is impossible with a live-axle
design located by a Panhard rod or Watts linkage. Here, the wrench changes
length as the car accelerates, decelerates, or corners, causing the car to feel odd,
due to the roll-couple changing. A car that feels most stable, consistent, and
predictable is one that has a constant distance between the RC and CG, at the
front and rear of the car.
An illustration by Smith
5
shows a diagram of a car sliced like a loaf of bread, with the CG of each
slice (centroid) interconnected. The centroid can be approximated with a straight line in the form of y
= mx + b, called the centroid axis. Using the loaf-of-bread analogy, pushing a stick through it along
the centroid axis and turning it will show how it will lean.
Look at the above figure, and picture what would happen if the front and rear
RC moved differently with suspension travel. This causes the distance between
the RC and the centroid axis to change, altering how the car feels. Trying to get
predictable handling from a car set up this way is very difficult.
Also note that the weight of the centroid doesnt matter; its all about where the
CG is that counts.
There is an RC at the front suspension and another at the rear. Each can be located wherever the
designer wishes (with IRS), and my research indicates that the RC axis should be parallel to the
centroid axis.
Now, more than one person has questioned this idea, pointing out that the chassis is (theoretically)
a rigid structurethere is no centroid axis, but only a single point at the chassiss CG through which
all cornering force is applied. There is a rule of thumb that goes: make the RC higher at the end of the
car that has a higher CG (in my case, at the back, where the engine is); conversely, set the RC lower
at the end of the car with the lower CG (like at the front with nothing but suspension, radiator, and
maybe fuel). The funny thing is that this agrees with Carroll Smiths centroid drawing.
A former Lotus chassis designer told me that it also depends how the car is
used. At the time, I said it would be for mostly track events, and he said the RC
height wouldnt really matter; the car would be so stiffly sprung that the RC
couldnt move around much. In hindsight, I misspoke; the car would be driven
quite a bit on the street, toowith softer suspensionso RC placement was
indeed an issue. This is what happens when trying to design a car for both road
and track compromise.
I chose 2 front and 4 rear, which, per the spreadsheet, matched the slope of
my cars centroid distribution. Did I make the right choice? I believe so; the car
is very balanced, enough so that I dont use anti-roll bars.
The case of the live-axle
For independent suspension, its easy to put the RC wherever we want, but its not as easy for the
live-axle. Since builders continue to use the live-axle, it is discussed herefor completeness. The live-
axle must be constrained laterally, and there are four usual methods of doing so: the Panhard rod,
Watts linkage, WOB link, and Mumford link. All serve the same purpose, yet each has its own quirks
and benefits.
Panhard rod
A Panhard rod is the simplest way to locate a live-axle. It should be made as long as possible and
sit level at normal ride height. The problem is that the RC is too high, made worse because it moves
up and down with the axle. Since the rod moves in an arc, as the axle rises and falls, it also moves
sideways. This is the least desirable method of getting the job done.
Watts linkage
Since the Watts linkage is symmetrical, the axle does not move sideways like the Panhard rod.
Unfortunately thats the only benefit over the Panhard rod; the RC is still too high and moves with the
axle. It also requires a low chassis mount, which may or may not be available.
WOB link
While more complicated than the above links, the WOB link allows moving the RC to where the
designer wants, and the RC moves with the chassisa good thing. Like the Watts linkage, there may
not be a chassis tube where the designer wants the RC to be (usually about 23 above ground).
There are sizable forces fed through the chassis mounting points, so plan accordingly.
Mumford link
The Mumford link is about as good as it gets for the live-axle. Its big advantage is that the RC is
located along projections of the long links, rather than requiring a physical point to be at the desired
point. With a desired RC of 23, thats a big deal. Another plus is that the RC moves with the
chassis. I recommend this mounting method. I discussed with my brother the option of swapping his
Panhard rod for the Mumford, but he didnt have a chassis tube passing behind the rear axle. Time
will tell how his Stalker handles using a Panhard rod. Other owners seem perfectly happy with their
cars!
My brother asked, Okay, just how much better will handling be if I go to all
the effort to change it? I dont know, and Im not sure it can be quantified. The
only way to know is to have two identical cars, one with a Panhard rod, and the
other with the Mumford link. Of course, as noted before, one driver may feel
theres a night-and-day difference between the two, and another may feel that they
both act about the same.
My feeling is, since the Mumford allows setting the RC where we want, which
gives more predictable handling, it is what I would have used if I had chosen a
live-axle.
Whatever. So where do I put my roll centers?
This is extremely subjective; there is no one right answer. We have seen how it depends on how
the weight is distributed in the car. In general, I resist issuing magic numbers because there will
always be someone saying they are wrong. At the same time, many builders are simply looking for
help getting an idea of decent valuesokay.
For a Super-7 type car, place the front and rear RC about 3 above ground. This rules out using the
Panhard or Watts link, but thats what it takes to do it right. These values can be fussed over, but they
are at least in the ballpark.
I placed my front RC at 2 and the rear at 4. The rear RC is higher because of the CG of the
relatively tall engine. In the Super-7 example above, the engine is at the front, but the gas tank at the
rear is also rather high up (at roughly the same height as the engine), so the front and rear RC should
be at the same height, too.
Consider the case of designing a car for a given tire size, then later changing
tire diameter (I may face this someday). Say that the new tires are 2 larger in
diameter. It is true that, using the shocks adjustable spring seats, ride height can
be set back to the same ground clearance as before. So changing the tires didnt
matter, right? Wrong. Adjusting the spring perches changes all sorts of things.
The roll centers are now in a different place, the tires are at a different point in
their camber curves, and the caster and toe settings have changed. The car will
handle differently maybe better, maybe worse. This shows just how
interrelated everything is.
After my insistence on keeping roll centers stationary, I found that there are
some very successful cars for which RC locations werent a design concern.
These car designers were more concerned about keeping the tires upright in turns,
and they simply didnt worry about the RCs. (As said before, though, if these cars
have very stiff suspension, the RCs cant move around much.) It just goes to show
that there are many ways to design a car. Since I designed mine to keep the RCs
stationary, and the car feels very stable and predictable, it is what I recommend.
But Im well aware that there are many different approaches to defining the
elusive perfect suspension!
No matter how its designed, what is important is keeping the tires upright to
have even tire wear under as many conditions as possible. Only if the tire
surfaces are kept on the ground can traction be maximized.
13: Suspension Design
Here is a list of goals to help determine where to place the all-important suspension pickup points:
Upper links should incline upward toward the wheel and be about 2/3 the length of the lower
link.
The longer lower link should slope slightly downward toward the wheel.
Parallel, equal-length links results in camber change equal to body rollnot good unless the car
is so stiffly sprung that it doesnt roll much, or is only a cruiser.
Scrub, the sideways tire movement as the suspension moves vertically, is a bad thing.
Long suspension arms are a very good thing. They decrease scrub and help maintain the desired
camber curve over a wider range of suspension travel.
Large amounts of static camber limit braking and acceleration because the tire is not flat on the
road. As always, this may or may not be an issue, depending upon the application.
A wheel-to-shock ratio of 1:1 is desirable. When the wheel moves a given amount we want the
shock to move a similar amount. Shocks perform best when theyre being worked; the shock has
to move to do its job! If the wheel moves an inch, and the shock only moves 1/16, its difficult
for the shocks internal valving to deal with the tiny movement. Also, at such small
displacements, internal stiction may cause the valves to not operate at all, rendering the shock
essentially useless.
Shocks should be centered at normal ride height. This means that with all fluids and the driver in
the car, the shocks should be at about 40%50% compression. This is yet another reason to use a
spreadsheet, which can calculate how much the shocks will be compressed for a given mounting
orientation and load, without having to build the car first.
Roll centers should move with the CG and remain a constant distance from it. This keeps the
handling feeling consistent, since the rolling force pulling on the CG has a consistent lever-arm
length. The effect is that the car is easier to control because it acts in an intuitive, natural,
predictable manner.
Anti-squat and anti-dive can be problematic; Smith said it binds the suspension, and I believe
him.
I used front rocker-arm suspension to move the shocks inboard so they become sprung weight (a
good thing). The disadvantage of this is that rocker-arm suspension takes up more space than
traditional outboard-mounted shocks.
Overall geometry
We would like zero camber change during acceleration and braking, perfect camber compensation
when the car rolls in a turn, and no scrub. It just isnt going to happen, though; any suspension used
today is a compromise. Based upon the primary use of the car, we go from there. For a cruiser, it can
be almost anything. For a drag-race car, we need the wheels to stay upright to the road under
acceleration and braking. For a road racer, we want the tires staying upright in the corners. Designing
suspension that excels at one trait, though, will cause it to lose out to the others. The saving grace, if
there is one, is that every other designer is in the same boat. Everyone has the same tools available;
its up to us to decide whats most important for our specific application.
Caster
Caster provides the self-return-to-center action of the steering wheel (think of how a caster-wheel
chair moves on its wheels). The more weight the front wheels carry, the less caster is needed. For a
mid-engine car with a light front end, this means more caster is better. How much caster to use is very
subjective and argued about all the time; I used 5, but 78 might have been better (for me). The
amount depends on how much you want; like many other parts of design, dont let yourself be talked
into using a particular value. My advice is set it to the minimum you think you want but leave room for
adjustment so that you can dial in more than think youll ever need; that way youre covered either
way. Caster is increased by moving the top upright pivot toward the back of the car and/or moving the
bottom pivot toward the front. How this is done depends upon the suspension type, but keep the
needed adjustment in mind during design.
A common trap for builders is to use an existing frame design and change the
suspension. A typical example is an American builder using Ron Champions
book plans to build a chassis. The problem is that the builder does not have the
front uprights Champion specified, instead using domestic parts from a Miata, S-
10, or Pinto. The thing is, it wont work as planned.
Changing the uprights changes the entire suspension design. If the upright is
changed, the resulting new suspension must be checked, almost certainly
requiring a redesign. This is why, even when using an existing chassis design,
you still need to know suspension design. Unless you can source the exact
uprights that the chassis was designed for, the inboard pickup points and steering
rack mounts will have to move.
Some builders believe that using the entire suspension off a car guarantees
success. The problems start if these parts are mounted on a car that has a different
width than the donor car. Using all the original components does not guarantee
anything, because the suspension pickup points are now closer together (side-to-
side), lowering the RC. The biggest problem isnt that the RC moved, but that the
builder doesnt know that it moved.
Also, be aware that substituting ball joints that are a different size than the
ones they replace moves the suspension pickup points and alters suspension
behavior! A similar problem comes from using commercially-available A-arms.
Unless they are designed specifically for your car, or you are reverse-engineering
your suspension to use them, they change the suspension geometry.
If you use different uprights from what the plans specify, leave out all tubes
directly involved in mounting the suspension arms and steering rack. Not only
will the brackets have to move up or down, or side-to-side, theres a good
chance theyll have to move inboard or outboard as well. Once again, it shows
how everything is connected to everything else.
Rising/falling rate suspension
The goal is to have constant to slight rising-rate suspensionnot falling-rate. This gives consistent,
predictable handling. The way to ensure this is to examine the suspension geometry. As the
suspension compresses, the angle between the shock and actuating arm must move toward, but not
past, 90. Geometrically, this means the ratio of the actuating distance and suspension travel stays
slightly above 1:1. Falling-rate means that from wherever we start, the ratio decreases, so that the
suspension has to move a disproportionate additional amount to reach equilibrium. This will give a
subconscious sense that the car is trying to tip over or sinknot a good characteristic for instilling
confidence!
In a truly worst-case scenario, falling-rate suspension can go so far over center
that its spring rate drops enough that it cant return! In other words, the
suspension can self-compress so far that it cant get back to its neutral position!
Knowing the dimensions of the wheels, tires, track width, ride height, and roll centers, we can now
choose the inboard pickup points.
I spent weeks searching for the elusive (and impossible) magic combination of keeping the RC
fixed and the wheels upright in bump and droop and having no scrub. Like every designer whos come
before, I learned the hard way that I couldnt achieve all three goals at the same time. Like everyone
else, I ended up with a solution that was somewhere in between, in order to get on with things. I
leaned more toward keeping the RC stationary over the other two goals, but hey, you have to pick
something and get on with it.
Rear suspension
At the rear, it became apparent that the suspension arms werent going to be as straightforward as
at the front; the engine was in the way! The problem is related to using a Mini shellits a small car!
I could either change engines or change my thinking, so I did the latter. The rear uprights were already
fixed in position, dictated both by the required alignment with the transmission output ports and the
requirement of the wheels having to stay inboard of the fenders. Yet another part of the car had just
designed itself, and it wasnt even designed yet.
The engine was placed on the chassis table, positioned at the previously chosen ground clearance
of 4.75, under the pan. The rear upright mock-ups were placed in line with the transmission and in
line with the wheel cut-outs in the shell. With the main roll-hoop also mocked up, the rear suspension
virtually designed itself; there was really only one place for the links to go.
While the rear suspension is technically A-arm, the arms had to be rotated to miss the drivetrain;
the forward links became almost trailing links. For the left-side wheel, the lower forward link just
missed the alternator belt when in full compression, and the upper link had to miss a chassis tube. The
forward links attach to the main bulkhead cross-tube. For the right-side wheel, the transmission
housing was in the way. Since the left and right side suspension links have to be the same, the
limitations of both dictate the final layout. The only real choice was where to attach the forward tubes
vertically. The suspension model and CAD (or String Computer) would determine the attachment
points.
This rear suspension layout was necessary due to the small size of the Mini. On a slightly larger
chassis it might have been possible to have real A-arms coming in below the drivetrain, but I
simply didnt have room. For the rear-most lateral links, there was more freedom since they came in
behind the transmission, a completely open area. Using the guideline that longer links are better, they
practically met at the center of the car. That left the upper rear lateral links to make the geometry
work.
Closely related to rear axle design is half-shaft lengthit must be correct. If
theyre too short, when the suspension goes to full droop, the CV joint can pull
out of its cup and disassemble itself. If theyre too long, when the suspension
goes to full bump, they can bottom out in the CV cup, bending or cracking
something. I spent a lot of time making sure they were right, and it wasnt easy
(including making the first set the wrong length). Also, dont assume, like I did,
that the left and right factory axles are exactly the same length; for some reason
the Honda axles differed by 0.4!
Finding the right length is especially a problem if the stock axles are replaced
with shorter ones, which they were in my case. It makes the angular motion much
more extreme with suspension travel, increasing axle thrust.
I do not recommend cutting and welding axles! I was talking to some guys who
worked on the NPTI IMSA car. They said with the car on the dyno, they had put a
strobe light on the engines half-shafts and were shocked to see them twisting 90
under the engines output torque! Think about whether welded axles can stand up
to that before making a decision. Yes, that was on an 800 hp car, but still. I had a
set made by Summers Brothers in Ontario, CA, near Los Angeles.
Rear suspension was designed so the wheels toe in slightly under both bump
and droop. The idea was that, under acceleration, I wanted toe-in, to avoid the
back end moving around in an unsettling manner. Under braking, the rear wheels
move into droop and I wanted toe-in again to keep the back end predictable. This
turned out not to be such a great idea and is covered in detail later. Suffice it to
saykeep things simple.
Selection of inboard suspension points
The actual choosing of the inboard suspension points is next. We already know the outboard pivot
points. We know the target roll centers, steering range, keep-out zones, desired shock ratios, and
steering rack size. Its time to work with all the known constraints and connect the dots.
This is where you use graph paper, Staniforths String Computer, or specialized CAD, and have at
it. I spent about a month looking for those elusive points in space to make my pile of steel and carbon
handle like a race cartesting, measuring, testing again, double-checking, setting aside potential
candidates, moving pickup points just a bit to see if things got better. If not, Id go the other way
around and around.
The goals are listed at the top of this chapter; if there is one Id call most important, it would be to
have the roll centers not move relative to the CG. (This means as the chassis moves up and down, the
roll centers move up and down with it, not remaining motionless relative to the ground.)
Dont feel bad if you dont attain suspension nirvanano one has!
Rocker-arm suspension
It can be a little overwhelming when first designing rocker-arm suspensionit was for me. I found
it best to break the task into smaller pieces, for easier chewing.
In the figure above, the key is realizing the upper arm is both the A-arm and the rocker arm. The
trick is to concentrate on the outboard portion first. Since we have already selected the inboard
suspension pickup points B and D, we can next determine the shock mounts.
The bottom shock mount E is (in this configuration) typically along a bottom chassis tube. In this
example, pivots E and D are separatebut they dont have to be. In my car, I placed the lower shock
mount, E, and the inboard A-arm pivot, D, on the same bolt. It just depends what you are trying to
achieve.
This happened because once I had achieved the shock motion I wanted, I saw
that points E and D were right next to each other. Instead of having two bolts so
close that they were practically touching, I realized they could be combined onto
one common bolt. Point A was then moved slightly to compensate for the move
from point E to D. The end result was one less bolt and a simpler layout, while
still achieving 1:1 spring travel, with very slight rising-rate suspension.
With point E fixed in place, its now easy to choose the last point, A. One of the design goals was
to have 1:1 spring movement with the wheel, so point A is moved left or right until its movement
equals the tires movement. This is easy, since all we need to do is make AB = BW. We can also
move A up and down if needed; nobody says the upper link needs to be straight. Tipping link AB
upward or downward is okay as long as BC doesnt change. This is usually necessary, too, because
the shock may be fairly tall; we dont want the lower shock mount, E, hanging out the bottom of the
car! In fact, my upper rocker arms did end up being V-shaped for this very reason, with the bottom of
the V being point B.
Why is the shock tipped at an angle? Recall that we want rising-rate suspension. Placing it at this
angle means that, as the suspension compresses, the angle between link CA and the shock body moves
toward 90 but not past it.
Thats all there is to it. If a rocker ratio of something other than 1:1 is desired, the ratio is simply:
shock ratio = AB / BW
This simple rule assumes the rocker-arm is nearly 90 to the wheel, which it
usually is. As yet another indication of the interconnectiveness of everything,
note how everything changes for the worse if wheels with a different offset are
substituted!
Heres a pretty wild suspension setup, yet as odd as it looks, it works very well. When the lower
A-arm moves upward, say, one inch, the rigid link pushes the right side of the rocker arm up one inch.
That moves the left side of the rocker arm down one inch. But the bottom of the shock also moves up
one inch, so what this does is mechanically amplify A-arm movement. With the pivot on the upper
rocker arm in the center, shock travel is doubled for a given A-arm movement. Its a great way to get
good shock travel where space is tight! I dont know if he invented it, but my buddy Lee Kaiser said
he came up with this in the 1970s.
Final notes
If this chapter seems evasive, in that Im resisting handing over magic formulas for the perfect
suspension design, its because there arent any. Thats right; there is no perfect suspension. Every car
is an example of how the engineer got all the conflicting issues working fairly well with each other
but never perfectly. A very big part of the design is what the driver wants, to get his (or her) best feel
for the best handling car. Why? Because each person has his own subjective preference for what he
thinks is best. All the above information is to be taken as in general; if you find your springs are
too soft, fine, swap in stiffer ones. If its a race car youre building, and adding more (or less)
Ackerman gives a faster lap time, then thats the correct amount for your car.
In one of Staniforths books, he mentions a fellow who got the chance to drive
Sennas F1 carand found it nearly undrivable! Theres a strong lesson there,
how one drivers setup is completely wrong for another. Lee related a story of
a father and son co-driving the same car at the autocross. Lee got a chance to ride
along with each of them and observed:
The son was so smooth it didnt seem like he was going that fast. He never
jerked the car around or spun the tires accelerating or braking; it was like he was
cruising along at 30 mph only he was going twice that. He handled the car so
smoothly youd think he was warming up for the next lap, until you heard his lap
time usually top time of day.
The father was the exact opposite. When you got settled in the passenger seat
you learned to lean away from him and hang on. Thats because his arms were
constantly flailing all around, the car never pointing straight ahead. Much of the
course seemed to come at you through the side windows, because he was never
pointed in a straight line, the car hurling this way and that, never aiming at the
next turn, yet arriving there anyway, usually sideways.
The most interesting thing about these two wildly different drivers was that
they got virtually the same lap times.
My buddy Ron Schramm brought up a good counterpoint, though: while the two
cars may have the same times, which car would be more reliable, and which
would have longer tire life?
So while a designer struggles to produce the perfect suspension, all he can do is use the sum of his
experience and whats been developed before, mixed with his drivers preferences. Whats perfect
for one driver is a problem for another, and whats poor handling to one is world-class performance
to another. All you can aim for is getting as close as possible to what the driver wants and balancing
that against what actually makes the car faster (and they arent always the same!). So there you have
it, the X-factor in suspension design: the driver.
Summation
Weve come a long way, having just completed the nastiest part of the design. The inboard
suspension mounting points chosen now serve to drive the chassis design, an adventure all its own.
(Image credit Robert Bowen.)
14: Springs
Springs choose themselves. Once constrained by all the requirements, one magic number pops out
of the equations. Of course, all the work is in coming up with the constraints.
Another factor is the driver. If the calculated rate is considered wrong by him,
and lap times verify his complaint, then they need to be changed, regardless what
the equations say.
A spring needs to maintain the car at the desired ride height somewhere in the middle of its
compressive range, which is also affected by where the shock mounts are. If anti-roll bars are not
used, the springs are the sole resistance to body roll, so they must be stiff enough to meet the amount
of lean the designer allows.
How much the spring is compressed, versus how much the wheel moves, is called the installation
ratio. As viewed from the front, if the spring is directly over the wheel, the spring rate will equal the
wheel rate, so the installation ratio is 1:1. This means the spring compresses the same amount as the
wheel. This is almost never true, though, due to packaging issues. It is likely that the spring will be
inboard of the wheel and probably not vertical. In this case, the spring has to try harder to support
the car at the same level (compared to if it were located directly over the wheel). Actually, its even
worse than that: if the spring is located halfway along the A-arm (installation ratio 0.5), it not only is
compressed half as much but also has twice the force on it, due to leverage. In this example, the
spring must then have four times the spring rate to have to same wheel rate.
This is the biggest reason not to feed suspension forces into the middle of
suspension links. Because the leverage force is squared, the bending force on the
suspension can be enormous.
There is a common perception that stiffer springs go with heavier carsnot
necessarily. Differing installation ratios mean its quite possible for a 1500 lb.
car to use 1000 lb. springs, and a 4000 lb. car to use 500 lb. springs; it all
depends how theyre mounted. There was a time when 1500 lb. F1 cars used
10,000 lb./in. spring rates!
There are several ways to select spring rate: trial and error, past experience, or two mathematical
approximations.
Method 1:
The rule of thumb is:
wheel rate = * sprung weight
Moving terms around:
spring rate = sprung weight / (2 * installation ratio
2
)
Example:
A car has 300 lb. on each front corner, typical for a Super-7 type car. Well assume 50 lb. of that
is unsprung weight (wheel, tire, brake, etc.), leaving 250 lb. as the sprung weight. Using the
equation, we want a wheel rate of half that: 125 lb. Lets say, though, that the springs are installed
per the figure above, with the spring located halfway between the inner suspension pivot and the
tire centerline, giving an installation ratio of :
spring rate = 250 / (2 *
2
) = 500 lb.
This is a very general guideline for a road-going sports car; a track car might have twice the rate.
While its no guarantee of a perfect value, which is subjective anyway, it a good first cut. This rule is
not appropriate for cars with wings or under-chassis aero devices, because the effective vehicle
weight (due to aero-downforce) is much more than when sitting still.
Method 2:
The natural frequency of a cars suspension is just that. Without shocks, and assuming friction-free
pivots, the car is just a big weight sitting on springs. Bouncing the car will cause it to oscillate up and
down at some rateits natural frequency:
natural frequency (cycles per minute) = 187.7 * (wheel rate / sprung weight)
1/2
Staniforth
6
notes the following:
Chapter 13: Table 1
He recommends 85125 cpm as a starting point (though Kimini is about 140 cpm).
Both methods work, and while they produce slightly different values, it doesnt really matter, since
both are starting points only. However, the overriding factor is that if the equations say to use spring
X, and you find the car is so stiff it blurs your vision, then its not the right choice for you.
Just because springs off a donor car are used, it does not mean theyll work the
same. The only way they can is if the new suspension is identical to the donor,
including the A-arms, with the same pickup points. The spring must also have the
same pickup points as the donor. Since the new car will be lighter, different
springs are needed. Lastly, if the new car is either narrower or wider than the
donor, the resulting suspension geometry will be different. Using all the donor
parts does not get the builder around having to understand whats going on.
Due to differing installation ratios, two cars may ride the same (they have the same wheel rate), but
the spring rate can be vastly different depending on the suspension. So, when your buddy says his car
will handle better than yours because hes using stiffer springs, that means nothing unless the two cars
have exactly the same suspension.
For looks, some people install wheels with huge positive offsets. This places
the tires far outboard, and then the owner wonders why the fender wears a
groove in the middle of the tire. Its because the clueless owner unwittingly
decreased the installation ratio of the springs. The wheel offset lowers the wheel
rate by about 25%, so the car compresses the suspension 25% more, and the car
rides that much lower.
Imagine the vastly increased bending loads on the spindle and wheel bearings
due to this, along with higher steering effort, and on and on. Its just the price of
being cool, I guess.
For adjustable shocks, the most common spring size is 2.5 ID. A few shocks can use smaller
diameter springsthe problem is finding them. To make life easy, stick with standard 2.5 ID
springs.
Calculating spring rate is easy. Just stand the spring on the floor, put a ruler
next to it, and apply some weight on top of it (your body weight is good). Then
measure how much the spring compresses:
spring rate = force applied / distance the spring compresses
I chose 12 in., 150 lb./in. front springs and 12 in., 225 lb./in. rears, arrived at using equations in
Milliken. (The rule of thumb came up with virtually the same values.) Instead of reprinting all the
equations, refer to the spreadsheet link in Appendix D.
Later, I shortened the front springs to 10 in. to gain more clearance for the
steering shaft.
My car designer buddy, Lee, felt I picked spring rates too high. He felt Id end
up using 100 lb. springs at the front and 180 lb. at the rear; otherwise Id end up
with tons of oversteer.
Later, spring rates were increased about 30%, because I didnt like the soft
ride. The fronts were changed to 200 lb. and the rears to 300 lb. Its another
example of how subjective many of these numbers are.
Lee said, no matter how hard you try, it always seems to take three tries to get
the spring rate where you want. Im currently on my second guess.
While these are the values I chose, other cars with different suspension designs will use different
values. Vehicle weight, CG location, wheelbase, suspension geometry, and opinion all make a big
difference in deciding whats used. Just remember, spring rate is not wheel rate!
Finally, be sure to check out circle-track suppliers for springs. More and more weekend oval-track
drivers are using high-end shocks, which use 2.5 ID springs. This means theres a pretty good supply
of springs out there, now more affordable than ever. Figure on spending $35$55 per spring.
Be sure that under full suspension travel, in compression, the springs dont
coil-bindthe springs compressing so tightly that the coils touch. If that happens,
it means the spring rate will go to infinity. At best, the car will lose traction at
that end; at worst, something will bend or break.
Natural frequency Type of car
6080 Comfortable road car
80100 Road-going sports car
100125 Race car
125+ With downforce
15: Shocks
When I started the project, I didnt give much thought to shocks, other than their being just another
component to select. I figured there were plenty that would work and didnt think much about the cost.
In hindsight, the shocks caused more angst than any other part of the car (well, okay, the doors were
terrible, but thats different). Everything I read, everyone I spoke tothey all said the same thing: do
not cut corners on shocks. In a car where handling is the primary goal, nothing makes more of a
difference (good or bad) than shocks. Shocks you buy for your family car cost about $50 eachnot so
with shocks for a real sports car. The right shocks can be very expensive, $250$1,500 each. It takes
high-quality shocks to make a car handle at its best, and if its outside the budget, we need to consider
where we can cut costs elsewhere; they are that important.
What you get for all that money is adjustability. The obvious is the threaded body, for adjusting
ride height and corner weight, necessary to balance the car. If the spring collars on the shocks arent
adjustable, how can corner weight be adjusted? If non-threaded shocks are used, ride height must still
be made adjustable through some other methodits critical to have.
Just as important is adjustable valving, the better shocks being double-adjustable. This means you
can select one damping setting on compression (hitting a bump), yet have a completely different
setting for rebound (recovering from that bump). Having both adjustable is the only way to maximize
handling of the car. Using any old non-adjustable shock is a completely blind guess, one unlikely to be
correct for the car whatsoever.
If your car does not use anti-roll bars, there are only two ways to adjust
handlingsprings and shocks. If you get cheap non-adjustable shocks, there is
then no way to adjust the transient handling on corner entry and exit. (Tires are
another variable, but since tire pressure should be set for even tread wear, they
should not be used to alter handling.)
Are there alternatives to expensive coil-over shocks?
Yes, sort of:
1. Buy used shocks.
Buy name-brand, used racing shocks and send them to the factory to be rebuilt (a factory race shock
rebuild is common practice). The factory can also revalve the shocks, so its okay to buy shocks off a
slightly heavier or lighter caras long as they are double-adjustable.
2. Use appropriate motorcycle shocks.
Thats the trick, finding appropriately valved units. While I could guess by the weight of the bike
what units might work, there was no way to know if the valving would be correct. A big unknown,
just like with the springs, is the installation ratio. It can make poorly valved shocks work okay or
make properly valved shocks completely useless. Some people go solely by price when buying
shocksending up with mystery shocks. (Kind of like the stuff you used to get at your elementary
school cafeteriamystery meat.)
If you can find shocks from a known source (car or bike), there is a better chance of at least
knowing the mass of vehicle it was designed for, plus the installation ratio. Unfortunately, most
people dont do this, the thinking going something like, I need shocks but cant afford the expensive
ones, so Ill buy cheap ones off eBaytheyll be fine. How can a person know what they are
getting?
So, what do they end up with? They dont know. How will they perform on the car? They dont
know. Are they too soft or too hard? They dont know. While this hit-and-miss method of design
saves money up front, its likely the shocks will have to be replaced with another set later on. But
buying better shocks means knowing what the old ones werean unknownhow can they know what
to replace them with? I saw this as another big time-sucking side project, with hidden expenses.
3. Convert regular shocks into coil-overs.
Check out how this guy converted ordinary shocks to coil-overs: georgecushing.net/Cheapo.html.
This fixes the spring adjustability issue but not the shock valving.
4. Dont use coil-overs.
Ride height can be adjusted not only with the threaded collar on the shocks, but also with
adjustable pushrods or rocker-arms. In other words, it is possible to use regular shocks without the
adjustable coil-over feature. Unfortunately, the only shocks I found having double-adjustability also
had adjustable spring perches. Regardless what type of shocks are used, some method of adjusting
ride height must be designed in. If not, there will be no way of adjusting corner weight. However its
adjusted, be it shock collars, adjustable rocker arms, or shims, each corner must be adjustable.
If I were doing it over again, Id check into shocks from the circle-track
community. Because their market is so much larger, their prices are often lower.
They frequently have the same features as the more expensive shocks; Fox and
QA1 (HAL) are two of them.
I didnt do any of the above. I wanted to build a car, not design and test shocks. Buying off-the-
shelf shocks intended for a 3000 lb. car wouldnt work unless they were modified, but how would I
do that? Yes, valving can be modified by the factory to some degree, but I wasnt sure by how much.
And even if I figured out how to modify them, how much was I supposed to change them? I had no
idea. So in the interest of getting on with it, I decided to buy new name-brand racing shocks.
Looking back, Id give Bilsteins a closer look, since they are gas shocks and
can be mounted in any positionperfect for actuating with pushrods. Also check
out AVO, JRZ, KYB, hlins, QA1 (HAL), and Spax.
First I needed to pick a manufacturer. While Penske and Moton shocks are the best, at $1,200
$1,500 each I simply couldnt afford themtime to compromise. Koni had been around a long time;
their shocks were standard issue on professionally built race cars back in the day. Since they were
installed on many lightweight race cars, I had a good chance at finding some for Kimini. Besides,
since I was making something of a retro car anyway, they would fit right in. Calling Koni confirmed
they could provide the proper shocksif I could tell them the spring rate. Luckily I could (see the
previous chapter, Springs), so I chose Koni 8212 aluminum double-adjustable shocks (image credit
Koni). Shock housing length and stroke must also be specified; six-inch stroke was chosen since the
shock motion ratio would be nearly 1:1. Six inches of suspension travel was already decided on. An
important component of the shock assembly is the rubber bumper on the shaft. In many suspension
designs, this is the only thing keeping the suspension from hitting the chassis, or the tire from hitting
the shell. I cut mine down, because it used up too much of the shock travel, but retained the tapered
portion. Whatever you do, always have suspension bump-stops; if they must be removed from the
shocks, they must be located someplace else in the suspensionbut never eliminated.
Such high prices for the shocks made me wonder about Ron Champions book,
optimistically titled, Build Your Own Sports Car for as Little as 250 and Race
It!. At the time his book was written, that worked out to about US$400. Dreams
of building my car for anywhere near that vanished in a flash, considering that
just one good quality shock cost more than his entire car!
Shock makers manufacture many types of shocks for many types of motor
sports. Be careful not to choose a shock made for drag racing. Drag racing shocks
are valved completely differently than those for street or road racing. Do not buy
them; even though they are cheaper, they wont work.
Regardless of suspension type, the shock absorber shaft should be centered at normal ride height.
This means that with all fluids and the driver aboard, the shocks should be compressed about 40%
50%. This is yet another reason to use a spreadsheet. It allows for calculating how much the shocks
will compress for a given mounting orientation and loadwithout having to build the car first.
As said before, no matter what shocks you use, the car will do pretty well. Putting an engine in a
car weighing half what the engine came out of is guaranteed to provide a fun ride. Even if top- of-the-
line shocks arent in the budget, the car will still provide endless entertainment. Quality shocks are
very important, yet we all have different budgets. Its true, you get what you pay for, but the difference
between a project car costing $10,000 and $40,000 isnt a factor of four in performance. Its that last
10% that costs silly amounts of money; as long as were happy with 90% of perfect, thats all that
counts.
16: Steering
Design Paralysis the inability to make any decision whatsoever due to numerous conflicting
issues.
Everything is connected to everything else. Me
The important partsthe drivetrain, wheels, tires, uprights, and brake discs and calipersand the
suspension points have been chosen; time now for how to steer the little monster.
Steering rate
What makes for good steering? Since Lee had designed the Rotus, I asked him what steering
parameters worked well for him:
Two turns lock-to-lock
+/-30 steering
About 1.752.0 rack speed
How to measure rack speed
Turn the pinion (steering shaft) exactly one turn and measure how far the rack
moves. Often, it can be measured while still mounted in the car. Measuring first
means that you dont waste time removing a rack that isnt what you want. (To do
this test, the steering wheel lock may have to be adjusted using extreme
prejudice to allow the wheel to turn.)
A rack slower than 1.75/turn is more appropriate for a cruiser, but it is very much personal
preference. My research led me to believe that the above values are about as far as youd want to go
for a street car. Much more and the car can change lanes when you sneeze. Keep in mind these are
desirable goals, but any manual steering rack will work; its just that some work better than others.
In the everything is connected to everything else column, the above goals
depend on wheelbase, too. The longer the wheelbase, the more steering angle is
needed to turn the car in the same radius (as a small car). Then again, the term
need is relative, too; it all comes down to what you will be happy with. I often
see arguments about how much Ackerman, toe, or caster is the right amount. It
all depends what the driver considers right; it all depends what you are happy
with.
Driving the car after it was done confirmed that the steering was just right.
Another variable in steering is the length of the steering arms on the uprights. Long arms mean
slower steering, with a more limited range of motion, but with better leverage (less steering effort).
Short arms give fast steering, with a wider range of motion, but with more steering force. Steering-
arm length can help turn a so-so steering rack into a very acceptable one (usually by speeding up the
steering). Also keep in mind that a rear- or mid-engine car will not have much weight on the front
tires, so steering effort may be quite light, even with short steering arms.
Rack length, measured between the steering rack ball joints, is important. The rack needs to be at
least as long as the distance between the inboard suspension mounts to make adjusting bumpsteer
easier. While a rack can be disassembled and cut or lengthened, I wanted to avoid it. Call it
efficiency, or laziness, but the idea was to get on with the design, not to modify yet another part.
Another good reason to leave it alone is that when it is time to replace it, a stock unit will drop right
in.
Bumpsteer
My sister had an old MG Midget. Going around a constant radius turn, one that you could go around
without moving the steering wheel, provided a good example of bumpsteer. At one spot in the turn
was a small bump. When the car reached it, all by itself, without the driver moving the steering wheel
at all, the car steered about two feet to the left as it hit the bump, then two feet back to the right as it
passed over it. Thats bumpsteer, and it made for an interesting drive. Its not a good thing; we want
a car that doesnt do anything odd. We want one that goes where we point it, and nowhere else;
eliminating bumpsteer accomplishes this.
Part of suspension design is accurately placing the steering rack so that no bumpsteer occurs. To do
this, the rack must reside in a magic location between the upper and lower A-arms, where all
bumpsteer is eliminated. Finding this magic location can take hours; this is where suspension
software can really helpto a point. You enter all the specs of the rack and where you want to put it.
The software can then tell you, before you even build the car, how much bumpsteer you will have. So
you mess with it for hours until it is as good as you can get it.
Build inaccuracies guarantee that there will always be some bumpsteer to
adjust out.
Steering rack placement
Like many design issues, it depends. The donor rack should be placed in the same position (front-
to-back) as it was on the donor vehicle. In my case, the Triumph Spitfire rack was a front-steer, so I
had to place it in the same position; likewise for a rear-steer rack.
Front steer and rear steer
This refers to where the rack is in relation to the front axle: in front of it or to
the rear. Its important, too; if it is moved from its original position, the car will
steer backward! A wrong rack can be turned upside down to correct its
left/right motionif the steering shaft can still reach the input shaft.
This answers where to place the rack, as seen from above, but theres still the question of where to
place it vertically.
Vertically, it is placed the same proportional distance between the pivots:
BC / AC = EF / DF
A simple example is if the outboard steering point, B, is 1/3 of the way between the upright pickup
points, A and C, then the inboard point, E, must also be 1/3 of the way between the links at its end, D
and F. This means that the rack can be one and only one length! Sometimes, though, it can be designed
in reverse, moving the rack up and down until point E intersects the line DF. Once the rack is fixed in
place, then the outboard steering point, B, is then required to be in one specific location.
Unfortunately, this is only feasible if custom uprights are used, or the steering attachment (point B)
happens to be where it needs to be. If you are determined to use existing uprights, then the rack has to
be the one that changes, with the existing steering arm being modified or even cut off and relocated.
I saw a novel way to lengthen a steering rack without having to modify the rack itself. In the figure
above, a frame that picks up the existing ends of the rack is fabricated and extends them out as much
as needed. The only requirement is that the frame be both stiff enough to avoid bending and have room
for it to pass alongside the rack. If the rack needs to be shorter, which is very likely, thats not as easy
a conversion. The rack must be disassembled and shortened, or a shorter rack should be chosen to
begin with.
Once the car is built with the rack placed right where the CAD software
specifies, dont expect perfect steering! Thats because its very, very difficult to
make the chassis exactly as the specs call out. Metal warps and twists when
welded; even when in a fixture, tubes get cut a little wrong, things bend, plus,
were only human. Dont expect to get closer than about 1/8 of perfectit is
one reason race cars use adjustable suspension!
Heres a big tip: Build the chassis and then attach position-critical
brackets. This way, it doesnt matter if the chassis is warped a little.
Yet another way to approach the steering issue is, as I did elsewhere, to design it in reverse. That
is, use a stock-length rack and design the suspension around it, making the suspension pickup points
accommodate the chosen unit.
Candidate steering racks
Finding a manual steering rack that meets all of the above specs is getting difficult, since almost all
cars now have power steering. Heres a list of manual steering racks that can be used in a light car:
Triumph Spitfire
Chevy Chevette
Dodge Omni
Ford Pinto
80s90s Ford Mustang
First-generation RX-7
Fiat X1/9 and 128
VW Fox or first-generation VW Rabbit
Ford Fiesta, German Capri
Classic Mini
Miata (see Tanner
7
for how he shortens a Miata rack and converts it to manual.)
Mk1 Toyota MR-2 (if manual steering)
Porsche 914
Toyota Tercel (80s)
MGB
Suzuki Swift
Geo Metro
Aftermarket dune-buggy, or circle-track (see circle-track suppliers in Appendix B)
I havent confirmed which of these have decent speed, but its a good start. While power-steering
racks can be converted to manual, theres still the issue of whether the rack has sufficient speed.
Sufficient speed?
It refers to how much the rack moves per turn of the steering shaft. Depending
on what youre building, this might be a big deal, or not. For a sports car, 1.5 to
2.0 per turn is about right. Note that a slightly slow rack can be sped up by using
shorter steering arms.
Looking for a rack wasnt too promising; almost everything in the wrecking
yards used power steering (this was before compiling the above list). While
many Cobra kit cars use the Triumph Spitfire rack, finding one in Southern
California was tough, but one day at the wrecking yard, I came around a corner
and couldnt believe my eyes: there was a 1979 Triumph Spitfire. There are good
reasons the Triumph rack is used in sports cars, too: its light and has 1.9/turn.
This one was in good condition, so I bought it immediately$40 (new ones are
around $350).
Combined with steering arms of the right length, I met all three goals: 2.0 of
rack travel/revolution, two turns lock-to-lock, and +/-30 of steering. Driving the
car validated these numbers; the steering was fast, with light effort, though
anything faster would have been too sensitive, at least for me.
Lastly, theres the question of using a steering-box instead of a rack-and-pinion. While anything can
be made to work, I dont recommend it because of the required space, lack of feel in the steering, and
increased weight.
Ackerman
A very common and hotly debated discussion is how much to use. Ackerman is how, in a turn, the
inside tire must turn tighter than the outside tire, so neither tire scrubs across the ground.
Ackerman was developed to prevent horse-drawn carriages from disturbing
the gravel as they turnedin the driveways of the upper class, no doubt!
Many of the original perceptions of how much Ackerman to use came from early stock-car race
teams. It is better understood now, but there are still plenty of conflicting opinions. I found arguments
either way, to use a lot or none, or even negative; it got so I threw my hands up. At first I was going to
have none, because Milliken suggested zero, but then I read some interesting comments by Carroll
Smith. In his earlier books he had a theory about why zero, or even anti-Ackerman, should be used.
Later on, though, he revisited the issue
8
.
He said that during a testing session they had some free time, and something hed always wanted to
do was to add or subtract Ackerman and see what happened. Much to his surprise, he found that the
more Ackerman they dialed in, the faster the lap times. The drivers also said that the car felt better
and was easier to control, and understeer decreased. Of course, his testing was on a race car, but it
made enough of an impression on me that at the last second, as much Ackerman as possible was
designed in. The amount ended up being such that the steering-arm projections crossed somewhat
behind the rear axle, giving perhaps Ackerman. Perfect Ackerman means they cross at the center
of the rear axle, but in my case, the steering arms couldnt extend any further outboard; they would hit
the brake discs. Based upon the intended use of Kimini, this seemed to be a good compromise.
A very practical reason for wanting a decent amount of Ackerman: since the
tires arent trying to slide across the road, it makes it much easier to push the car.
For a home-built car of questionable reliability, this could be important!
17: Anti-roll Bars
An anti-roll bar (sometimes called a sway bar) does just that: it keeps the car from rolling
(leaning). Think of it as a big hinge, across the car, side-to-side, with one side of the hinge attached to
the front or rear wheels. It works by cross-coupling the suspension; when the car leans in a turn, the
outside tire thats compressed also compresses the inside tire, the result being that the car doesnt
lean as much. Unfortunately, we dont get something for nothing; by adding the bars we just made our
independent suspension not so independent. It means that when we drive down a straight road and
hit a bump with one wheel, that bump force is now also fed across to the other wheel.
The perfect sports car suspension is not upset by bumps; if one wheel hits a bump, it deals with it,
not bothering the other corners of the car with its own problems. The stiffer the bar, the more we
approach, in effect, a solid-axlenot desirable for a nimble sports car.
Now we can see some more of the interconnectiveness of the suspension system. If we had
placed the roll centers as high as the CG (a bad idea), the car would not lean at all. Since the car
doesnt lean, adding anti-roll bars would have no effecttheres no roll! On the other hand, placing
the RC down near ground level gives the bars lots of motion to restrain, so this is good for tuning
purposes. A designer once told me to think about anti-roll bars this way: They destroy grip at their
end of the car to make the other end handle betterin comparison. The car may go fasterbut grip
is given up in order to balance the car. If the car can be designed to be balanced without them, grip
would be at a maximum.
This is just what I attempted to do: design the car so bars would not be needed. If it could be pulled
off, the car would handle at its very best; plus it would be lighter, simpler, and cheaper.
At the autocross, Kimini was so nearly balanced that bars were deemed
unnecessary.
Often, though, bars are needed anyway, due to unforeseen circumstances such as the driver not
liking the initial handling, for example. So just how stiff do you need them to be? It all depends how
much you want to reduce roll or transfer grip to the other end of the car. One common misconception
is that anti-roll bars decrease weight transfernot true. For a given car with a given CG height and
track width, in a corner of known radius and at a known speed, some amount of weight transfer will
occur. Nowhere in the math is there a factor for anti-roll bar size. Whether or not the car leans has
nothing to do with weight transfer.
How to find anti-roll bar stiffness:
Roll-bar installation ratio K = I / (I + L)
Track = T
Outside bar diameter = Do
Inside bar diameter = Di (for a solid bar, Di = 0)
Effective arm radius = E
Bar length from centerline = D
Bar arm length = A
Stiffness, lb/in per degree of vehicle roll = S
Note that the tubing diameter is raised to the 4
th
power. It means that using a tube that is just a
little bigger makes a huge difference in stiffness. For example, going from a 0.75 diameter bar to
one that is 0.8125only 1/16 thickermakes the bar 38% stiffer. Be careful changing anti-roll bar
diameter!
This equation shows why its important to build a chassis from large diameter thin-wall tubing. For
the same weight as small diameter thick-wall tubing, it is much stiffer.
18: Forces
Suspension forces
Forces in suspension tubes can be much higher than expected; it pays to know just how high. In our
quest to keep weight down, its all too easy to make suspension links or brackets too small, risking a
nasty failure.
When designing the suspension I ran into a common dilemma. During suspension design I wanted to
know what the forces in the suspension would be when the car cornered. Thanks to the spreadsheet, I
knew what the car weighs and how much force is on each wheel in a turn. But how hard will the car
corner? I had to pick a number, so I guessed1.0 G. That allowed me to calculate the forces in all
the suspension links, and that seemed to be the end of itbut it wasnt, due to the unknown nature of
bumps.
I saw a Locost Seven that used the typically thin A-arm tubes ( OD) and 1/8 wall tubing for the
lower A-arms. With this particular car, though, the tubes had been fabricated already bent, with the
lower shock mount attached near the middle of the tube. For a Super-7 car, each corner sees about
300 lb. static load. Since the shock/spring was roughly in the middle of the tube, the spring must exert
four times that force to hold it in position, or 1200 lb.
Four times?
Moving from the outboard end to the middle of the tube does two things: the
spring is now being compressed half as much and is subjected to twice the force,
yet still has to support the same weight. Each action requires doubling the spring
rate.
The springs were also tilted inboard about 30 off vertical, so the actual spring rate ended up
around 1400 lb., all to supply a wheel rate of 300 lb./in.
The angle of the spring doesnt have much effect as long as its near 90, but
the further away from 90 it gets, things go downhill in a hurry. The factor to
apply is: spring rate / sin(A), where A is the angle between the shock body and
the actuating arm. In the above example, the inclined spring rate must be 1.15
times stiffer to apply the same force as a vertical spring; hence the 1400 lb.
The most dangerous aspect of this suspension is exposed during braking and corneringand hitting
a potholewhen shock loads of about 5x the static load feed into the shock absorber mounts. Its the
worst possible situation: a small-diameter tube, already bent, under heavy compression, is subjected
to a bending load of 1400 lb. * 5 = 7000 lb.! It is a failure waiting to happen.
Colin Chapman called such tubes pre-failed.
I describe this simply because if a suspension design looks bad, it probably is. Spending just a
little time calculating some numbers can quickly show if itll be okayor if it can kill someone. Sure,
suspension arms can still fail, but at least some effort can be made, some research done, and some
numbers calculated, instead of just blindly using some tubing and having absolutely no idea if it will
work.
Applying bending loads of any kind is poor design, so try not to. Put shock absorber brackets as
close as possible to the outboard end of the lower A-arm, or even directly on the upright, which can
be done on the rear suspension since it doesnt steer. Larger-diameter thin-wall tubing is stiffer and
lighter than smaller thick-wall tubing. If the installation requires the suspension arm to be bent, it must
be of an even larger diameter and/or have greater wall thickness to deal with the additional stress. I
used a Strength of Columns table in a mechanical design book to determine the tubing diameter.
Yes, it is possible to use bent tubes, but they have to be much stronger to
withstand bending forces in tension and compression. This means the tube must
be of a larger diameter, thicker-walled, heavier, and more expensive. Another
aspect of bent tubes is that its nearly impossible to find design data. There are
no easy-to-find tables titled Strength of bent tubes!
Regarding bumps and potholes, most amateur designers are in the same boat; just how much force
does a bump apply? Lets say that a car is traveling along a smooth road and hits a pothole. What
force is fed into the suspension arms at that instant? Its impossible to answer because there are no
calibrated potholes. How large and deep is the pothole, and how fast is the car traveling? On top of
that is tire pressure; the softer the tire, the more it will take the hit instead of the suspension. In the
end, I had to guess; I picked 10 G as the ultimate shock load the suspension could handle.
I still wince whenever Kimini hits a hard bump. Even changing lanes on the
freeway, hitting the lane divider bumps (Botts dots) really sends a pounding
through the chassis.
Suspension pivot forces
Rod-ends are so strong that amazingly small units could be usedbut shouldnt. The reason is
twofold. First, we can never really know what force is fed through the bearings when hitting a
pothole. Its better to choose a stronger unit than what the equations give as the bare minimum.
Second, Teflon-lined bearings do wear out, but larger parts last longer because they arent being used
anywhere near their limits.
Rod-ends are very strong with the bore oriented axially to the load; this is the strength value listed
in catalogs. Not quite as easy to find is the strength when the rod-end is oriented the other way, with
the force trying to push the ball out of the housing. They are much weaker in this direction, on the
order of 5x10x weaker. An impulse load, such as hitting a pothole, can pop the ball right out of the
housing. If you must use a rod-end in this orientation, at least add a large washer to the mounting bolt
to contain things if the ball pops out. Better yet, use them the right way or just use automobile ball
jointsthey usually have a larger range of motion anyway; they just arent as sexy.
Proper suspension mounting
This illustration shows several problems to avoid when designing suspension. A sure way to
guarantee suspension failure is to have a shock or an anti-rollbar feeding loads into the middle of
suspension arms. The only way to avoid problems is to move the force to one end of the tubeor use
thick, heavy, A-arm tubing to combat the side-load. As said earlier, applying a 7000 lb. bending load
into a suspension arm is not asking for troubleit guarantees it.
Another error is attaching suspension brackets to unsupported chassis tubes; feeding suspension
loads into the tubes with offset brackets is even worse. Doing so produces a twisting force on the tube
for no good reason at all. The way to prevent this is to line up the applied force so it passes through
the center of the tubes.
Another goal is to always feed heavy forces into a junction of three tubes. Why? Because only then
is the mounting point rigid in three dimensions. Anything less and the mounting point will flex. Or,
once again, the heavy-handed method of using thick-wall tubing to fight bending forces can be used.
With a little thought and care, a correctly designed bracket (smaller and lighter) will provide a much
more reliable mount.
Finally, place all critical mounts in double shear! Note the bracket at far-right; it must be much
thicker than the two thin brackets at left, making it heavier, and it still bends. Single-shear is bad
newsdont do it.
Toe links
Steering links at the front of the car, and toe-control links at the back, are two of the most highly
stressed points in the suspension (right up there with shock mounts and rear lower A-arms). This is
not a place to use a weenie little bracket in single-shear, attached with welds that look like, as my
buddy Craig calls it, dog-hurl. Place these and all highly loaded points in double-shear!
Drivetrain forces
Next, consider the force fed into the chassis by the drivetrain, specifically the differential. The
differential will always have the maximum force on it because its last in line at multiplying engine
torque through gear reduction.
Consider the case of the Miata, currently a popular donor for Super-7 type cars. Many people use
the entire rear suspension assembly as is; they figure that the factory has a nice setup, so why mess
with it. Part of Mazdas suspension design is a large brace called a Power-Plant Frame (PPF) running
between the differential and transmission. Rigidly connecting the differential to the
engine/transmission helps contribute to the Miatas excellent handlingbut mostly its there to deal
with the tremendous torque fed into the differential. The problem with Super-7 type cars is that
theres very little space for the PPF to run alongside the driveshaft, leading some builders to leave it
out.
A typical Miata engine puts out 115 ft. lb. of torque, and with the transmissions first gear reduction
ratio of 3.1:1, there is 115 x 3.1 = 356 ft. lb. of torque heading through the driveshaftbut we arent
done yet. The differential ratio is about 3.9, so now there is 3.9 x 356 = 1390 ft. lb. of torque to deal
with! Its actually worse than this, because that torque is applied, removed, and even reversed.
Several builders have found this out the hard way when either the frame mounts break or the
differential housing itself cracks. Since this doesnt happen on the donor car, it tells us something. If
youre planning to use a Miata donor, try to find space for the PPF. If it just wont fit, fabricate the
differential mounts so they can handle 1400 lb. of force, because thats what theyll see.
Even if youre not using the Miata as a donor, be aware of that torque multiplication. It all has to be
fed into the chassis somewhere; exactly where depends on the particular design. On a straight-axle, it
is fed into the trailing links. In a de Dion or independent rear suspension it feeds into the differential
mounts. Kimini, with the FWD drivetrain mounted mid-engine, has the torque fed into the front and
rear engine mounts. The point is, whatever mounts are used for the suspension and differential, they
need to be really strong. Use at least 1/8 bracketsand thicker if theyre not triangulated. Use AN-
bolts and nylon lock nuts, and really tighten them down. The danger is that if they start slipping, the
bolts will slowly elongate the mounting holes, resulting in even more play.
Small differentials, big engines
People building cars from scratch use engines and differentials from many different carspretty
obvious. What comes up periodically is someone using a smaller differential with a big engine
typically coupling a Miata differential to a V6. The concern is whether the differential (intended for a
four-cylinder engine) can last long, being fed more than twice the torque it was designed for. I say yes
with a couple of conditions.
If the car being built is much lighter than the Miata, and if the tires are of similar size as the Miata,
it will be okaythe key is how much torque the differential sees. The Miata differential never sees
the higher torque because it gets vented off through the tires. In fact, in a Super-7, the differential
will see less torque than it did in the Miata because the Super-7 is lighter.
Theres some amount of torque coming into the differential and some amount being bled out
through the tiresthe differential sees the difference between the two. As long as that torque is less
than what the Miata differential saw in its original application, itll be fine. Taken to an extreme, even
a V8 wouldnt make any difference to the Mazda differential lifeall that extra torque just makes
more tire smoke.
The irony is, it doesnt matter how big the engine isthe traction (and torque handled by the
differential) is determined only by tire traction. Off the line, a Miata-powered Super-7 will, for the
first 50 ft. or so, accelerate just as fast as a V8-powered Super-7. The tires can only transfer so much
torque; feeding 10 times that amount into the drivetrain will only result in more tire smoke, not a
faster car.
Of course, all of this goes out the window when the owner slaps on 10 wide racing slicks. Now
the differential will see nearly all the torque fed through it, because the tires dont bleed it off. Even
putting sticky racing slicks on a stock Miata can cause the differential to fail.
The point isit depends. It depends what the differential is being put in, what the car weighs, and
what the tires will be. It may turn out that a small, light differential will be perfectly reliable; it all
depends on the application. If you are building a cruiser, youll be just fine.
Range-of-motion limits
When designing the suspension, be very aware what angles all the pivot points move through. All
joints, be it rod-ends or automotive ball joints, have a specified range of motion. Rod-ends are
typically around 812, which isnt very much, while ball joints are up around 20. This is another
good reason to use long suspension links, because the angles the rod-ends move through will be less
than when using short links.
This range-of-motion issue must be dealt withbefore the car is driven. This means installing the
suspension arms, uprights, and shocks (without the springs). Move the suspension through its travel
and make sure the only thing stopping it is the bump-stop (which you must have). If this is not
checked, one of the rod-ends may be acting as the suspension limiter. Hitting a dip in the road may
mean that the weight of the car is trying to bend a rod-end or bolt check! Since rod-end shanks are
threaded, theres already a nice stress-riser waiting to be given a chance to crack. If it breaks, youre
on an exciting ride, meeting a hedge, wall, or oncoming traffic.
If a rod-end is the limiting factor, there are high-misalignment rod-ends just for this situation. In my
case, the rod-end connecting the front upright to the steering tie-rod was getting too close to its end-
of-travel; a high-misalignment rod-end saved the day. You cant just swap one type for the other; the
bolt holes are smaller for a high-misalignment part with the same bore. In other words, if you replace
a regular rod-end (one that a bolt will pass through) with a high-misalignment rod-end, the
is actually the size of the threaded shank. The bore will actually be 7/16; thats how they get the
additional travel.
Also be aware that rod-ends arent the only things that can limit suspension travel. In the same
picture above, the steering-arm bracket on the upright had to be trimmed back enough to ensure that it
couldnt hit the threaded shank of the rod-end, stop nut, or steering arm itself.
Production ball joints have quite a bit more angular travel than rod-ends, so dont discount using
them just because they arent sexy. I used factory Honda Accord ball joints for locating the rear
suspension uprights for this reason (more travelnot because they are less sexy; I dont have a
problem with sexy, really).
Lastly, it isnt always suspension arms, ball joints, or rod-ends that can limit travel. Notice how
tightly stretched the flex-line is in the picture below. This is the front suspension under full droop and
maximum steering lock. This is very poor design, and I had stern words with the designerme. This
was a case of not moving the suspension through its full range of travel. I think I had eyeballed this
and judged it to be fine. I dont know how sharply stainless flex-lines can be bent or how many times
they can be bent this sharply, but it was alarming and embarrassing to find this. They were replaced
with longer hoses.
What size tubing do I use?
This requires a couple of mechanical engineering books. Two texts are needed, a Statics and
Dynamics text and a Strength of Materials text. Both are available on-line or through university
bookstores.
The Statics and Dynamics text is used to calculate forces in suspension links and chassis tubes.
Then, the Strength of Materials text is used to select the proper size tubing.
In Kimini, all suspension A-arm tubing is straight 1.0 OD, 0.065 wall. Due to higher forces at the
rear of the car, the lower rear suspension arms are 1.25 OD, 0.065 wall.
A larger-diameter, thin-wall tube can be stronger and lighter than a smaller
tube with thick walls. A chassis made of fewer larger-diameter tubes will be
stronger, lighter, and simpler than a chassis made with many small-diameter
tubes.
The roll-cage tubing size was dictated by the SCCA rule book, 1.5 x 0.095 wall DOM.
One good reason for having a spreadsheet is the ability to list the weights of
all components. This is valuable because the SCCA assigns roll-cage tubing
sizes based on the weight of the car. Without having a good weight estimate first,
how do you know what size to select?
19: Chassis Design
Chassis design is connecting the dots around all the hard-points and big hardware.
Ron Schramm
Safety
All through the design, try to consider what would happen in a wreck. Okay, I realize its nearly
impossible to know absolutely what will happen, but its better to at least think about it. For example,
its pretty obvious that its a bad idea to have a one-piece steering shaft aimed straight at your head.
In the 1950s thats exactly how steering shafts were designed. I saw several
sad pictures of deceased drivers with their heads pinned to the roof of the car by
the steering wheel.
In the case of the steering shaft, add a U-joint (or two) to break the shaft into segments avoiding
having one long steel shaft aimed at your head. In the event of an accident, the two (or three) segments
will compress into a Z instead of spearing you. It also means restraining the upper end of the column
with a split-collar clamp, to not let it slide toward the driver.
Another option is to use a collapsible column assembly, either from a donor
car or a hot-rod supplier.
Carefully consider the seat-belt mounting brackets. Worst case, there could be 10 times your body
weight on the beltswill the brackets hold?
The fuel tank or cell: where will it go in an accident? Is it acting as a crushable structure? Is it so
rigid that it will get pushed into the passenger compartment in a crash? Will the filler tube get ripped
off and spew fuel everywhere? In my case I tried to place it so that it would get pushed back between
the seats.
Okay, enough of the gloom and doom. With roll centers already chosen, we can move on with the
rest of the overall design: ride height, ground clearance, axle alignment, and the consequences of tire
diameter.
Ride height
Ride height is related to the center of gravity; many people attempt to lower the CG by lowering
their car as much as possible. While the intent is good, some dont understand what theyre doing,
only caring about looks at the expense of performance. Some owners drop their cars so low that they
ride on the bump-stops; this is modifying a car for vanity, not handling. With no suspension travel, any
bump hit while cornering will cause roll stiffness to go to infinity, and the car will slide. Again, in an
effort to make the car look like a race car, the modifications actually make it less of one. There must
be suspension travel to corner as fast as possible; the tires need to stay in contact with the ground.
Since we are designing a car from scratch, we can make the car as low as we want and still have
plenty of suspension travel, but we have to draw the line somewhere. Since the car will be street-
driven, it has to negotiate real-world obstructions like potholes, speed bumps, driveways, and
railroad tracks. What if a truck drops a brick in the road? Will the car pass over it without hitting?
Sooner or later, the car will have to drive over some road debris.
On the forums, I saw someone bragging that he had 1.5 ground clearance
under his oil pan. I think that anyone driving a car on the street with 1.5 ground
clearance needs to give up drugs.
Even on the racetrack, when we go off course, there may be large bumps or rocks. Quite often, the
soil to either side of the track is several inches lower than the track itself. Either we design the car to
deal with these issues or they will deal with the car.
There are four closely intertwined issues concerning ride height: oil pan ground clearance, axle
half-shaft alignment, wheel and tire diameter, and suspension travel.
Oil pan ground clearance
If the oil pan is the lowest point on the chassis, we risk damaging the pan and losing all the oil.
Then we get to buy a new pan if were lucky or a new engine if were not. Or maybe we crack the
pan and cover the rear tires in oil and go for a wild ride, destroying the car. We want to protect the
pan, and this means placing it higher than the bottom frame rails.
Some people deal with a tall engine by cutting down the pan, making it
shallower so they can lower the engine. This is fine for a street car, but its a
really bad idea for a car to be used for autocross or on track. The car will likely
pull over 1 G, which means the oil will be at a 45 angle. Will the oil pickup still
be submerged if the pan is 1/3 shallower?
Axle half-shaft alignment
Whatever ride height we select, its important that the half-shafts come out of the transaxle at near
90; they are not intended to operate at extreme angles when transmitting power. Other than a small
offset to prevent wearing a groove in the bearing races, they should be nearly straight. CVs cannot
pass big power at extreme angles and last very long. They should stay more or less straight from the
transaxle to the hubs, especially during acceleration.
This is the main issue in deciding vertical engine placement when using a
FWD drivetrain in a mid-engine configuration. Technically, its possible to rotate
the entire drivetrain around the axle outputs, but it doesnt work well due to the
oil pan orientation and shape of the top of the cylinder head. If the engine is
tipped too far, the oil may have trouble returning to the sump from the head if the
return holes are on the high side of the head.
Wheel and tire diameter
In addition to being lighter and cheaper, small wheels and tires also help lower the car. We cant
just pick any wheels and tires without considering the entire designthe big picture. The radius of
the tire dictates the location of the half-shaft axle ports on the transaxle, as detailed above. This is a
good example of how everything is connected to everything else, and are we happy with the results?
The axles should be fairly straight with the driver onboard. Determining how much higher the
drivetrain must be mountedsitting there all by itselfis a bit of a guess, unless we already know
spring rates. A mock-up helps find the proper axle position, with a slight downward angle toward the
wheel. Shown here, the outboard end of the axle is at ride-height, which for the chosen tire was
10.75 (its radius) above ground. The moment of truth was measuring from the ground to the bottom of
the pan, showing 4.75 clearance. Proper axle alignment forced this number, which turned out to be a
good compromise between a low CG and providing real-world ground clearance. Based on this, 4.5
became the final ground clearance, with the bottom chassis rails below the pan for protection. In
addition to protecting the pan from impacts, the under-tray could now extend under the engine for
smoother airflow beneath the car.
Suspension travel
Given this 4.5 ground clearance, how much suspension travel can we have, how much did we
want, and how much did we need? Since were desigining a street-driven car, there has to be enough
bump travel to deal with everyday concerns, yet it should not be so stiff as to knock out teeth. But how
much is enough? After reading the books (which all deal with race cars with very little suspension
travel), talking to designers, and working with the spreadsheet, I determined that 3 of bump travel
looked like a good compromise.
With 4.5 of ground clearance and 3 of suspension bump, that leaves 1.5 worst-case ground
clearance. I also tested the case of someday using racing slicks, which are 1.5 smaller in diameter
than the street tires, so the car would sit 0.75 lower. This brings minimum ground clearance down to
0.75, just enough room for some rub strips.
This 3 of bump travel means 3 of clean travel, not including approximately 1 for a bump-stop.
The special conical bump-stops on the shock absorber shafts give a gentle rising rate instead of a hard
sudden stop. Under worst-case conditionson the track, using racing slicks, braking hard, and hitting
a bumpthe front of the car will just miss the ground, as it should! The chosen shocks had 6 of
travel, leaving 2 for suspension droop, which I considered sufficient.
Next time I might shift it to 3 of droop and 2 of bump + 1 of bump rubber.
That allows the inside tire to remain on the ground longer during hard cornering,
but the answer wont be clear until its driven hard. It can always be changed
later by moving the upper shock mounts.
Human factors
Since I started with the shell, the car had to be designed inward, from the shell. But it also meant
designing outward from the driver, engine, fuel tank, and suspension. Without being comfortable, the
driver wont be able to extract maximum performance, and the passenger wont be happy, either.
Even when simply cruising, who wants to hit his head on the roof or his knees on the steering wheel,
get his shoes fouled in the pedals or cooked by the exhaust, or hit his knuckles on the dash when
shifting?
Here are some points to consider regarding human factors:
I sat in the seat on the floor of the living room, and while holding the steering wheel where it
was most comfortable, my wife took measurements. This dictated a number of important
dimensions: seat inclination, distance between seat and pedals, and the relative location of the
steering wheel. (I forgot to wear my helmet when we did this. Because of this, the seatback
pushes my helmet forward a little too much.)
The shifter position was placed such that, with my eyes closed, it was right where I expected it
to be. Interestingly, this ended up being in-line with the steering wheel.
Outward visibility is dictated by the height of the seat. If the seat is too low, we cant see the
front edge of the hood and may end up hitting cones at the autocross; too high, and the top edge of
the roof will impede visibility, requiring the driver to tip his head to see traffic lights.
The angles of the seatbelt straps are non-negotiable, as dictated by the manufacturer, so we have
to set the mounting points to suit the seat placement.
The rearview mirror placement depends on obtaining the optimum viewing angle.
There must be enough room around the pedals so your shoes dont interfere or get stuck behind
the pedals. This is a very common problem with front-engine cars, due to the intrusion of the
transmission into the foot-well area.
A detachable steering wheel may be needed in order to get in or out of the car. I normally leave
mine in place, but its very handy to be able to remove it when working in the foot well.
The steering shaft must miss your shoes when you move your foot from one pedal to the other,
and when depressing the pedals.
Using a momentary turn-signal switch will keep you from looking like an elderly driver, driving
for miles with the turn-signal lights flashing.
All of these items arent a big deal, until one of them is wrong. Its a nagging annoyance, like
having a rock in your shoe. If you arent comfortable, the car isnt fun to drive, and if on track, you
will go slow because of it.
Regarding the type of pedals to usehigh-pivot or low-pivotPaul Van Valkenburg noted that if
you chart a point on your foot as you rock it forward (pressing on the brake pedal), it better matches
the arc taken by a high-pivot setup. Yes, good old-fashioned hanging pedals better match the foots
path than the different must be better low-pivot pedals. I drew it out, and sure enough, hes correct.
(No comments on my drawing please!)
Overall chassis design
Triangulate everything
If the chassis flexes, the shocks, springs, and anti-roll bars will have little effect on how the car
handles. If the handling cant be changed, it means we are forever stuck with whatever spring rate our
flexible chassis gives us; I tried to give it my best shot the first time. Besides, if I designed a nice stiff
chassis, I could bail myself out of whatever suspension design goofs I committed.
Theoretically, the perfect chassis has each high-stress point located in three dimensions [read three
tubes 90 apart]. If an important point isnt located in three dimensions, theres going to be flex,
however small. Frequently, its not possible to truly make the tubes 90 apart. Thats fine, just do the
best you can to locate the points so nothing is bending. Some of the important points are the inboard
suspension brackets, the shock absorber brackets, and the engine and differential mounts.
Use paneling to stiffen open bays
The car must have paneling anyway to keep nasty stuff from getting inside. Have it do double duty
by stiffening the open areas of the chassis.
SCCA and tubing size
I reasoned that the SCCA knew a whole lot better than I what roll-cage tubing to use. The 1997
Rule Book said a car of my weight should use 1.5, 0.095 DOM steel tubing. Okay.
Keep feet behind the front wheel centerline
Some early F1 cars had the drivers sitting so far forward that their feet extended ahead of the front-
wheel centerline. The terrible leg injuries were enough to make F1 designersand me avoid doing
this.
Build a full-size wooden mock-up
My buddy Cecil Napoli forced me to do this, and I hated it. It seemed like a complete waste of
time; I was building a car, not a kids push-kart! That was my mindsetuntil it was done; then I
understood why it would have been impossible to build the car right the first time had I gone straight
to steel. Nothing would have fitit really needs to be laid out full-scale in some form.
Since the body shell already existed, the outside dimensions of the chassis were fixed. By this time,
all the major components were on hand: engine, wheels, tires, steering rack, fuel cell, pedal cluster,
seats, and steering wheel. While the engine, wheels, tires, and steering rack have already been
discussed, we need to cover the other major parts.
Why mess with the physical parts? Do it in CAD. Yes it could have all
been done in CAD, which would have taken a long time to learn and lots of time
to measure everything. While I could have then used CAD to move the parts
around virtually, I just used the real parts. Thats one reason why there are no
complete plans or drawings for Kimini.
Another option would have been an accurate scale model made of balsa wood.
While this works just fine for many people, I didnt do it. I didnt think it would
be accurate enough, other than to get a general idea of where things should go. To
get any more accurate than a general idea meant taking lots of measurements off
the parts and making scale-sized componentsvery similar to using CAD. No, I
stuck to working with the actual parts to save time. Of course, saving time is
relative when the project took 10 years to complete!
General Component Placement
Fuel cell
After seeing a horrific video of a car bursting into flames with the driver still inside (and not
getting out), I chose a real rubber-bagged, foam-filled, steel-boxed fuel cell. It was the second-most
expensive part in the car, but I dont regret it. Due to the size and shape of the car, not many off-the-
shelf cells fit. Holding down costs meant using an existing cell, and there was only one promising 10-
gallon unit: a 10 x 10 x 24 cell from Fuel-Safe.
Complaining about the high price of a fuel cell reminds me of someone arguing
about the expense of a parachute.
There were two places it could go: crossways behind the seats or fore/aft just ahead of the shifter
assembly. I had a concern that the CG of the car could end up too far to the rear, with the consequence
of understeer. Of course, good chassis design dictates that components should go as near the CG as
possible, especially the fuel cell. The Mini itself ended up deciding the solution; placement of all the
major parts showed that there simply wasnt room, at least for a standard-size cell, behind the seats.
The concern with it back there was that heat from the header would boil the fuel, so it was mounted
forward of the shifter.
If I were doing it again, I might consider a custom cell behind the passenger
seat, if I could afford it. I also wouldnt use the manufacturers steel box, making
my own to save money. I made my fuel cell removable from the front of the car,
which turned out to be a mistake. So much goes in front of it that if I ever have to
remove it, Im going to be very unhappy. In hindsight, it would have been best to
make it removable from below.
Seats
Seats consumed more time to select than they should have, because I wanted to sit in them first; this
is not easy, since no dealers have all models in stock. That meant driving to the few places that had
the ones I was interested in, to try them out. When my wife and I got there, the clerk was surprisingly
reluctant to take them out of the box. Well, gee, sorry for the inconvenience, but if Im spending
$1,100 for seats, I want to sit in them. I ended up buying two Cobra seats, not because I think theyre
the best, but because they actually had them and let us sit in them. I bought a Kevlar wrap-around
racing seat for the drivers side, while my wife picked her own seat for the passenger side. I think she
appreciated participating in the design process.
Mount the seats to the chassis, not to the floor panel. In an accident, as the
chassis deforms, the seat will begin moving sideways with the chassis, not sitting
there like a target, waiting to get hit all at once by the collapsing frame. Another
reason is that the floor panel is very thin and may not be strong enough to support
the weight.
Very inexpensive fiberglass seats with no padding are out there on the market; I do not recommend
them. My experience is that, with no padding, perspiration gets trapped between you and the seatit
has no way out. The friction from cornering and engine vibration on your damp skin will rub your
back raw. After my ride in a Hayabusa-powered Super-7, with fiberglass seats, I couldnt figure out
why my back hurt. Looking in the mirror showed why; I had a long red abrasion along my spine. If you
get stuck with this type of seat, consider having seat covers made.
Steering wheel
Its just a steering wheelhah! Its a very visible, tactile componenta very personal choice. It
needs to be small enough to miss your knees during shifting and braking, when your feet come up off
the pedals. Make sure the one you get mates with the steering wheel quick-release mechanism. I
bought a Momo, along with a 6-screw/3-screw steering wheel adapter from www.ioportracing.com.
Getting it all together
I built a full-size wood mock-up of the main cage, inside the shell to make absolutely sure that the
steel chassis would fit. At this stage it was easy, making sure to follow the contour of the shell so the
cage wasnt too far away, to make supporting the shell easy later on. I used the SCCA rulebook as a
guide for where the tubes should go. The roll cage was made as large in cross section as possible, for
maximum stiffness. Since the SCCA rules specified 1.5 tubing, 1.5 wooden closet dowels were
used for the mock-up.
After the basic mock-up was built inside the shell, it was moved to the chassis table and placed in
the appropriate position relative to the engine. Mock-up construction continued by first lining up the
rear tire cutouts in the shell with the axle outputs from the transmission (marked in pen on the wood).
This located the drivetrain relative to the shella major milestone. Since the alternator was the
forward-most component of the engine, it had to clear the main roll-hoop bulkhead by an inch or so,
which fixed the bulkhead relative to the engine. Placing the seats just forward of this bulkhead fixed
the steering wheel and pedal cluster. The big question was whether I would have enough legroom. I
had to fit between the engine and front suspension. Placing the pedal cluster in position and sitting in
the seat showed that there was just enough room behind the suspension arms. I guess thats what I get
for building a Mini! Note the steering rack in front of the fuel cell.
Once all major components were in place, extra wood bracing was added to complete the overall
framework. After many weeks of making sawdust, the mock-up was complete. Building it was the
only way to make sure that all components were comfortable with each other, fitting together as a
system. Cecil was happy I finally came around to his way of thinking. Building only one carand
only wanting to do it oncemeant that everything had to fit the first time. Thanks to his insistence
upon making the mock-up, everything did.
The plan was that the tubes forward and rearward of the main roll cage would be thinner 0.065
wall material. The thinking is that these areas are crush zones and should have slightly weaker tubing.
The engine support tubing also uses thinner wall tubing, the idea being that the entire drivetrain will
separate from the car in a big wreck, to absorb the energyso I dont.
The roll cage is nothing special; the forward down-tubes come down just forward of the door
openings, convenient to attach door hinges to. It then became a matter of connecting the dots,
running the chassis tubes near all the suspension links so the brackets wouldnt have far to reach.
There were the doors with their hinges and latches, mounts for the windshield, seat mounts, seatbelt
mounts, steering mounts, radiator mounts, etc. I decided early on to have a transmission tunnel
down the center; its a bit misnamed since theres no transmission in it, but it gets the idea across. It
carries everything fore and aft: coolant, fuel supply and return, and shifter and throttle cables. Another
goal was a completely flat underbelly on the car for aero downforceas optimistic as that is for a
brick-shaped Mini. This meant that the center tunnel had to be accessible from above, which was fine
with me; I didnt want to crawl under the car if I didnt have to.
At the front, the foot wells needed enough depth for pedal travel and enough width for my shoesit
was a tight fit. To the outside was the inside fender well, whose position was already dictated by the
+/-30 steering angle chosen earlier. To the inside is the center tunnel containing the fuel cell. I had to
add a recessed area so that when my foot was on the gas pedal, it would miss the brake pedal. The
transmission tunnel was made symmetrical and on the centerline of the car.
Lesson learned
In hindsight, Id make the center tunnel off-centerforget passenger comfort;
the driver comes first! Shifting the fuel cell to the right would have given more
leg room; I just didnt think of it at the time.
Another design goal was to run radiator exhaust air up and out the top of the hood. Doing so kept
air from going under the car and gave it an easy exit into a low-pressure area. Another restriction was
that the fuel cell was immediately behind the radiator; it needed to be kept out of the radiators hot
air, to keep the fuel from boiling in hot weather. It would also extend the life of the fuel cell bladder,
and besides, the radiator exit in the hood looks cool. This means the radiator has to be far forward,
and tipped forward, so exhaust air can exit upward near the leading edge of the very short Mini hood.
Thats because back near the base of the windshield is a high-pressure area that would prevent good
flow. (There are test results concerning this later in the Cooling System chapter.)
At the rear, things were a little trickier. Here, the engine was surrounded by many sub-assemblies,
but it had to be removable for service. The mounting of the engine is why many chassis designs are
not very stiff around the engine bay; tubes cant block the engines removal or installation. Most
designers seem stuck on the idea that the engine must come out the top, so they leave a large open bay,
weakening the chassis. I didnt want it to come out from above for a different reasonIm lazy. The
answer was to make the engine removable from below and behind. A dolly is placed under the
engine, which is then bolted to the dolly, unbolted from the chassis, and rolled out from behind. A
bolt-in engine tray made this approach work well. This way, I didnt need an engine hoist.
Jumping ahead for a moment, here is what the engine tray looks like, sitting on the floor of the
garage. It bolts to the chassis along the bottom main hoop and to the left and right of the engine. Being
removable, it carries only one engine mount; the other three support the engine while a dolly is rolled
under the engine for removal or installation.
The area above the engine was now free of constraints; it could be designed for pure strength and
rigidity. The rear end of the chassis terminates at the rear shock mounts, with the same points being
used as attachment points for the under-tray. The other attachment points are along the lower rear
bulkhead floor tube. The engine tray has an aluminum panel riveted to it, providing enormous
resistance to horizontal movement, and the rear mounts keep it from moving vertically. The engine
tray tubes are positioned near where the rear inboard suspension pickup points go, solving how those
would mount.
Months later, the wood mock-up was completed. Note the single diagonal members in the doorway;
the second diagonal tube was added later in steel. It was left out here to make access easier.
Measurements were taken, and then the mock-up was disassembled for firewood. This freed up
space in the other half of the garage, so my wife could once again park her car inside. Its important to
keep our wives happy.
So where is the chassis design?
We just did it! All the months of picking suspension pickup points, measuring where the shocks
will mount, determining suspension travel and the locations of seat, seatbelt, and steering rack mounts
its all done. Actual construction of the chassis just means connecting them all together.
20: Materials, Fasteners, Paint
Steel tubing
Diameter
I used SCCA requirements for the roll-cage tubing size. The bottom outside main chassis rails are
the same size as the roll-cage tubing, 1.5 OD, 0.095 square tubing (so the floorpan was easy to
attach). For the rest of the tubes, the choice was driven by a Strength of Materials table in a
mechanical engineering textbook. Tubes, ahead and behind the roll cage, have the same 1.5 OD but
have a thinner 0.065 wall (so they crush first in an accident). Remember that a large-diameter thin-
wall tube is stiffer and lighter than a small-diameter thick-wall tube.
About metal in general, always use cold-rolled! Hot-rolled is just that, and it
has a very stubborn burnt crusty skin on it that has to be removed with a grinder
sandpaper isnt good enough. If its not ground off, itll bubble up contaminants
when welded. Cold-rolled steel is far more pleasant to work with and is of
similar cost.
Round versus square
This is another endless argument; short story: it is personal preference. Square tubing is easy to
attach panels to, because theres always a flat surface for rivet holes (or adhesive). Square tubing is
also stronger diagonally , and some people say its easier to cut by hand.
I agree that a single cut in a square tube is easier, but things get trickier when
handling a two-dimensional cut, where the cut is angled in two axes. Cutting a
square tube for a three-dimensional intersection is very difficult! For round tubes,
I used a tubing notcherthe type that uses a hole saw.
Round tubing is equally strong in all directions. A 1.0 square tube has a 4.0 circumference while
a 1.0 round tube is 3.14 around, so the round tube is 22% lighter for the same size and is easy to cut
with a tubing notcher. I think that a chassis made from round tubing is more attractive, but how much
of it will be visible after its done? I used a mixture of both: square tubing where panels attached, and
round everywhere else. The real answer is, use whatever you want; it doesnt really matter.
I drilled rivet holes in the roll cage for some panels but could never determine
if that was okay to do. I was told it was okay, but still. If the cage ever fails, itll
no doubt be right along a row of holes. If you dont want to drill holes in the
cage, it means welding on long strips of metal, then drilling the strips. I didnt
want to do that.
EW-tubing
The most common tubing is EW (electrically welded). EW tubing is made by bending a flat strip of
metal into either a circle or rectangle, and then welding the seam. Its easy to tell if its EW tubing,
because itll have a weld seam down one side. Be aware of the seam when placing the tube; try to
keep the seam away from rivet holes. Its not a disaster if you forget, but since the weld joint is
thicker than the rest of the tube, correct-length rivets may end up a bit short if they pass through the
thicker weld seam. And finally, the weld seam is harder, due to localized heat-treatbeing heated
red hot and cooled, its harder to drill through.
I had no problem with new EW tubing, but later, I bought some 2 x 3 surplus material for the
trailer and got a surprise. The seam at the left side of the picture is properly welded; the seam at
center and at right is not! Its as if the seam welder got turned off, explaining why the metal was
surplus. The point is, carefully examine any material youre buying, same as if youre buying wood at
Home Depot. Check each piece for bends, rust, twists, holesand split seams!
DOM tubing
DOM (Drawn-Over-Mandrel) tubing is very nice material. It is very smooth inside and out, with no
seam to mess with drill bits, and dimensional control is very tight. The downside is price, 35 times
as expensive as EW tubing, and that it#8217;s not as widely available. My roll cage is made of DOM
its the only material the roll-cage manufacturer (Autopower) uses. I used EW tubing everywhere
else for cost and availability reasons.
Steel? Bah, Ill use aluminum; its much lighter .
Short answer: A very bad idea.
Long answer:
First off, the SCCA does not allow aluminum rollover structures, which says
something. Second, it is aluminums very nature that makes it unsuitable for car
structures, differing from steel in a very important way. A piece of steel can be
bent slightly back and forth, literally millions of times, and never break. A piece
of aluminum bent back and forth (even a tiny amount from vibration) work-
hardens. That means that the longer it is flexed, the harder it gets. Steel can be
bent a certain amount and never fatigue, but if aluminum is bent any amount, it
will harden and eventually crack.
So why are airplanes and race cars made from aluminum?
Because the engineers know how many cycles the chassis or fuselage can last
before it needs replacing; we amateur builders dont have any idea.
But what about some cars with aluminum chassis or suspensions?
The automotive engineers calculated that they will last the life of the car
(200,000 miles). Note that these applications typically do not use welding,
either, and if they do, it is heat-treated. We amateur builders have no idea what
were doing.
Okay, Ill use aluminum only in places where it isnt structural.
The problem is, any vibration work-hardens aluminum. Even a non-structural
panel will crack along the rivet holes given enough time.
Fine, Ill use thicker aluminum so it wont crack.
Aluminum will always crack; the only question is when. Once it gets thick
enough to not crack for a really long time, the weight savings goes away.
Welding an aluminum chassis will be fine.
When aluminum is heated, it becomes dead-soft. In fact, this is how aluminum
is prepared for forming; its called annealing. Welding an aluminum chassis will
result in a very soft, flexible joint at every weld point. The chassis must be heat-
treatedwhos going to do that, and to what specification?
Some people insist on making an aluminum chassis anyway, yet the facts are in
the books; if we ignore what the books say, okay. Its like running out onto thin
ice, vaguely knowing something might happen, but simply not worrying about it.
In the early days of aviation, the first British commercial airliner, the Comet,
experienced several mysterious crashes. Suspicion eventually fell on the
aluminum fuselage. It was tested by pressurizing the chassis over and over (with
water!) to simulate pressure changes during takeoff and landing. It revealed that
the aluminum became brittle over time (at a corner of a window), after being
exercised by the pressure changes. Metal fatigue is vell understoodby
aeronautical and automotive engineers, not by us.
Even more insidious is how metal fatigue in aluminum is cumulative. That is,
flexing a part a million times may show no problem, but the smart engineer
demands that it be discarded anywaybecause its life is nearly up. Problem is,
later on, someone buys this same part in a surplus yard and uses it in a new
application (perhaps even testing it and finding no stress cracks). The trouble is
that the second buyer forgets where it came from, and in his mind, the part is
new. While no cracks are found, it only means it is okay right now.
Aluminum remembers, and with its nearly expired fatigue life, the presumed
safety net simply isnt there. The buyer assumes this new part is good for a
million cycles because he thought it was new when purchasedwhich can be
catastrophic. It pays to know where aluminum that will be used in a critical
application came from.
In the interest of full disclosure, I did use one aluminum panel structurally, the
panel beneath the engine, preventing the rear suspension from moving laterally. It
will need periodic inspections to watch for cracking. So yes, I too think itll be
fine, but will keep an eye on it.
Paneling
Paneling refers to flat materials used in the build: Lexan, steel, stainless, and aluminum.
Lexan
Scratch-resistant Lexan was used for door windows, rear side windows, the back window, and the
engine bulkhead window, between the engine compartment and the passenger compartment. Benefits
are that it is lightweight and dampens sound, plus theres no worry about getting shredded by broken
glass. The material was purchased from McMaster.com, though I hear some Home Depots also carry
it. The stuff is really tougha sheet can be bent completely over on itself and will not crack; if it
does, its not Lexan. The material is soft, and because its not real stiff, 3/16 is used instead of 1/8.
A stock Mk1 Mini windshield was used. Id love to have used Lexan, because the windshield is
high up and contributes to raising the CG of the car. Its also likely to get cracked at a track event, so
why use glass? Its expensive and heavy but has one big advantage: durability to cleaning. Itll get
cleaned often, too, because I dont like a dirty windshield. Just a few cleanings with a paper towel
can leave a Lexan windshield hazy.
The Lexan was cut with a variable-speed saber saw, which worked great. Contrary to what some
people say, Lexan will not crack. Drilling near an edge doesnt cause any trouble either. The material
is, for want of a better word, soft , like plastic, while Plexiglas is brittle like glass. Some people
claim it does crack when bent, but if it does, they were sold Plexiglas. Do not use Plexiglas; it can
shatter, leaving big sharp shards like glass.
Lesson learned
Even though the Lexan used is scratch-resistant, using a simple paper towel is
enough to scratch it; I havent yet found the right way to clean it. Right now Im
using microfiber towelsthe softest I could find. Being Lexan, and perhaps
because of the scratch-resistant coating, it doesnt clean like glass. The water
beads up differently, and its hard to get it clean without leaving streaks and a
haze. The jurys still out on how smart the decision was to use the stuff. Its
expensive, tooaround $250 to do the entire car. For now Im treating it as an
expensive consumable.
Steel
Ordinary cold-rolled 1020 steel is used everywhere, except for the roll cage, which is DOM. Some
people seem to have fallen in love with 4130, thinking its some sort of miracle metal. Unless its
properly heat-treated, it is no better than 1040so says Carroll Smith.
Stainless
Thin (0.032) 304 was used anywhere impact from rocks or sand was expected, such as the floor
pan. While Id expect it to machine the same as mild steel, I got cuts over my hands only when I
handled the stainless. Its also hard on cutting tools. (Tip: do not let stainless get hot when cutting or
drilling. It will quickly get very hard, so hard that the cutter or drill bit wont be able to finish the
job.)
If I were building another car, aluminum would be used for the floor. But if I
ever run over a piece of road debris that tries coming up through the floor, Ill be
happy I stayed with stainless!
A hand punch is very helpful for punching rivet holes; it produces much cleaner holes than a drill
and doesnt leave nasty sharp metal cuttings all over the place.
Aluminum
Flat panels are 0.050 6061, though 2024 T3 could have been used. Panels requiring bending were
0.050 5052 H14 when I could find it, or annealed 6061 when I couldnt.
Annealing aluminum
Ill illustrate how one panel was formed. Any panel needing more than slight forming was annealed
(softened) using the Sharpie pen method. The panel is marked with a Sharpie pen and the panel
heated with a propane torch (I used MAPP gas since it gets much hotter).
Be aware that MAPP gas gets really hothot enough to melt a hole right
through an aluminum panel if not careful. Never mind how I know.
When the black mark disappears, presto, the metal is annealed. The metal is now very soft; it is
amazing how easily it can be worked in this state. As the metal is worked, it begins hardening again
but can simply be re-annealed as necessary. I didnt bother having the finished panels heat-treated;
normal vibration causes them to harden on their own. Of course, that means, given enough time, all the
aluminum panels will crack. Thats why an aluminum suspension or chassis should never be
considered without serious expert guidance.
As an example of working aluminum, the bulkhead behind the drivers seat needed two recesses for
the drivers seatback. The panel is 6061, normally too stiff to do much bending, but no problem if it is
annealed first. Two pattern pieces were made from hardwood, in the shape of the desired depression.
With one of the forms behind the panel and the other in front, the two are clamped tightly together.
A hardwood dowel is used to hammer the aluminum down into the wood form below. A matching
form above is needed to prevent the material from deforming upward. The material can be worked
only so much before it hardens, but it can be annealed again and again. Theres no need to anneal the
entire panel; just heat the part of the panel needing softening.
Heres the panel in place behind the seats. The depressions are there to clear two of the seat ribs,
which had been preventing it from butting up against the bulkhead. The small bolt hole at the top of the
upper recess is for bolting the seatback to the roll-cage cross-tube. Doing so made the seat much more
rigid, like it was truly part of the car, not a flimsy added-on part. A lot more time could have been
spent making the hammer-formed depressions really smooth and pretty, but no one will see them once
the seat is installed.
Jumping ahead a bit, this is what is stuffed inside the engine bulkheadfiberglass home wall
insulation. Its great stuff: it doesnt weigh anything; it costs nothing; it keeps the passenger
compartment cool; but most importantly, it is quiet and prevents oil canning.
The dash panel was fabricated from 2024 and not annealed. It was bent over my knee first, then
around a 6 PVC pipe (left over from a Koi pond filter project, but thats another story). Between the
pipe and a big rubber hammer, it turned out pretty well. Here, the rivet holes are being drilled along
the top and bottom edges. Clecos are wonderful and I highly recommend them. In fact, theres really
no way to rivet anything togetherand have it turn out nicewithout them; theyre that useful.
Below are all the panels; theres an entire car sitting heremore or less. This pile of aluminum
panels represents hundreds of hours of labor. Its all heading out to the powder-coater to be painted
black, both for looks and corrosion protection. The square hole in the large upright panel is the oil
drain access, so the engine tray doesnt require removal. At this point Id forgotten to add another
cutout for the transmission, which was added later. The dash and center tunnel covers were coated
with black wrinkle; they turned out really nice.
Lest you think an expensive sheet metal brake was used for bending the panelshere it is. This
tool was used to bend virtually every panel on the car. It consists of a scrap piece of I-beam, two
clamps, a backup bar, a rubber hammer, and thats it. For panels that needed to be pretty, heavy paper
was placed above and below the panel to prevent scarring. The shot-filled rubber mallet was my best
friend throughout the project. This particular sheet is a part of the radiator inlet duct.
Fasteners
My advice is to buy Carroll Smiths books and read themseveral times. They cover fasteners so
completely that theres little for me to add, except for what I used.
Bolts
Use AN fasteners! Bolts from Home Depot or Lowes have their placein mock-ups during
construction, perhaps (only because theyre convenient and cheap), but not in the final build! Thats
because they have very poor quality control, wide tolerances, and no traceability. Most annoying is
the poor dimensional control; used as suspension pivots, rattles are guaranteed. To prove the point,
place a rod-end onto a Home Depot bolt to see how poor the fit is. Now multiply that play by
the number of rod-ends in the carits pretty bad.
AN means Army/Navy. It is a spec that was created around World War II
when manufacturers agreed on a common hardware part numbering scheme.
It was disappointing to read in the forums that some people believe that
tightening down bolts real tight means it doesnt matter if theyre undersized.
The same reasoning leads to the belief that it doesnt matter if the thread goes the
full length of the bolt. The logic is, since the bolt isnt moving, the thread cant
act as a saw and start cutting through the bracket.
All it takes is hitting one pothole to start the undersized bolt moving around,
and thats the beginning of the end for the bolt and the bracket it passes through.
AN-bolts have a much tighter fitalways use them. They have a consistent part number scheme,
too: diameter is measured in 1/16 increments. An AN-8 is 8/16, or . An AN-3 fitting (be it a bolt
or hose fitting) is 3/16. It makes it easy to know what size the part is without having to measure it.
Rod-ends
Only use Teflon-lined bearingsthey have virtually no play; plain metal ones will rattle. The
small amount of play in each rod-end can translate to a lot of slop at the outside ends of the
suspension arms. As with the bolts, multiply by the number of rod-ends to get an idea how itll result
in a noisy, clunky, inexact suspension if low-quality parts are used. They cost a bit more and do wear
out, but theyre worth it.
This is the right way to mount rod-ends. Always use washers on each end of the boltit shows you
care. Spacers keep the suspension pivot point (that you worked so hard to accurately place) from
sliding back and forth along the bolt. The spacers allow the bolt to be fully tightened without bending
the mounting brackets in toward each other, and the Nylock nut will not vibrate loose. Always ream
holes to final size; drills are notorious for creating oversize holes that are not round! Always mount
brackets in double-shear. And finally, do not allow threads to pass through the brackets; they will act
as a slow cutting saw.
In certain applications, its wise to use left- and right-hand rod-ends. Why?
Consider the front steering tie-rod links. Imagine that the front toe needs
adjusting. If the link has two right-hand (or two left-hand) rod-ends in it, turning
the link will screw it further into one rod-end and unscrew it from the other! In
other words, you can turn it all day and the overall length wont change. Using
one left-hand and one right-hand rod-end is the right way to do it. Now, turning
the link either screws both of them in, or both out, at the same time. Carefully
choosing where to use left-hand rod-ends makes suspension adjustments much
easier.
And dont forget to order lock nuts (left- and right-hand) for all rod-ends!
Small fasteners
I used stainless hardware everywhere; it looks nice, doesnt rust, and wont seize up in a few
years. Small stainless flat-head screws were used as mounting studs, be it for cable tie-downs,
subassembly mounts, or whatever. They were tack-welded on upside down, with the head flat against
a chassis tubecheap, easy, light.
My brothers Super-Stalker had instructions saying to fasten the floor pan on
with sheet-metal screws. This seems like a really bad idea, since they are
notorious for loosening upunless they are dipped in adhesive prior to
installation.
I used scratch-resistant Lexan for the side and rear windows, but I knew theyd have to come out
eventually for replacement. This meant using fasteners instead of rivets, but what to use? Turns out
there are some really nice flat Allen-head screws that look just like rivets, unless examined really
closely. Each stainless screw uses a nylon washer and matching nylon locknut; the entire car needed
about 100 of each.
Buying small hardware in quantities smaller than 100 doesnt save any money; buy in bulk from
someplace like McMaster.com. It costs little more than buying just a few and guarantees you wont
run out in the middle of the project. Get several hundred of each length of screw, and dont forget nuts
and flat washers, too.
Bushings and bearings
For a home-built car, bearings are usually used only in the wheel hubs; all other rotating pivot
points use either rod-ends or bushings shouldered plastic, rubber, or oil-impregnated brass. Check
McMaster.com or the circle-track suppliers listed in Appendix B. Bushings are very reasonably
priced since they are manufactured in such high quantities. For true ball bearings, I suggest searching
for ball bearing with Google. Bearing manufacturers have very impressive catalogs with tons of
technical information to help select the correct part. (Note that bearings have a universal part
numbering system, so you can get just about any bearing from any bearing supplier.)
Rivets and adhesive
All permanently mounted metal panels had a bead of silicon run along the joints and were attached
with either Cherry Q or CherryMAX rivets.
The silicon keeps the panels from rattling, seals against water intrusion, and
gives an overall clean appearancelike someone cared.
Cherry rivets are amazing fasteners, having nearly the strength of solid rivets, yet with the
convenience of pop rivets. The most important aspect, though, is that they are structural ; the mandrel
(removed from most rivets after installation) is retained. This leaves a pin passing through the
clamping point, making the connection much stronger. The downsides are price and availability; they
cost around 2030 times that of solid rivets. After a long search, several suppliers were found that
had stock at reasonable prices; they are listed in Appendix B.
Lesson learned
There are approximately 2500 0.156 diameter rivets in the car! Cherry Q or
CherryMAX rivets are already hard enough to find, and picking an odd size of
5/32 made it that much worse. In hindsight, I would have used either 0.125 or
0.187. Why? Theyre much easier to find in stock. (Rivet sources are listed in
Appendix B.)
The reason I chose 5/32 is because a designer whom I respect told me to.
What happened was that he was giving his engineering answer to the problem,
not the practical solution. That is, the rivet size he recommended was correct,
from a strength standpoint, but if no one carries them, whats the point? The
lesson is, if calculations show that a certain size part is required, make sure that
its available before designing it in; otherwise, use a more available component.
An emerging alternative to rivets is using industrial adhesives; there are some pretty amazing
products on the market3M VHB tape, for one. The stuff is strong, so strong that it can be used
instead of rivetsalmost. The problem, in a car at least, is that various liquids, such as oil, gasoline,
antifreeze, and brake fluid can soften the adhesive. The safest thing to do is to install some rivets at
the corners of the panels to prevent them from peeling off. Its good insurance, especially for the
floor, where all the spilled liquids end up; plus, they are the panels most likely to be ripped off in an
accident. We dont want our feet hanging out the bottom of the car!
As Ive stated elsewhere, do not use liquid metal type adhesives in place of
welding! Somehow, that industrys marketing weasels have convinced some
people that it is just as strong; it simply isnt so. Maybe someday, but not yet!
Imagine what car inspectors would say at the track.
Removable panels
Rivnuts (from Aircraft Spruce) were used to attach any panel that had to be removable. The Rivnut
installs in a hole just like a rivet, but when the threaded installation mandrel is unscrewed, it leaves a
threaded hole. These were used on the center tunnel covers and some radiator ducting.
Thread locking
Use nylon lock nuts on everything; vibration guarantees any fastener not locked down will come
loose. This type of lock nut is available in sizes all the way down to #2, so theres no excuse not to
use them everywhere.
Lesson learned
In the few places where I didnt use nylon lock nuts (I didnt have any at the
time), the fasteners did indeed loosen.
Locktite has some nice thread-locking compoundif used correctly. Its great for locking parts
where you cant get at the backside, or where nuts cant be used, like the screws in the terminal
blocks on the engine computer board. For situations where locking fasteners and Locktite isnt
appropriate, theres always safety wire. At above-right, the threaded ends of the screws retaining the
muffler baffles arent accessible; plus theyre subjected to high heat, which would fry Locktite. Old-
fashioned safety wire comes to the rescue; its about the only thing that works here. It gives absolute
confirmation that the parts are locked and havent backed out; plus, I think its attractive. (Note how
the wire is wrapped, tending to keep tension in the tightening, not loosening, direction.) Safety wire
and safety-wire pliers are available from aircraft supply companiesmine came from Aircraft
Spruce.
Safety wire really is needed here, too. For some strange reason the muffler
came with only three of the six screws having drilled heads; guess which ones
backed out? Shorter screws were ordered, all with drilled heads. Being shorter,
less thread was exposed to exhaust gas on the backside, which made them easier
to remove. Also, always use high-temperature anti-seize on the threads. I always
did, and even then, on two occasions a screw got jammed so tight that it broke
when removed.
Heres a specialized thread-locking device very similar to a safety pin. Its perfect for acting as
a cotter pinwhere a cotter pin wont fit. I used them on all the taper-pin castle nuts. Nylon lock nuts
do not work here because they can spin the taper pin instead of threading onto it.
At right is the ubiquitous Dzus fastener. The spring on the left is riveted to the fixed panel in my
case, the chassis. The screw-like portion is riveted to the removable componentthe nose and tail
sections.
The large screw-like portion of the fastener has an over-center slot machined into its shaft. As
the screw portion is pushed down, it engages the spring in its slot and then is turned 90, latching the
spring into the slot, applying tension on the spring, and locking the fastener in place.
Some people say they are a real pain to install. Thats because the screw portion comes in about
six different heights, as do the springs. Choosing the right length fastener and proper spring
combination can be a pain. I took the easy way out and ordered a bunch of each size. I first installed
them without rivets, swapping parts around until the reach and tension was just right. This was nearly
the only way to do it, since the composite thickness in Kiminis shell varied quite a bit. Often, two
adjacent fasteners ended up different sizes, due to local variation in the composite thickness.
There are many different types of fasteners that perform the same quick-release function. I like the
Dzus fasteners because they are simple and have that 1960s look.
Plumbing
I used Earls plumbing components, and frankly, I was disappointed. Twenty years ago, they were
really high qualitynot anymore. Perhaps they no longer manufacture their own products; perhaps its
because theyve moved production offshore. But regardless, quality isnt what it used to be. For some
parts, this can work, but not in this case; its a risky decision. Once customers become unhappy,
theyll switch to a competitor: Parker, for example. Worse, these lost customers will tell others of
their experience. I hope Earls realizes the error of their ways; otherwise they are destined to go out
of business, probably never understanding why.
Paint
I went round and round on what the best coating was for the chassis and suspension. Choices were
to leave them bare or paint, anodize, plate, or powder-coat them.
Leave them bare
Cheap, light, and simple is a hard combination to argue with. Carroll Smith said he was shocked by
how much weight paint added to a race car except I wasnt building a real race car. Upkeep was a
big issue, and a dune buggy owner said it was a mistake to leave the panels bare. Bare aluminum
corrodes very quickly, with the panels getting covered with an ugly whitish powder. Yes, aluminum
can be polished to a high shine, but its laborious, and its impossible to get into every corner. Plus,
having mirror-like panels inside the car is bad news for glare.
Ive seen cars where owners have painted the instrument panel light colors
even white. The suns reflection off the inside of the windshield must drive them
nuts, but as long as the driver thinks hes coolwhatever.
Spray or brush-on paint
Painting a tubular three-dimensional object can be tough; making sure every side of every tube has
even coverage is trickysurprisingly difficult to ensure. While the pros have the technique down, I
didnt. Being a cheapskate by nature, I wondered if brushing the paint on was a viable alternative to
spraying: it would get around the issue of even coverage; it would be cheaper because I could do it
myself; the garage wouldnt have to be converted into a paint booth; and touch-up would be easy.
There are several paints claiming to encapsulate rust: Rust Bullet, POR-15, Chassis-Saver, and
Glisten-PC. I bought a small bottle of Rust Bullet to test, and while it worked as advertised, it left
very visible brush marks. Perhaps if Id tried a higher-quality brush it would have turned out better,
but the brush streaks made me lose interestit looked really bad. Later, I tried POR-15, a competitor
to Rust Bullet. Unlike Rust Bullet, its a four-stage process, so it requires a lot more time. The POR-
15 went on okay, maybe even a bit smoother than Rush Bullet. The problem was with the required
topcoat; brush strokes were clearly visible.
If it had been sprayed on, it may have been perfectly acceptable. I just didnt want to deal with the
setup and mess. It put me back where I was before, unable to guarantee good coverage, and not
wanting to build a spray booth in the garage. Because of how the brushing turned out, I didnt go with
that approach.
Regardless of the type of paint you use, wear gloves! Some of the new coatings
cure from moisture in the airor from your skin. Once it gets on your hands, it
will not come off until your skin wears off. Really.
The picture above shows my environmental test sample. The silver-colored portion of the tube
was painted with Rust Bullet, and the other end was left bare. It was left out in the elements for more
than two yearsit held up fine. The ding in the corner is where I hit it with a wrench to see how well
the paint would hold up to impact. It lifted the paint in that spot, leaving some surface rust. Overall,
its good stuff, but still visible are the brush marks. Whether this is a concern is up to you.
The above brush method is offered as an alternative to spraying the steel
chassis and panels. As for painting the shell , I didnt want to try it myself; there
was so much work, time, and money in the car, I didnt want a crummy paint job.
The shell was taken to a real paint-and-body shop and shot with Sherwin
Williams paint, single-stage Volkswagen metallic Reflex Silver, paint code
LA7W.
Chassis color
This is up to you! I learned a lesson on my old Datsun about using black paint on the roll cage. The
problem is that it sucks the light right out of the carnot everywhere, of course, but down in the foot
wells, where its hard to see to begin with. The lesson learned, I chose a much lighter color this time.
Of course, you can go too farusing too light a color. This happens when going for looks over
function. Using neon green on the roll cage will get a lot of attentioneven out of the corner of your
eye while driving. Its simply too bright and draws the eye to itnot whats wanted while driving.
Ive seen comments from a handful of owners, wishing theyd painted their roll cages a less irritating
color. I appreciated the honest comments; rather than insisting it was great, they dropped their egos
and admitted it wasnt a good idea. Given that, I decided to use navy gray and have no regrets.
Shell color
Several factors helped me decide this. First is that I had a composite body; since I didnt use high-
temperature-cured epoxy or an autoclave cure, there was a real risk of it softening in the sun. Even a
white surface will reach 160F in the desert sun, and it only gets worse with darker colors. The next
factor was how, in my opinion, the perfect color on one car looks totally wrong on another. I had a
1969 Datsun Roadster, and I felt its perfect color was fire-engine red. Once painted, it indeed was
perfect. But I would never paint Kimini red. Why? Its very smalltoy-like in fact. The last thing I
want is a car that looks like a toy, which rules out strong colors like red, blue, or green. No, it needed
to look all business, and thats exactly what metallic silver does. But the most important quality of
matching paint to a car is the cars shape, and Kimini is all curves; metallic colors are excellent for
enhancing the contours of a curvy body. So, between needing a light color to keep heat down, metallic
to show the curves, and non-colored paint looking very business-like, the final color pretty much
picked itself. Metallic gray also complements the gray chassis, black panels, stainless floor, and red
suspension.
So which of the thousands of shades of silver is the right one? I wanted to be able to go into a paint
shop in five years and know theyd still have the same exact match; turns out thats a non-issue. These
days, paint shops can color-match any color: they just type the paint code into their computers, and out
pops the magic mix. So, it doesnt matter what color I liked, as long as I could supply the painters
with the magic paint number.
Brake caliper paint
There are reasonably priced kits available for painting brake calipers in just about any color
imaginable. While it doesnt do anything for performance, having clean, shiny brake parts goes a long
way toward having a really nice-looking car thats pleasant to work on. Being able to work on the car
without getting your hands black with grease is a worthy goal and shows you care. If you dont care
for the me-too thing of bright red calipers, theres also black, silver, blue, green, etc.
Anodize
Anodized aluminum panels look really niceif done right. It means having all the panels sanded,
preferably in one direction, or polished, then having them all anodized at the same time to ensure the
same shading (and to keep the price down). One problem is that if you use different types of
aluminum, like I did, each type will end up a slightly different shade. I also learned that there are
different types of anodizing. For example, take black anodizing, which I found to be very susceptible
to ultraviolet light, fading very quickly. Simply setting parts out under fluorescent lights for several
weeks is enough to cause an obvious color shift.
As an example, my wheels have anodized centers. They were kept in the garage during the build, in
black garbage bags, but not always. This particular wheel ended up on top when they were stacked
out of the way. The discoloration is due to UV exposure from the overhead fluorescent lights and UV
from the welder. The lesson is to request UV-resistant anodizing. If these wheels ever need to be
disassembled, the centers will be powder-coated black.
Plating
I dont know enough about plating to offer advice. I was scared away from it because Id heard
stories about hydrogen embrittlement or internal corrosion due to parts being dipped in chemical
baths. In the past, Ive had parts nickel-plated, and while the parts came out really nice, the parts
werent structural. For suspension components, I didnt want to risk it.
Powder-coat
There were two places that did powder-coating in town. One place acted as if there was only one
type of powder, giving me the corporate blank stare when asked if there were other types of
powders available. Coincidentally, an unhappy customer was complaining about chipped paint on his
dune buggy when I walked in. The second place had a completely different attitude: when asked about
different coatings, they produced a brochure clearly explaining the different qualities. Powder-coating
seems to have gotten a bad rap as being brittle hogwash. Its because the wrong type is being used.
It seemed like the perfect coating to me, so I chose powder-coating for the chassis, suspension, and
all aluminum panels. Its not the cheapest or the lightest, but it looks great and requires little
maintenance. (Theres also something to be said for dropping the parts off and picking them up a few
days latercleaned, painted, and wrapped.)
The panels are prepared by the vendor for paint by media blasting, something like sandblasting, but
using plastic grit instead of sand. Ive heard that if very thin panels (20-gauge) are blasted, the
blasting can distort the aluminum slightly. Fortunately, all my panels were 18-gauge, so it wasnt an
issue. The aluminum panels were painted with three different types of black. A polyester wrinkle
finish was picked for the dash and the top of the center tunnel to minimize reflections. Polyester was
picked due to its UV resistance. The wrinkle finish came out really nice!
The external door hinges would be in direct sun, so a semi-gloss TGIC polyester was chosen. All
the internal panels were covered with low-gloss epoxy; since theyre out of direct sunlight, fading
wasnt as much of an issue.
Suspension arms were coated in red epoxy. Epoxy is the most impact resistant. This is important,
because the arms are in the line of fire from rocks and sand. Epoxy powder-coat is pretty much
perfect, except for its poor UV resistance, so it cant be used where the sun hits it. The red color was
just vanity, something to nicely contrast with the black aluminum panels, stainless floor, and gray
chassis.
I am well aware of the risk posed by powder-coating the suspensionthat
cracks may be hidden under the flexible paint. On the other hand, plating has its
own demonsspecifically, corroding from the inside out. This seemed worse
than surface cracks; if it is powder-coated, the crack might be spotted, but if the
crack is on the inside, it most certainly wont be.
The chassis was coated with epoxy, since almost all of it would be out of direct sunlight. Even
areas exposed to light through the windows would be okay, since both glass and plastic absorb UV.
Besides, being navy gray, what color could it fade to?
Lastly, the stainless floor was left bare; itll take a beating from rocks, sand, and the occasional
speed bump. If glare is an issue, some trunk carpet will be added.
A year after the build, Im still very happy with the powder-coat. It has held up
extremely well, putting up with everything I threw at it literally and
figuratively. Id definitely use it again.
Regardless if powder-coated, brushed, or sprayed, it is important to keep paint out of all holes,
bores, and threads. The picture here shows a hole I forgot to mask off. This is the top tapered ball-
joint mount on the rear suspension upright. I forgot to mask it for the paint shop, they didnt catch it,
and I missed it during final assembly. The end result was that the taper pin of the ball joint didnt fully
seat. The thickness of the paint caused the ball joint to stick up about 1/8 higher than it should. This
caused the ball joint to sometimes scrape the inside corner of the wheelthats how close I cut
things. This is a good example of how vigilance is always needed!
Lesson learned
I had powder-coated the uprights, along with the brake calipers. I knew to
keep the paint out of the mounting holes, so there was no problem mounting the
calipers. Later, after the first trackday, I was shocked to find that all four (two
per side) brake caliper mounting bolts were loosevery loose. What had
happened was that I had painted the faces of the mounting surfaces. When the
caliper was tightened down, there were a total of four layers of paint holding the
caliper away from the bolt. During the trackday, the brakes got hot, hot enough to
soften the paint. The paint that was acting as a spacer between the caliper and
mounting brackets melted out from between the mounting facesinstant loose
bolts.
I discovered another lesson related to powder-coating, on a front toe control
link. The threaded holes had been masked as they should have been, so I thought I
didnt have to run a tap through the threadswrong. Apparently, blasting grit
was still in the threads, and when I screwed in a rod-end by hand, it got stuck.
Stuck was being politeit was downright frozen! The tube had to be clamped
in the lathe, the spindle locked, and a 3 ft. bar (!) run through the rod-end to work
it back and forth. I came this close to throwing the toe link and rod-end in the
trash, but I was too stubborn and cheap, instead taking 20 minutes to get it out.
The lesson is, run a tap through every threaded bore before installing anything!
The Eastwood powder-coating kit
This was purchased with the thinking that it would save money, but Im not sure I recommend it.
Theres nothing wrong with it, but it just wasnt that useful. Yes, it can be used to paint various small
parts, but you need an electric oven big enough to hang the parts in while they cure. Its great if you
have lots of small stuff to do, but not so helpful if you have larger panels to coat. The final proof was
when I took all the large panels to the powder-coating shop. If I had brought along a few more small
parts, it wouldnt have cost any more to do them also. So, I could buy the kit and spend a bunch of
time coating all the little parts myself, or take those same bits along with everything else, and have it
all done at the same time, for free. Seems like a cheaper, more efficient way of doing things to me.
Colored films
There are now plastic films that claim to be colorfast, lasting for up to 10 years and allegedly
sticking well to metal. It is applied much like a decal and is claimed to be very flexible, even able to
be installed over rivets. One type is from: grafityp.co.uk/graficast%20series.htm.
The phrase up to always bugs me. Its used in the context of, make up to
$60,000, or adds up to 50 hp. Marketing weasels count on human nature to
only hear the big number. Imagine the same engine product that guarantees to add
less than 50 hp. If the product adds only one horsepower, it has still fulfilled
its promisenice, huh?
De-rusting steel parts
Someone figured out what is in a popular steel and iron de-rusting agentmolasses!and showed
how to make your own.
Yes, thats right, molasses, like you use for cooking. Use it at roughly nine parts water to one part
molasses, and let the part soak for as long as it takes to eliminate all the rust.
To be fair to Don Kemp, the guy who figured it all out, his Web site is
http://www.wr6wr.com/newSite/articles/columns/wp0906/wp0906.html.
21: The Chassis Build
With chassis construction finally begun, it was simply a matter of sticking all the parts togetherin
a very over-simplified, understated, under-estimated, overly optimistic manner of speaking.
Using the wood mock-up as a pattern, the bottom rails were clamped down to the chassis table, and
the engine and fuel cell set in place fun stuff. Note that the fuel cell is being used as its own
template, to find where the side rails go. Also visible is the center-string, which saved me many times
from building a bent chassis. On the right is one of many rulers, which you can never have too many
of! After it was all clamped down, it was finally time to start welding!
The chassis table also provided a great place to store tubing: below it on the floor (though I
learned the hard way that, if left on the floor, moisture quickly rusts the steel).
The basic roll cage is beginning to take shape, with large C-clamps a necessity to keep things put.
A goof at this stage would forever compromise the accuracy of the chassis. As stated before, since it
was a one-off car, the actual parts were used to figure out where to run chassis tubes. The engine is
already in position, with the chassis being built up around it. The horizontal tube in the main roll hoop
is at shoulder height, in anticipation of attaching the seatbelt shoulder anchors to it.
Behind the drill press, on the wall, is a large poster from J&L Supply. It is a
drill and tap charta very handy reference and great time saver! A chassis
builder said to expect each chassis tube to take about one hour of preparation
he was right. That was good to know; it would have felt like it was taking forever
otherwise. A tip on welding tubing together: drill a small hole in each tube for
the escaping hot air; otherwise itll blow holes in the molten metal during
welding.
Regarding seatbelt anchors, the bolts should mount directly to a tube or plate, without any spacers.
The drawing above shows the wrong way (above) and the right way (below). The important thing to
note is how the spacer feeds too much twisting torque into the tube, allowing the seatbelt to, in effect,
stretch, letting your head get that much closer to the steering wheel in an accident.
At this point in the build it was great fun welding in major tubes and literally watching the
project grow up off the table.
One goof: the rails on each side of the fuel cell warped due to the welding
heat, enough that the cell wouldnt fit! I had to use a hydraulic jack to force the
tubes back apart.
Tired of welding, I welcomed the diversion of cutting and drilling the first of the stainless floor
panels. Stainless is nasty stuff, producing downright evil metal bits with diabolically sharp edges; my
hands looked like Id been playing with razor blades. Stainless offers some protection from road
debris coming through the floor. Doing it again, however, Id probably use aluminum, simply because
its lighter and easier to work with. This is the drivers side floor panel viewed from below (the
chassis has been rolled over for easier access). The stainless here is fairly thin, about 0.032an
effort to keep the weight down. For some reason, a panel full of Clecos looks really cool.
The secret to a clean line of rivets is:
1. Put the panel in place, marking where the tubes are on the back side; I use a
Sharpie pen. (Do not use a scribe on aluminum; the aluminum will crack
along the scribe line.)
2. Remove the panel and mark the centerline of all the tubes with the marker.
3. Using a ruler, mark the center point of each rivet along the centerline, being
sure to use consistent spacingbecause you can, and it looks nice.
4. Center-punch each point, then punch or drill the holes.
5. Clamp the panel back in place, on the chassis, and using the holes in the
panel, drill through the chassis tube. Be sure to install Clecos as you go,
otherwise the panel will slowly shift, with the result being that half the holes
wont line up.
At this point the chassis was still fairly light easy to move around to a more convenient working
position. The white panel at the bottom is stainless, too, with its protective sheet still in place.
Cleco fasteners were indispensable throughout the buildthere really isnt any alternative. Just
buy themyou wont regret it. Dont forget the special Cleco pliers, used to install and remove them,
because regular pliers wont work. I bought mine from Aircraft Spruce.
Heres the front of the chassis partway through the build. The fuel cell is in its central position,
with the two coolant lines temporarily in place above it. At front-right is a suspension rocker arm; the
outboard upper ball joint is a tie-rod end from a Chevy truck. The steering rack brackets have already
been attached (at the front-bottom of the fuel cell). It took a really long time to place them accurately,
to ensure there would be no bumpsteer. Its best to deal with it now instead of later.
The fuel cell slides out from the front; better would have been to have it removable from below
next time.
When one part becomes tiring to work on, theres always something else that needs attention. This
is the aluminum coolant header tank, fabricated from 3 pipe. The concave end caps were hammer-
formed from annealed aluminum and welded on. The two nipples are coolant return lines, one from
the radiator (later eliminated), and the other from the top coolant bleed on the cylinder head.
Normally, the cylinder head vent is only opened to bleed air from the cooling system when its
refilled. This setup leaves it open all the time, continuously venting air back to the header. The
radiator cap flange came from Summit Racing. The header tank is mounted so that it is both the highest
point in the cooling system and near one side of the chassis for easy access.
Dont forget to leak-test anything that will have fluids in it! I found a leak in this assembly and
several in the coolant pipes! Since leak testing was done before installation, it wasnt any problem to
fix.
The chassis is finally removed from the chassis tablepsychologically, a big step! Components
were used as their own templates wherever possible; this was much faster than modeling in CAD.
Though the basic chassis was now complete, it was far from finished. Now it was just a matter of
fabricating and attaching about a million brackets.
After the chassis was taken off the table, the table was disassembled, and leftover tubing was used
to make a three-wheel dolly. The front caster wheel is visible at the bottom of the picture. This
allowed the entire frame to be moved around for better access.
Here, the engine is shown suspended from the chassis. Below the engine is the engine tray (or here,
the lack of it). This way, the engine can be removed without lifting it out the top of the chassis. The
left, right, and front engine mounts support the engine with the tray removed. The engine tray bolts in
along the bottom of the main bulkhead. At the rear-most point on the chassis are the two large bolts
that support the back of the engine tray.
A close-up of the rear engine-tray support with the bolt backed off. The chassis (left side) has a
raised flange that the engine tray plugs into, so the junction doesnt move around absolutely
necessary because rear-wheel alignment depends on it! The flange also ensures that no shear stress is
taken up by the bolt, which simply keeps the tray from popping out. The clearance was tight enough
that it couldnt have any paint on it, so it had to be masked off.
Another view of the engine under-tray. The top of the V really needs a cross-tube, but since metal
shrinks when welded, and the alignment was so critical, I was afraid to add a tube for fear it would
never fit together again. The tubing could possibly have been braced during welding, but with a
tolerance of maybe 0.005, I didnt want to risk it.
After working on the chassis for so long, I needed a psychological boost like hearing the engine
start. For this to happen, it meant I had to confront the ECU, with its dozens of wires. For dealing with
this, a factory manual is an absolute must-haveno way around it. It took weeks to go through the
entire engine schematic, removing or keeping each wire.
Every single wire going to or from the ECU had to be examined to see where it went, what it had to
be connected to, and if it was still needed. The good news was that a great many wires were simply
no longer needed. They were cut off and saved for later, which became a great source of different
color-coded wires. Another perk was the 15 lb. of wire removed! The mess you see here is what was
leftthe bare minimum needed to run the engine. After a few weekends, all the grounds were
grounded and +12 V connections run as needed, and it was time to try to start the engine. So after
years of messing about with drawings, equations, and tedious welding, it was finally time to see if it
would run.
When the drivetrain was purchased, there was evidence that it had been hit
hard in the front right wheelone CV joint was shatteredso there was some
concern that the engine or transmission might be internally damaged
Unfortunately, because the engine was bought sight unseen, I didnt know what
had happened to it. Had I seen the damage in person, Id never have bought it, but
luckily everything turned out okay.
Finally, the big moment. First a bit of oil was squirted in each cylinder (the engine had been sitting
for about five years) and the engine cranked with the plugs out. No bad noises, no clanking sounds
good. I put the plugs back in and turned on the fuel pump. Having 37 lb. of fuel pressure in temporary
hoses made me nervous, so I kept a close eye out for leaks. Well, here we go. Setting ignition hot, the
engine was cranked againnothing. Sigh. Now what? Turned out the fuel pump wasnt primed. After
fixing that, I tried again. Id like to say it just sprang to life, and it tried, but Id forgotten to remove
the tape over the end of the rubber air intake hose. When it tried to take its first breath, all it could do
was suck the inlet hose flat! After removing the tape, I tried again, and it started! Initially it ran on
three cylinders, but after a minute it smoothed out, as all four began firing. With no water in the block,
I let it run less than a minute, but I was a happy boy; Kimini had taken her first breath!
The issue of running on three cylinders returned later. Apparently, the
electrical connector on one injector was intermittent. After the contacts were
cleaned, the problem vanished.
Tired of working for weeks on the electrical system, I turned my attention to the aluminum paneling.
In hindsight, this access hatch on the side of the center tunnel wasnt very functional. After the seat
was installed, I found it blocked the panel, making it impossible to install many of the screws (a case
of not thinking ahead). I ended up installing the screws I could reach and gluing the panel on with
silicon. The small tube sticking out on the right is a seat mount, which proved very difficult to access
once the center tunnel was finished. Also visible is a coolant line and the two (noisy) Facet fuel
pumps. They were moved above the fuel cell.
Construction paper was used everywhere as mock-up materialthe secret to
getting it right the first time. If you have a documentation department at your place
of work, this type of heavy paper is normally thrown out.
I learned the hard way about having tubes come in at angles less than 90 to a panel. It prevented a
single sheet from fitting; this one, at right, had to be cut into three pieces.
The sheet on the engine side is stainless, because stainless insulates against heat 12.5 times better
than aluminum. On the forward side of the bulkhead, shown here, aluminum is used, but it had to be
cut into three panels for the same reason. During final assembly, home wall insulation was packed
insidethe best thing I ever didand it was amazing how quiet and cool the bulkhead stayed. Of
course quiet is relative; a fair bit of noise comes through the carbon shell on the side near the
exhaust, but every little bit helps. Another perk of the insulation is that its very light and damps out
vibration, keeping the large panels from oil canning sounding like a huge drum.
The right side of the center tunnelI wish Id made the tunnel about one inch wider, for better
access, but there just wasnt room. The Mini shell dictated everything, and the small size caused the
seats to butt up against the chassis on both sides, and against each other. My notes say that there were
about six hours spent in each panel here. The fuel cell lives inside the forward end of the tunnel. The
idea was to isolate it from the passenger compartment in case of fuel leaksor worse.
Working forward, the wheel-well panels were next. One nagging issue was the rocker-arm pivots
(here, a large temporary bolt). The forward end is prevented from moving vertically by a vertical
aluminum panel and steel tubes behind it which form an X. I probably should have welded in more
steel instead of counting on the aluminum for its sheer strength. As I said before, aluminum cant be
counted on for sheer strength unless its designed for the intended load, and I had little idea how much
load gets fed into the front end. What makes the design seem safe is that the front axle centerline is
toward the rear of the pivot, so 90% of the force is fed into the rear pivot mount instead of the front.
Thats okay, since the rear pivot mount is in the foot-well panel, which is much stronger.
This picture below shows the front upright from a 1979 Nissan 280ZX. Doing it over, I might
consider uprights with bolt-on struts, which would have made it easier to adapt (or maybe not). A
bolt-on pivot, instead of the screw-in type shown here, would have made camber adjustment easier,
too. As it is, the top plug has to be removed from the upright in order to screw the tie-rod end in or
out. (In hindsight, a Mazda Miata upright would be an excellent candidate.)
A Tilton master cylinder assembly was used, simply because Id used them before. Originally, a
forward-facing unit was tried, to get the master cylinder reservoirs out from under the base of the
windshieldit wouldnt fit, so the rear-facing type would have to do. Several sets of mounting holes
were added for adjustability but werent very useful since the brake and clutch pedals would bottom
out in the foot well if it was moved any further forward. If moved backward, the outlet hoses hit the
chassisits always something. (Remote reservoirs could not be used because there was no place to
mount them higher than the master cylinders.)
I have mixed feelings about Tilton; they seem to be very high quality, yet not
one, but two, master cylinders leaked after the car sat for a while. Tilton said that
they will leak if they sit. I wonder if Wilwood master cylinders have this same
problem. (It is possible the leaks were due to using synthetic brake fluid.)
A top view of the front of the chassis (upper-right,) with suspension, fuel cell, brakes, radiator
lines, radiator, and ductwork. Next time, Id use rubber elbows for all the coolant pipe bends instead
of welded bendstheyll crack eventually. Rubber elbows also make final alignment easier when
routing the pipes and moving parts around.
This more or less finished the basic chassis, yet hundreds of hours remained to truly finish it. Id
been warned about this, that the little bits, especially brackets, take a really long time, and it would
look like nothing substantial has happened for all the time spentthis warning was very true! Well,
the little bits are just as important as the big ones, so on with the show.
During construction, full access was needed to get into all the various corners. Now that the chassis
was nearing completion, it was time to start buttoning things up by adding the final tubes. A relatively
easy addition was that of the door tubes, adding the second cross-tubes to the doorways to form an X.
The door tubes fully triangulate the doorways, making the chassis much stiffer, and they also serve as
side-impact bars. (Many people wrote, after seeing this picture on the Web site, saying they were
glad to see them finally installed.) The tubes were always part of the design but were left until last so
they werent in the way.
Next, the rear diffuser was added, the idea being that fast-moving air below the car would give
some semblance of low pressuredownforce or at least decrease lift. The main diffuser support is
the bottom center of the V, with additional tubes coming off from the sides. It may not be very elegant,
but I hope itll be functional. Anyhow, the thinking is it is better to try to make it right, rather than not
even bother. Its effects wont be clear until lift and downforce are measured at speed (which proved
difficult). With the area paneled in and angled upward about 14, data in Katz
9
says there should be a
bit of downforce. At that angle, there will be some drag, but since the car will be gear-limited
anyway, it doesnt matter. Aerodynamics is covered more completely in its own chapter.
Next was making accommodations for mounting the shell to the chassis. The most important piece
of the shell was the passenger compartment, since it was mounted first and permanently attached. A
tube was bent to closely follow the lower edge of the windshield. Its purpose is to attach the upper
dash, firewall panel, hold-down clips for the windshield, and Dzus fasteners along the front edge of
the passenger compartment.
Heres the windshield tube seen from the front, with the shell lifted out of the way. Already
attached are the small brackets for the windshield hold-down bolts and the mounting cups for the Dzus
fasteners. These will attach the nose to the passenger compartment and are mounted to the curved dash
panel at the base of the windshield.
From the inside again, this is how the curved upper-dash panel mounts. The windshield hold-
downs (not visible) will also serve to keep the carbon shell down against the front edge of the panel
here. Once the shell is mounted, a bead of silicon sealer is run along the edges to keep out dust,
fumes, and noise.
Here are the doorsthe evil, nasty doors. The doors and hinges were the least enjoyable part of
the entire build. To get the door to fit just right meant everything had to be near perfect in alignment.
Worse, factory Mini doors have a surprisingly big tolerancea door thats a perfect fit on one car
may not fit another identical Mini. A fellow wrote me who had worked in the Austin factory, on the
assembly line, specifically in the Rectification Department (his words). He said it was common
practice to get Mini doors to fit with large rubber hammers and wooden beams! This pretty much
guaranteed that my two doors were different and unlikely to fit as is.
These pictures sum up many hours of getting the door into alignment with the shell, and then
carefully welding mounts to the forward roll-cage down-tubes. The factory Mini external door hinges,
bolted to the door, were retained, but the mate, bolting to the original front fender, had to be replaced.
Since the front fenders (and nose) are removable, there cant be anything bolted to them. Also, note
the welded hinge mounts on the cage, how they end just short of the shell; had they extended through,
the shell couldnt be installed!
The mounting points are simple square tubing but are large enough so that spacers can be slid in
underneath the hinge for vertical alignment of the doornot glamorous, but functional. After the
hinges, everything else seemed easier by comparison. Here, cardboard is tucked in behind the hinges,
as patterns for steel panels to seal up the area. Steel is used, since rocks kicked up from the tires will
hit it directly. (Stainless should have been usedI forgot.) The steel panel is now welded in. The
passenger compartment portion of the shell will be permanently riveted to this panel, making for a
very rigid mount.
Some items were easy to place, such as these doorjamb backing plates. Since the door would be
slammed shut many times, the door latch and guide mechanism would take a beating. Since all the
force is in the in and out direction, a simple 1/8 plate did the trick. The extra holes lighten the
bracket and look nice.
An important goal was keeping the passenger compartment sealed against noise, dust, dirt, and
exhaust fumes. This meant fabricating flanges just about everywhereanywhere the passenger
compartment had openings, other than the doors. These pictures show some of the flanges being
attached to the main roll-hoop. Their task is to seal the gap between the main hoop and the shell.
This flange segment, in the two pictures above, is placed behind, and to the right of, the passenger
seat. The flange segments took a long time to make, but there was no way around it; doing it again, Id
do it the same way. The flanges stiffen the shell, (which is rather wobbly on its own), keep poisonous
exhaust out of the passenger compartment, and eliminate dust intrusion. Lastly, they make it quiet
enough to carry on a conversation without yelling (much). The large hole in the above flange is where
the Honda engine harness passes from the passenger compartment into the engine bay. The huge
grommet/seal thingy on the factory harness was retained, hence the big hole. Using it was the easiest
way to pass all the wires through.
At right, the bulkhead window flange is being mocked up to seal the gap between the cage, side
window, and roof of the shell. A surprising number of people think that I sit in the same compartment
as the engine. No, the Lexan window provides critical isolation against the possibility of fire. Yes,
the Lexan will melt and burn, but having the partition in place buys some critical time to pull the car
over and get out.
Backside view of the engine-bulkhead window frame. On the top, composite flanges extend down
from the shell and will be riveted to the steel flanges extending upward. Here, both the roll-cage-to-
shell flanges and bulkhead window flanges are being tack-welded. Note at the base of the upper
flange is a small 90 bend that keeps the thin metal from distorting during welding.
Endless bracketsthese stabilize the back corners of the passenger compartment. With the bracket
holding the passenger compartment flange, the Dzus fastener spring rivets also hold down the
passenger section composite. Its important that theyre located rigidly, since any air getting under the
leading edge of the engine cover could cause it to be ripped off by the wind. The tube near the tire
turned out to be a bit too close, rubbing under hard cornering.
All these busy tubes at the bottom are the diffuser framework below, their support tubes, and the
Dzus mounts along the curved tube, matching up with their mates along the bottom back edge of the
engine cover.
Front bumper tube with Dzus mounts. The idea is that it will take most of the force of hitting cones
at the autocross, the fear being that the force would otherwise crack the composite. The thin curved
tube in front of the radiator was bent by hand. It frames the Mini-shaped radiator inlet in the nose, to
which radiator ductwork is attached. Since it hugs the inlet, it minimizes air leakage around the
radiator. (This picture was taken before I realized it was now impossible to remove the fuel cell! The
tubes attaching the bumper were later cut and modified to be bolt-on.)
Next, a tube was hand-bent to follow the curve along the top edge of the windshield cutout. The
four tabs are the chassis-to-shell bolts along the front edge of the roof. The same bolts will attach
aluminum hold-down clamps for the windshield. Overkill? Probably, but the SCCA requires them,
because air coming in through the side windows (in a spin) can pressurize the air inside the car
enough to pop the windshield out.
Proof-of-concept drive
Technically, the car was now drivable, and to be honest, I was scared of it. I wouldnt be going
that fast, but there was the very real chance that something might break, fall off, crack, unscrew, leak,
or catch fire. Everything was suspect, but regardless, I pushed forward.
The purpose of driving this dune buggy now was to prove out the design, that the steering
steered, the shifter shifted, and the brakes brokeso to speak. It also forced me to account for every
bracket for every component. Of course, it also served as a tremendously needed psychological goal:
getting the car to move on its own. At this early stage, no paneling had yet been installed; I didnt
want to drill out any rivets before painting it. That meant there was nothing below my feet, so a scrap
of plywood was clamped to the chassis to temporarily serve as a floor. A little gas, the cooling
system filled, the brakes bled, and I was out of excuses.
Letting the clutch out slowly, for the very first time, to back the car out, was a monumental moment.
It was with almost childlike wonder, feeling the car move under its own power. It was like magic,
and if the car had fallen apart after just a few feet, it would still have to be considered a successit
had moved.
Since we have a sloping driveway, even backing out was traumatic (and it was surprisingly hard to
push it back into the garage). But nothing bad happened, so I backed out onto the street, making sure to
turn up the hill, in case I had to coast back home. I drove about 100 yards or so, driving over one
neighborhood speed bump, and it was cool watching the front rocker arms do their rockering. It
was surprising how soft the suspension was. I did a U-turn and found that the fast steering and short
wheelbase made it quick and easy. No smoky burnouts or drifting on this trip; I just wanted to get a
sense of what I had on my hands. One incentive to take it easy, besides getting marooned, was the
awareness of how many laws I was breaking: no license, registration, seatbelt, turn signals,
stoplights, headlights, wipers, speedometer, just to name a few. If the cops felt like it, they could have
really made life unpleasant. I pulled back into the driveway, got out, and breathed deeplyit was
nerve-racking. But since nothing terrible had happened, I took it out again, this time with camcorder
on board (the video is on the site).
Being braver, I set out to go around the block. This entailed going to the corner, turning down a
short, but steep, hill, down around the block, and back up another hill to our house. Turning down the
hill, I immediately heard the fuel pumps start clatteringa sign they were sucking air. It wasnt a
concern, since the fuel accumulator took care of such situationsor so I thought.
I got to the bottom of the hill, turned right, stepped on the gas, and the engine started coughing and
sputtering, just like in a Warner Bros. cartoon. It quitI had run out of gas. It was far from amusing to
coast to a stop in the middle of the streetat the bottom of a hill wondering how I was going to get
back up the hill and home; I vaguely heard a little voice in my head saying something about poor
planning. I didnt dare leave it sitting there. After thinking about what must have happened, I let the
pumps run a while, figuring they would eventually pick up fuel again. Finally, they did, and it started
right upa close call.
Once safely back in the garage, I thought about what had happened. First off, it was pretty obvious
that there wasnt enough fuel in the tank. Turning down the hill meant it had all run forward away from
the pickups. The accumulator tank should have handled the momentary fuel loss, but it didnt, so the
fuel pump delivery rates became suspicious. The high-pressure pump pulls fuel from the bottom of the
accumulator tank, returning it back to the fuel cell. This allows for the use of the large volume of fuel
in the cell to keep the fuel cool. Unfortunately, the high-pressure pump had emptied the accumulator
tank faster than it could be filled. The solution was to return fuel to the accumulator tank instead of the
fuel cell, so thats what was done.
The next day, after I re-plumbed the fuel system, I took the car out for another test drive. This time,
the fuel system did its job, running on the fuel in the accumulator tank, as it should. I hadnt added any
fuel to the tank, just to make sure it was working as designed.
If I had filled the tank completely before the first drive, I never would have
found this problem. Oh, sure, it would have been discovered at some point, no
doubt far from home. I was happy to discover it now.
It was good that the fuel situation was solved, but the engine still wasnt running quite right. Posting
the symptoms to Honda-tech, I was asked the obvious question, What codes are being flagged by the
ECU? As an electrical engineer, I was a little embarrassed to realize that I had never checked! Sure
enough, there was one error code, saying something was wrong with the MAP sensor. Some
bonehead, me, had wired it wrong! While I intended to do another drive after fixing it, to confirm the
engine now ran properly, I never got around to it. The assumption was that it really was fixed (it
was), but either way, it wouldnt be allowed to hold up the project.
I welded in the missing door bars and various other bits and pieces, and it was time to finalize the
welding. This needed to be done now, because the car was going to come completely apartfor
paint! Disassembling the car was easy, but boy, I didnt realize just how many welds werent
finished. That took a full day to complete.
And so, after years of design, fabrication, and blood, sweat, and tears, the day had come to
disassemble the chassis for paint. It was a huge step, like walking out of class after a big midterm
exam. These pictures also show how the engine tray is removed before the engine is rolled out.
The bottom-left picture shows how Id boxed myself in; the fuel cell could not be removed! The
front tubes were cut off and later bolted back on. Of course, this wasnt the end of it, as the final
exam of putting it together for real was still looming!
Subsystems
Subsystems have their own sections because they didnt really fit anywhere else. They are arranged
in order of placement, more or less, from the front of the car toward the back.
Regarding my comments in the following chapters, if they seem a bit negative, it isnt because Im
unhappy with how things turned out. Instead, I want to share my thoughts with the advantage of
hindsight, something I hardly ever see in other books. To me, reading, how I should have done it, is
far more helpful than, I did it this way, implying that the designer/builder somehow knew the best
solution straight off, without saying how he knew, how well it worked in the long run, or if hed do it
differently next time.
So heres how the various subsystems were builtwarts and all.
22: Front Grill
Composed of stainless mesh welded into a frame, the front grills job is to stop rocks kicked up by
cars ahead. On track, the speed differential between the grill and rock can easily be over 100 mph, so
it needs to be fairly tough. The mesh is 0.048 diameter stainless wire, with 0.25 spacing, meaning
that the actual openings are about 0.20, spaced four squares per inch. It was purchased from
McMaster.com.
The mesh is first clamped to its frame, bent previously by hand to fit the front grill cutout. I feel it
would have looked nicer if the mesh were cut at 45 to the frame, but since the material comes off a
roll with the wire aligned to the edges, a much larger (more expensive) piece would have been
needed.
Each strand is TIG-welded to the frame, making a very strong, yet flexible, assembly. It is flexible
enough that it can be bent back into shape if damaged and wont fall apart. The assembly is crowned a
bit so that the center bulges outward about one inch. This allows debris, like plastic bags or leaves, to
drop off when the car slows. These tips come from Alan Brickey, a top-notch fabricator who worked
on the Nissan GTP car.
The grill was originally going to have mounting tabs, but that seemed too complicated. Alan said,
If youre not going to remove it often, which you arent, why not just mount it with silicon seal?
Good idea, so thats how it is mounted. Some people asked why such large-diameter wire was used.
Its to stop rocks, but other things, too. I once came in off track with a 3 long, diameter bolt stuck
in my Datsuns headlight! To this day, I wonder what car could continue driving without such a large
boltor maybe it didnt.
23: Cooling System
As delivered from the factory, the Mk1 Mini has a rather unusual radiator placement, off on the
left-side wheel well. Air comes in through the front grill, passes through the engine compartment, and
exits through the radiator into the left wheel well, so the engine fan is actually pushing air out of the
engine compartmentneat. But since Kimini is a custom car, the radiator could be placed anywhere,
so its up front in the traditional position. Hot air leaves the radiator up and out a hood vent.
Aluminum pipes carry coolant between the engine and radiator and through the center tunnel. The
radiator is from Griffith, their small aluminum Ford unit. (Copper-brass radiators are much heavier
and have lower thermal performance.)
Sometimes a radiator must be modified to move a coolant port. Its
understandable to not want to weld to the tanks, thoughthe concern being that
the welding heat will ruin the epoxy holding the tanks on. I wanted to move a port
and called Griffith to ask about it. Their weldor said, Go ahead and weld it;
thats what I do. I welded it okay, but I put wet towels around the area to keep
the tank cool anyway.
Do not use epoxy to attach a plugit wont work. The problem is both the heat
of the coolant and the pressure cycling flexing the plate. It will gradually weaken
the bond, and when it goes, itll go in a big way.
Most racing radiators are fairly tall, but the cross-flow Ford unit was perfect for Kimini. The
radiator was modified to move the lower hose inlet to the top of the tank, necessary since both coolant
pipes come in at the top corners. This raised the question of how well the radiator would work;
would the coolant simply go straight across the core, instead of distributing downward across the
surface?
Ordinarily, the only time there is a cooling issue is during a long stop. Coolant flow apparently
slows so much at idle that the electric fan is only marginally effective, yet with engine speed
anywhere above idle, its not an issue. The situation doesnt happen often, though, so an electric
water pump is not used (theyre available for the H22A1, but theyre expensive).
Regarding electric coolant pumps, how do you know if they quit? During a
race, the engine will fail very quickly without coolant flow. An engine-driven
mechanical pump will at least start leaking or making noise to give some
warning; an electric pump just stops.
Coolant pipe size
Some of the aluminum coolant pipes have bends welded on to make them low maintenance. The
idea was to avoid having to replace a bunch of rubber elbows in locations hard to reach, but this may
come back to bite me. Since aluminum work-hardens with vibration, the aluminum will someday
become brittle and crack; the big unknown is how far away someday is. It all depends how well
isolated the pipes are from vibration. In hindsight, it wouldve been wise to just use straight sections
and automotive hose for all bends. As it is, if and when a pipe cracks, itll probably have to be
repaired in-place or cut out due to all the components preventing easy removal.
Lesson learned
I probably should have used slightly larger-diameter coolant tubing. The stock
Honda coolant lines are about 1.25, so that size was used throughout. In
hindsight, long coolant lines add fluid flow resistance. Other than being slightly
heavier, a slightly larger size is a good idea, especially due to the long runs
needed in a mid-engine car. Going to 1? for the longest runs between the front
and back of the car may have been smarter.
The electric fan is from Spal and is the largest unit that would fit. While twin fans look nice, they
are expensive, weigh more, and dont move as much air as one large fan. Even if they look cool, they
arent visible through the front grill. Spals Web site has excellent datasheets showing how much air
each unit moves.
The fan was put on the forward side of the radiator. This moved the fans CG
slightly forward and also keeps it out of the hot aira good thing, since fans push
air better than they pull.
Originally, a swirl tank was to be used as part of the cooling system, but I changed my mind
because there was little room. For now, only a header tank is used, filled by a bleed line from the top
of the cylinder head (from the stock Honda bleed point). A line from the bottom of the header tank
goes to the water pump inlet, which was originally the heater core return. While a bleed line from the
top of the radiator is recommended by Smith, there was no easy way to run it from the front to the
back of the car without the hose dropping down low through the center tunnel. Unfortunately, this
prevented air bubbles from being pushed back to the header tank, so for now, there is no bleed line.
The header tank has a 16 lb. cap on it, with an overflow tank for expansion.
There is a second radiator cap on the radiator, at the front of the car. Originally, both the front and
rear radiator caps were 16 lb. The front cap kept venting coolant after a hard run, but not the rear one.
The suspicion is that the water pump provides a couple of pounds of pressure to move the coolant
through the radiator; maybe it was enough to raise the radiator pressure slightly over 16 lb. To fix it, a
28 lb. cap was swapped in, while the 16 lb. cap remains on the header tank.
No expansion tank was used, until one day coolant was found spitting from the overflow hose after
some hard driving. Over time the cooling system would pump itself dry if nothing was done, so an
expansion was added, slightly higher than the header tank. Since it had to be mounted high, it was
attached directly to the Lexan bulkhead window.
After driving Kimini for five months through cool spring weather, I finally autocrossed it during
warm weather. During the event, the temperature kept climbing, though it wasnt a concern, since
there wasnt much time for the engine to cool between runs. Later, on the way home, the temperature
didnt drop. With a 180F thermostat, the car had always been well behaved until that day
apparently finally hitting the limits of the cooling system.
It was disappointing to find my ultimate cooling system had met its match in
only 95 heat at 80 mph on the freeway. So what to do? The first improvement
was getting rid of the antifreeze, replacing it with distilled water and Red Lines
WaterWetter.
WaterWetter and Royal Purple
This seemed like a no-brainer. Antifreeze is a poor thermal conductor, and if you get rid of it, the
cooling system becomes nearly 50% more efficient! Normally, you cant get rid of antifreeze because
of its corrosion protection and water-pump bearing lubrication, and of course, its freeze protection.
Since I live in southern California, I didnt need the freeze protection, so replacing the antifreeze was
investigated further.
A Google search on both WaterWetter and Royal Purple (the two seem to be nearly identical)
showed that virtually everyone was happy with them. While I dont trust much of what I read on the
Web, it is very good at finding unhappy people complaining about things. The good news is, there
were very few negative comments.
Theres a bit of controversy surrounding WaterWetter. Almost everyone finds
it works just great for them, but a few have issues. There were a couple of
reports of it clogging up the cooling system with goo. Several more posts
reported that WaterWetter left a white powdery residue in the water passages.
The suspicion was that this was due to antifreeze being part of the mix. Red Line
said it was possible that this goo could form if WaterWetter is mixed with certain
antifreezes. Since I dont plan to run antifreeze, I should be okay. In the end, it
was up to me to decide whether to use it or not, and based on hundreds of people
being happy with it, I decided to give it a trythat is, at least until I improved the
cooling system through other means.
Red Line claims the following temperatures for a Chevy 350 and a 160 thermostat:
Chapter 23: Table 1
Thats a very impressive improvement26 for just changing the coolant
mix; its almost good enough to not bother looking for any other cooling
improvementsalmost.
Another point about WaterWetter is that some racing organizations require it,
or rather, they do not allow antifreeze on track. Thats because when regular
coolant is spilled, it acts much like oilWaterWetter does not.
The other possible improvement to the cooling system is to enlarge the exit duct to also dump air
into the wheel wells. The radiator inlet is about as good as it can get, with its ram air, so the trouble
was very likely on the outlet side. But this modification would have to wait until aero testing could
confirm air flow; no point adding it if the flow wont improve. This is dealt with further in the
Aerodynamics chapter.
After using WaterWetter for a while, Im a bit less impressed with the
increased cooling efficiency. Or maybe Ive just been running the car in warmer
weather; its hard to tell. Regardless, the cooling system still needs improvement.
Exit ducting
Another question is, how restrictive is the radiator exit-grill? To keep continuity of design, the
same aluminum diamond mesh was used as at the back of the car. With about a 50% open area (and
even less with airflow), it cant be helping things. Below is an alternate layout, suggested by two
Web site readers with better styling taste than myself! It was initially rejected as cutting the airflow
down too muchnow Im not sure. I definitely like the aggressive look of the large louvers, and Im
considering changing to that style, if airflow testing shows that the mesh is too restrictive.
Theres another issue thats been eating at me, that I moved the radiator inlet
pipe from its original position down low, up to where it is at the same level as
the exit (my radiator is a cross-flow type). Im beginning to suspect that the
coolant is not distributing itself across the entire core; instead, maybe its just
going straight across the top?
The lesson is, keep the radiator inlet and outlet diagonal to each other. Even
better, make it dual-pass (see below).
So what else could be done? In the summer heat, the car ran hot, though hot is relative, running at
210 instead of 180 as it did in cool weather. Is 210 too hot? No, but to me it meant that the cooling
system was running at 100%+ capacity, the coolant temperature now rising along with outside air
temperature. This didnt bode well for future events at desert race tracks, where Id be hard on the
gas for long periods of time; I dont like running with zero margin.
At some point in diagnosing the cooling system problem, the question came up,
How hot is too hot? The consensus seems to be as long as its not boiling.
Putting a number on that isnt easy, but it seems to be around 230. So,
technically, I have nothing to worry about. What bothered me was the fact that the
coolant temperature rose with air temperature. Running hard on a desert track in
high heat seemed like a recipe for an overheated engine. Even though it had only
reached 210, I wanted more margin.
Multi-pass radiator
A few improvements could be made, like putting the radiator inlet back down where it was
originally, to better spread the coolant across the core. Another change is to convert the single-pass
radiator to dual-pass. Since the radiator is a side-flow type, cutting a slot in one tank and inserting an
aluminum divider would force the flow to cross the core twice instead of once. I learned that a dual-
pass radiator is about 40% better at rejecting heat than a single-pass unit.
Unless I was willing to put the inlet and outlet on the same side, it meant it
would have to be a three-pass arrangement, which has too much back-pressure.
Summed up:
Use larger-diameter tubing for carrying coolant from the front of the car to the rear.
Use a two-pass radiator.
Cooling efficiency mostly comes from how low a pressure you can achieve on the outlet side of
the radiator. Rarely is the inlet side the problem.
The irony of this is that while the cooling system can be improved to the point
where itll run cool no matter how hot the day, I may not! The truth is that driving
the car in temperatures greater than about 90 is just plain uncomfortable, and
with no air conditioning and small sliding side windows, its simply not fun.
Adding air conditioning isnt something I want to dotoo much complexity,
weight, and cost.
Evans NGP+
Evans NGP+ coolant (www.evanscooling.com) offers a solutionmaybe. Instead of ethylene
glycol, it is made from non-toxic propylene glycol. The entire cooling system is flushed out and
replaced with it, with no water added! It has a crazy high boiling temperature of around 375F
(191C.) Because of this, the cooling system pressure can safely be lowered to between zero and 7
pounds, so no more burst hoses.
However, switching to NGP+ is not a cure for an existing hot system; it just means that it wont
boil. Said another way, if I switch to NGP+, my coolant temperature will remain just as high. Not
having the coolant boil is a good thing, but it doesnt solve the original problem. In fact, this coolant
carries less heat than water-based coolants, so it will run even hotter than it did before the switch it
just wont boil.
I dont know enough about the NGP+ system to offer a recommendation either way, and at $32 a
gallon (as of 2007), its not a trivial thing to try out. Ill do some investigating first, to make sure
theres no downside to the product. My concern is that because its roughly three times thicker than an
ordinary water/ethylene glycol mix, it will further slow down coolant flow to the radiator. Im also a
little concerned that it may plug up water passages. I have no proof of this, but its something I need to
look into first.
My concerns aside, my brother said many people who use Subaru engines in aircraft use it with
absolutely no problems.
50% glycol/50% water 228
50/50 with WaterWetter 220
Water alone 220
Water with WaterWetter 202
24: Fuel System
Fuel lives in a 10-gallon fuel cell manufactured by Fuel Safe, part number PC-110. As Ive said
before, being scared of burning to death meant using a high-quality cell with a real bladder and foam.
Fuel Safes top-of-the-line unit was chosen, partly because it fit. Seeing a video of a driver nearly
burning to death due to his simple aluminum tank splitting open made the decision easy. Cost strangely
became unimportant when choosing the cell, but just so you know, it was $800.
Fuel enters the system through this sweet-looking filler from Aircraft Spruce. At around $100, it
isnt cheap, but it fit the 1960s look of the car and looks great.
There are two fuel pickups in the rear corners of the bladder, the idea being that one or both will
always have fuel under acceleration. Fuel is pumped out of the cell by two Facet lowpressure pumps
(the small cube-like devices in the picture below) into a small fuel accumulator at the top of the
photo. Each Facet pump supplies enough fuel to run the engine, in case one pump picks up air, or if
one fails. A line at the top of the accumulator returns fuel back to the cell to ensure that the
accumulator is always full. A second line, from the bottom of the accumulator, runs to the high-
pressure fuel pump located behind the engine.
After passing through the fuel rail and fuel-pressure regulator, the low-pressure return line goes
back to the top of the accumulator. The returning fuel from the engine could either run to the fuel cell
or to the accumulator; initially it was run to the fuel cell. The thinking was that returning fuel would
be warm and should be cooled by returning it to the large fuel mass in the cell. Unfortunately, the
high-pressure pump moves fuel faster than one low-pressure pump can supply, quickly emptying the
header tank. Because of this, returning fuel was instead run to the accumulator.
There was some concern that the fuel would boil, going round and round,
getting heated in the engine, sent back to the accumulator, then back to the engine.
It hasnt been a problem, probably because cool fuel from the cell is constantly
cycling through the accumulator.
What did I learn from all this? The Facet pumps are very noisy (watch some of my early test drives
to hear them). I ended up removing the rubber isolator mounts, which quieted them down a bit (the
panels they were mounted on were acting as sounding boards). They were still noisy and were later
removed completely from the panels, left to hang from the AN-6 braided line, which probably isnt
wise, but since they didnt weigh much, it didnt seem like a problem. All the plumbing around the
low-pressure pumps and accumulator is expensive, awkward to access, and heavy. The way to get rid
of it all would be to use an in-tank accumulator, with the high-pressure pump in the accumulator. The
only drawback of the Fuel Safe accumulator is the cost, but having done it the hard way, Id probably
just use their accumulator and be done with it. Short of that, Id at least install a baffle between the
front and back of the cell. If the tank is less than half full, under hard braking, the fuel runs forward
and uncovers the pickups. (I can tell because the pumps get loud when theyre sucking air.) The
engine doesnt quit because the accumulator tank is doing its job, but a baffle would be a simple
improvement.
Whether or not an accumulator tank is needed will depend on the shape of your
fuel tank. An external one will be cheaper but adds to the complexity, with more
hoses and hose-ends, potentially additional pumps, and increased noise.
Another issue with using a real fuel cell is the foam; it doesnt last forever, eventually breaking up
and clogging the filters and pumps. Its not a big deal, if youre aware of it, but if not, itll be a cause
of confusion as to why the car starts running badly after a few years (due to a plugged-up fuel filter).
When the cell is removed to replace the foam, I may replace the external pumps and hoses with the
Fuel Safe accumulator and in-tank pumpa much simpler, cleaner, lighter, and quieter solution, but
its not cheap.
I highly recommend to people building cars, especially Super-7 type cars, that they seriously
consider a real fuel cell. The gas tank in a Super-7 is in a very bad location, at the extreme back end
of the car, with virtually no protection to the rear. It pretty much serves as the crushable structure
when hit from behind. Yes, a Seven is lowlower than most production cars bumpersbut if the
car coming up from behind has its brakes on, its bumper will be much lower and could strike the tank
directly. The Super-7 has only a few small chassis tubes protecting the tank. In my brothers Stalker,
he located the tank forward of the axle to at least try to give it some protection.
The fuel cell
Fuel cells get a bad name because of their high costbut consider the consequences of using a
plain aluminum fuel tank. In a crash, the impact will compress the tank, using the fuel as the hydraulic
medium to distribute the shock load throughout the tank. The weakest seam will split, spraying fuel
into the airan explosive mixture. Since the driver is only about two feet forward of the tank (in a
Seven-type car), he will very likely be engulfed in an explosive cloud. I saw a video of this
happening to a sprint car that had been rear-ended at about 10 mph. The plain aluminum fuel tank, in
the tail of the car, split open, spilling fuel around the drivers seat and under the car. He was in an
inferno for over a minutehow long can you hold your breath as you burn? It just about made me sick
to watch, and it convinced me right then to use the best fuel tank possible.
There are several very important reasons to use a real cell; here is what fuelsafe.com says (text
credit Fuel Safe):
What is the difference between a fuel bladder and a (hard rubber) plastic cell?
Fuel bladders are designed to freely deform and absorb energy under impact, like an air bag in a
car. The more energy the cell absorbs, the lesser the chance of a rupture. A so-called (hard
rubber) plastic cell does not have the ability to freely deform or absorb an impact. These cells do
not have nylon to reinforce the plastic, which is needed to resist a puncture. Most (hard rubber)
plastic cells are made from a crosslink plastic, which is a very slippery material. This material
allows a projectile to slip into and through the cell with very little or no resistance at all.
What is the foam inside the bladder for and how much of my total capacity does it require?
The foam is one of the most important components of a fuel cell. It is designed to reduce the slosh
of the fuel and the chance of an explosion by reducing the air volume in the cell. If the cell does
ignite internally, the foam absorbs the expansion and the energy of the explosion. At that point, the
oxygen is used up and the flames go out. The cell must be filled with at least 80% foam to perform
effectively. The foam only requires 3% of your total capacity.
They have another important note: never use PVC pipe as part of the refueling
pipe. It produces static electricity and may lead to an explosion.
If you use a fuel cell, be sure to check whether the liner and foam can handle
alcohol. More and more states are blending ethanol into pump gas. Even if your
state doesnt use it now, it may be worth buying a cell that can handle it.
25: Brakes
I hadnt planned to have a chapter on brakesit seemed so simple. Get some calipers, a couple of
master cylinders, and brake pads, plumb them all together, and voila perfect brake system! It just
didnt work out that way. My tale of woe is (I hope) to make you aware of what it takes to have a fine
brake system. The overall goal is to remove as much compliance as possible from every brake
component. I only wish I would have figured that out earlier.
I used as many donor parts as possible to keep costs down. This was fairly easy, at least with the
front suspension. Since 280ZX struts were used, it was a no-brainer to also use the Nissan calipers
that came with them. They were cheap, and I didnt have to re-engineer anythinga real time saver.
At the back, the Honda brake rotors were used, since they fit the factory half-shafts. Since the stock
Honda calipers were enormous 13 lb. affairs, which wouldnt fit in the 13 wheels, I replaced them
with lighter, second-generation front (yes, front) RX-7 calipers. At nearly half the weight, and with a
much lower profile, they worked great.
With the calipers selected, choosing the correct size master cylinders was next. A very big help is
an Excel spreadsheet created by Mike Polan: 7builder.com/Downloads/MikesBrakes.xls. What the
spreadsheet cant be expected to incorporate are factors such as expanding brake lines, flexing
calipers, brake fluid choice, and how much pedal pressure you feel is correct. This is not a negative
comment about Mikes spreadsheet but an observation that some factors are difficult or impossible to
measure. At the time, I didnt know any better, so I dutifully entered my numbers, and the spreadsheet
selected a 0.70 master cylinder diametercool.
A short honeymoon
During initial test drives, the brake pedal was soft, but I didnt worry about it, chalking it up to
having air in the system. The first hint of a design problem was noticed after I had finished the car
the soft pedal was still there. Soft here means that the brake pedal acted as though there was a big
spring somewhere in the system, consuming about 2/3 of the pedal travel before the brakes would
lock up. Bleeding the brakes several times failed to improve matters, and the soft pedal became
frustrating.
Over the following months, my annoyance with the brakes grew; the problem was always the
subconscious worry under braking, the little voice in my head asking, Is it going to stop this time?
After yet another brake bleeding session had no effect, I decided that the issue had to be solved. The
cause of the spongy pedal was unknown, but suspicion fell first on the flex-lines, which were used
throughout the car. Could they be expanding and causing the soft pedal?
Hard-line vs. flex-line
Asking car designers, builders, and drivers only served to muddy the waters further. Some said
switching to hard-line would absolutely help; others said it wouldnt make any difference. A technical
rep at a manufacturer of flex-lines said, yes, it does expand with pressure, but in the range of use
(<2000 lb./sq. in.), it would expand about 0.001%a very small number! But Im always suspicious
when requesting information from a source that wants my money; how much could I trust them? Given
that skepticism, I decided to switch to hard-line anyway, to find out for myself.
Switching to hard-line was a pain because the car was already done. If hard-line had been installed
as the first component on the empty chassis, it would have been easy. Now, everything seemed to be
in the way!
Doing it over again, I would definitely use hardline. Its a combination of
cheap, simple, light, and reliablehard to beat. Im not sure what I was thinking
when I went for expensive and heavy flex-line. (It was the lack of familiarity
with hard-line; I thought that it would be hard to install. It was not, as long as its
installed on an empty chassis.)
There is one annoyance with hard-line fittings: I found that I have to really
tighten the hard-line junctions; otherwise they leak slightly. Even when using
flare wrenches, I was always afraid I would round off the hex pattern on the
fittings.
After completing the swap and bleeding the system, what was the result? At first, I felt it was an
improvement, but over time, I think it was only psychological. In other words, the improvement (if
there was one) was so small that my imagination led me to think there was one. I dont think there was
any improvementvery disappointing. Even though there was no noticeable improvement with hard-
line, I still recommend it over flex-line due to its low price and light weight. Flex-line must still be
used where suspension motion occurs, though.
Related to the use of hard-line, the brake flaring kit from Harbor Freight is
really poor. Granted, I havent used any other brand, but anything has to be better
than theirs. I do not recommend it.
At this point, I was a little fed up with the brake issues. I just wanted to drive the car instead of
messing with stuff. But the soft-pedal issue never went away and was a constant annoyance.
Flex-line life
Stainless flex-line is used to connect the hard-lines to the calipers. There have been some reports
that the line doesnt last long, due to sand or grit getting under the stainless braid and wearing through
the Teflon tubing. While I have used stainless flex-lines for years without problem, I can see how it
can happen. If you plan to use the car in a dirty environment, consider adding shrink sleeves over the
tubing.
Caliper flex
Since brake-line expansion wasnt the problem, the next suspects were the calipers. That suspicion
proved correct during one of my marathon brake-bleeding sessions, where, on a whim, I unscrewed
the flex-lines from the calipers and screwed plugs into the lines. Huh, well what do you knowa
hard brake pedal! Replacing the calipers, though, is an expensive proposition, plus, the new calipers
wont line up with the original mounts, necessitating new mounts or adapters. I put the idea aside due
to the expense, plus I wasnt sure that replacing the calipers would fix the problem. New ones are not
always guaranteed to be stiffer.
Trackday trepidation
The brake issue came to a head after the first trackday, where I found that I was subconsciously
braking early; my brain just wouldnt allow me to brake late with a soft brake pedal. The track,
Streets of Willow Springs, is really twisty; the brakes on every car took a beating. Over the course of
the day, my brake pedal slowly got softer. I thought it was due to the brakes overheatinga little
surprising in such a light car with relatively low power. What was strange, though, was that after the
car cooled off, the pedal remained soft. Even more surprising was that when I was loading the car
onto the trailer, the pedal went to the floor! It pumped right up, but it was unnerving; what the heck
was going on? No leaks, and the pads were still in good condition, but something was going on.
I was told later that it was due to the brake fluid boiling. Once that happens,
the bubbles do not go back into suspension. Someone else said that, no, it was the
brake pads cooking, and now they were compressible. But changing out the fluid
brought the pedal back to its original firmness.
Synthetic brake fluid
Talking to several people about what had happened at the trackday, I casually mentioned that I was
using synthetic brake fluidboy did that get a reaction! The big surprise was how universally
disliked synthetic brake fluid is, at least when used on track. I heard stories of otherwise leak-free
systems developing leaks (which mine did), perpetually soft brake pedals (which I had), and brake
pedals that went to the floor, even after the car cooled off (my problem, too). Well, this certainly
sounded like a clue! Could this be the cause of all my problems? I checked to see what brake
component manufacturers said, and many, indeed, recommended not using synthetic brake fluid on a
race track (including Tilton, maker of my master cylinders). Manufacturers noted that synthetic is
much more compressible than normal brake fluid. Ah hah!
I was asked, How can you have synthetic brake fluid? Its not like theres
natural brake fluid out there. What, do you get normal brake fluid from a
tree? Apparently, the term synthetic refers to silicon brake fluid, a type I knew
to avoid, yet the stuff I used didnt have that word anywhere on the label.
Having had enough of my marginal brakes, I set out to fix the problem once and for all. Brake fluid
would be switched to normal fluid, with the high hope that it would fix the soft pedal. If not, new
calipers would be installed.
Which brake fluid?
Asking around and doing a lot of reading, I came up with this short list of brake fluids that everyone
seems to agree is good stuff while still being affordable and available:
Ford PM-1
ATE Superblue
Castrol LMA
There are other, much better, brake fluids, such as Castrol SRF and Motul, but they can cost up to
ten times as much. Based on this, I flushed out the system and replaced the synthetic fluid with Ford
PM-1, with a dry boiling point of 500F. So after the fluid was changed, did that fix all my problems?
No!
In the back of my mind, something was bugging mehow the brake pedal got hard when I had
plugged the flex-lines right at the calipers. I guess it just took a long time for the implication to sink in.
Since the pedal was hard, it meant that the entire problem was contained in the calipers, not in the
lines, not in the fluid, and not in the master cylinders. Logically then, the problem had to be flexing
calipers, air (still!), or compressible brake pads. I dont believe that compressible pads are an issue,
and since the system had been bled like crazy, I doubted it was that either.
Several people suggested that I hadnt bled the calipers correctly, or I hadnt
oriented the calipers correctly to purge all the air. They are correctly oriented;
plus, I even tried unbolting them, moving them around, and even whacking them
with a mallet to try and knock any bubbles out of them no improvement.
That left a suspicion that the calipers were flexing.
Flexing calipers
Most production brake calipers are of the slider type. This means the assembly consists of two
parts, where the single piston presses on one brake pad, and a sliding frame connects the back side of
the piston, up and over, to the opposite pad. This type of caliper handles misalignment quite well and
is inexpensive, hence its widespread use. Could the sliding assembly be introducing compliance into
the system? To find out, it meant I had to bite the bullet and buy new calipers.
Some of the popular brake manufacturers are AP, Stoptech, Brembo, U.S. Brake, Outlaw, and
Wilwood. Unfortunately, most of the calipers are too large to fit 10 discs under my 13 wheels. But
Outlaw and Wilwood do have smaller calipers. Outlaw, though, proved to be hard to find, so
Wilwood was investigated.
After reviewing their different units, I settled on the forged billet Dynalite caliper. (Another unit
looked even better, the radial-mount Powerlite caliper, but it was not as widely available.) Wanting
to stick with easily available parts, I ordered two Dynalite calipers$250 for the pair, plus $60 for
pads, from summitracing.com.
When they arrived, I marveled at how light they were; they would remove a whopping 16 lb. of
unsprung weight from the front of the car. Unfortunately, the mounting ears didnt line up with the
existing mountsnot unexpectedso the existing mounts were cut off and new mounts were
fabricated and welded on. Of course, this meant that my pretty red powder-coat would get burnt off
from the welding heat, but so it goes.
As part of the upgrade, I again used Mikes spreadsheet to find the recommended master cylinder
diameters. The size recommended very much depends upon the size, and therefore surface area, of the
caliper pistons. The Wilwood calipers I chose have four 1.38 diameter pistons. Based on this, the
spreadsheet specified a 0.813 cylinder for the front, and a 0.875 for the rear. But since I had a
balance bar, I felt that perhaps the two existing 0.813 units might work fine.
So, after all the work and expense, did the caliper upgrade finally fix the soft pedal problem? Yes!
There was a tiny doubt that the upgrade would workabsolutely. Just because
theyre pretty, light, and forged is no guarantee that the new calipers would be
better. In fact, some race teams in the past found that lightweight aluminum
calipers are actually more flexible than the heavy, but stiffer, steel units they
replaced. Also, several people wrote me, saying that their stock 280ZX calipers
worked just fine. So there was no guarantee that this was a certain solution, but it
certainly worked well in my case.
The new calipers worked much better, removing almost all of the compliance (the remaining
compliance may be the rear calipers). When I took the car out to bed-in the new pads, it entailed
doing a bunch of near-stops from about 60 mph. When I first started out, the brakes were only eh in
terms of stopping power, even though the pedal was now hard. In fact, I wasnt sure if it was the un-
bedded pads, or if I had used the wrong master cylinder diameter. The effect is the same: lots of pedal
pressure doesnt slow the car much. But after doing several braking runs, I found that it just kept
getting better and better. In an odd complement to how well the brakes worked, I made myself
carsick; the car kept slowing down faster than my stomach expecteda good thing, once I got used to
it! The brake pressure was now just right, so the master cylinder diameters were left alone.
Power bleeder
For bleeding the system, my buddy Dennis recommended a pressure bleeder. Since Id been having
brake problems for so long, I finally broke down and bought one. Its a little jug with a pump, kind of
like a garden sprayer. It worked like a charm, making the task of bleeding the brakes truly an easy
one-man operation. I used it for the first time to bleed the new front brake calipers, which of course,
brings up the following question.
Did the new Wilwood calipers really fix the brake compliance, or did the brake bleeder get rid of
some air that resisted removal by other methods? Ill never know; I wasnt going to put the old
calipers on to test it. Im just thankful that the problem was finally solved.
Pedal cluster
A Tilton master cylinder assembly was used. While it worked fine, there is an aspect of it I dont
care for: a leak anywhere in the system will cause the entire braking system to fail. Thats because if
there is a leak in one circuit, the balance bar transfers all force to the soft sidethe failed side.
Theoretically, the good circuit still works, but it means that twice the pedal force is needed to
activate it. Its questionable whether that amount of force could be applied in the little remaining
pedal travel. One way to correct it is to limit how much the balance bar can tip, though it might
prevent the brake balancer from working as designed.
Doing it over again, Id consider using a forward-facing master cylinder
model, to get the reservoirs out from under the base of the windshield (but then
theyd hit the rocker armsits always something).
Master cylinders
Tilton Engineering also makes brake master cylinders, in addition to the pedal assembly.
(Wilwood also makes identically sized master cylinders.) They are made in 1/16 diameter
increments, from 5/8 to 1.
Remote bias adjuster
Connecting the brake balance bar to the dash is this nifty Tilton bias adjuster. Once set up correctly
(after I finally read the instructions!), it made adjusting the brakes, on-the-fly, very convenient. The
knob is mounted to the dash, and a flexible cable is run to the balance bar. With this setup, I dont
have to remove the nose to set the brake balance.
Which way do I move the balance bar to adjust the brakes?
Move the balance bar toward the master cylinder that you want to do more of
the work. For example, if the rear brakes are locking up, move the balance bar
toward the front master cylinder.
Rattling brake pads
This was an issue while still using the stock Nissan 280ZX calipers.
While it wasnt a big deal, the constant rattling was an incessant irritation. Eventually it got the best
of me, so the brakes were disassembled to figure out what was going on. Apparently, with each
rotation, the discs would barely catch the pads and give them a little push. Then, the little spring clips
designed to keep tension on the pads, to keep them from rattling, would push them back the other way.
Back and forth theyd goclick, click, click. Finally, I realized that if I swapped the spring clips to
the other end of the pads, theyd keep the pads pushed in the direction the rotor was trying to push
them. Presto!no more rattling brakes.
Flex-line length
Make sure brake flex-lines are long enough! I was sure that mine were, and I still got bit. I have no
excuse for this happening, either, because I knew better. I had checked the front lines by eyeballing
them instead of actually moving the suspension through its full range of motion. I dont know how long
these flex-lines would have lasted, but it made me very nervous. The lines were replaced with longer
hoses when the brakes were upgraded.
Big brakes on a street car
There is a common impression that upgrading a brake system makes the car stop in a shorter
distanceno. Think of the braking ability of the car as a complete system, from the pedal all the way
to the ground. Upgrading the brakes typically involves larger rotors and new brake calipers. But this
misses the one key element in transferring this increased braking ability to the groundthe tires.
As an example, consider a car with anti-lock brakes. When braking at maximum effort, the brakes
lock and unlock the tires; they are right on the edge of adhesion to slow the car down. Now, the brake
rotors and calipers are upgraded. When the driver goes out for a test drive, whats going to happen?
The exact same thing: the tires, right at the edge of adhesion, will slow the car at exactly the same rate
as before, and the car will stop in the same distance. So when owners say that their new brakes stop
the car much faster, its the psychological effect of spending all that money that makes them think
theyre better! A brake upgrade provides a larger heat sink to dissipate heat. The larger pads will last
longer before fading, but it wont stop the car any faster! For a street car, its a waste of money.
26: Roll Cage
The SCCA knew far better than I did about suitable roll-cage tubing size, so the SCCA rule book
was consulted to find the correct diameterif its good enough for race cars, its good enough for me.
For the expected weight of the car, the rules specified 1.5 OD, 0.125 wall thickness DOM tubing.
The tubing requirement is solely dependent upon the expected weight of the finished car.
If youre starting from scratch, be sure to check the latest SCCA specs, since
they can change at any time.
The tubing closely follows the inside surface of the passenger compartment for several reasons: it
looks nice, makes adding shell mounting tabs easier, increases headroom, and gives a larger cross-
sectional area, increasing torsional stiffness.
As Ive said elsewhere, roll-bar padding was added anyplace I thought my body could impact it.
The correct stuff to use is not hot-water pipe insulation; your head will crush it down to nothing and
then itll be like hitting the steel cage directly. SFI-45.1 is the only stuff to use and is much firmer than
the cheap stuffalmost hardbut can save your life. (Summit Racing has it for about $13 per 36
section.)
One small irritation about the stick-on roll-bar padding is that the adhesive
fails. Ive now had two pieces simply drop off while I was driving, one hitting
me in the head (What the?). The answer is to either use tie-wraps, or
perhaps weatherstrip adhesive.
27: Electrical
The electrical system is a big topic and a huge time sink for any car project; it simply cant be
avoided. It has to be done right, or theres no point in doing it at all. It took months of unglamorous
work, figuring out where each and every wire went, but there was no way around it.
The four things that cause electrical problems are:
1. failed solder or crimp joints.
2. failed connectors.
3. wire insulation wearing through.
4. failed components.
Prevent these from happening and youll have a reliable, long-lasting electrical system.
Failed solder or crimp joints
I often read arguments about soldered versus crimped connections. In the aerospace industry,
where there is no cost limit to doing it right, all cable connections are crimped, so what does that tell
you?crimping is superior. This pertains specifically to crimp pins used in Mil-Spec circular
connectors. Other types of crimped pins and connectors dont measure up. The common D-shell
connectors (the old 9-pin serial connectors on the back of your PC) are very unreliable; the problem
is that the wires arent well supported. Probably the worst, though, is the common spade-lug
connector. The usual failure mode is that the wire simply pulls out because it wasnt crimped tightly
enough during assembly. The typical crimp tool sold in most stores isnt adequate; the proper tool is
the ratcheting type that applies the proper pressure every time.
For soldering wires together, always put a section of shrink sleeve over each connection. It not
only protects the bare wire junction from shorting against something, but it helps prevent the solder
junction from flexing due to vibrationa key element for failure.
Other than by corrosion, contacts rarely fail by themselves, but rather, its the
contact-to-wire junction. If the wires coming into the connector shell are not
supported, they will vibrate, flexing right at the connector-junction point. Given
enough time, the copper wire (like aluminum) work-hardens, becoming more and
more brittle, until eventually the strands crack through. The key to a reliable
connector is preventing the wires from flexing by properly supporting them. This
can be done by either using the proper clamping backshell or even filling the
finished connector shell full of silicon sealanything to prevent vibration.
Rigidly supporting the wires to the backshell ensures that when the connector is
unplugged, the wire/contact interface isnt put under tension.
Failed connectors
You get what you pay fordont use cheap-quality connectors. The military-style circular
connectors I used got lots of attention, with everyone wondering where to get them. The ones I used
were surplus, and I found them at Industrial Liquidators. Allied Electronics carries plastic-bodied
circular-type connectors that are very similar and affordable. The real problem, though, isnt the
connectors themselves; its the crimp pins. The proper AMP tool (Amphenol is a huge connector
manufacturer: http://www.amphenol.com/) for crimping pins is around $400 and has special inserts
for each style of pin. If you use circular connectors, it is possible to solder wires directly into the
crimp pins, instead of crimping them, as long as they are protected from vibration.
Wire insulation wearing through
Protect the wires! That means using common sense, not running wires around sharp corners or
though areas where they can get pinched. Put something over themanything is better than nothing.
Any time a wire goes through a panel, use either a bulkhead connector or at least a grommet.
There are two types of sleeving sold through most automotive supply stores, Home Depot, Lowes,
and Frys: spiral wrap and split sleeve. I dont like the spiral-wrap type because it takes forever to
put on: it has to be wrapped around and around (and around and around) the harness. Worse, if you
ever want to add a wire, theres no easy way to do it. Either the wrap has to be removed, or the wire
must be pushed under every turn of the wrap and pulled throughnot an efficient use of time. I much
prefer the split-wrap stuff, like whats used on new cars. Available in many sizes, it is inexpensive
and perfect for a self-built car. Thats because its very easy to stuff just one more wire into an
existing bundle.
As a final touch for a finished harness, when you are sure you dont need any more wires, wrap the
final assembly with high-quality vinyl electrical tape. The good tape, as well as the ordinary type, is
available through Home Depot or Lowes, so theres no excuse not to use it. After the assembly is
done, it needs decent support so it cant flop around, risking failure by vibration. I used rubber-lined
clamps from Aircraft Spruce, the same type of clamp used to support the coolant pipes. They are
sometimes referred to as Adel clamps.
Another failure mode is the insulation melting through. It is up to the designer to be sure that all
wires are far away from high heat sources, of the proper insulation type, thermally insulated, or all the
above.
Failed components
A rare failure mode is when a part shorts internally. These can be tough to troubleshoot, especially
if its intermittent. I had an odd failure that could have been really hard to find and which had the
potential of starting a fireexcept for a lucky break, so to speak. This is from the build diary:
I had been working on the dash, which is a sheet of carbon fiber (which is electrically
conductive). Id removed it from its mounting, so that it was sitting loose on the frame of the dash.
Having just finished a test, I turned the ignition key off, and something caught my eyea wisp of
smoke! That was odd, and looking over the dash, I couldnt find any problem. I carefully turned
the ignition switch back on nothing. Oh well, figuring Id been seeing things, like dust in the air,
I turned it off againthere was the wisp of smoke again! But this time, I saw where it came from:
right where the corner of the carbon dash was touching the steel chassis. That meant something on
the dash was feeding voltage into the composite, but what was it?
After removing one item at a time from the dash, to isolate the problem, I foundnothing. Okay,
fine, everything was put back, the ignition switch turned on, and no problem.
Went to turn it off again, andsmoke. Huh? The smoke only appeared when the switch was
turned off, but not when it was turned onwhats with that? Eventually, the ignition switch itself
became suspect, and indeed, it was momentarily shorting the main +12 V supply to ground, right
when the switch was halfway between Off and On; I was fortunate to find it. Once the panel was
screwed down to the dash tubes, the energy would have gone into making the wires smoke. I count
myself fortunate to have found it before more serious damage occurred. Oh, and this was a brand-
new switchmade in China.
Switches
I ordered toggle switches from McMaster.com, simply because it was convenient, and they had the
specs in the on-line catalog. The other place to check out is digikey.comthey have everything
electrical and electronic! The trick is to figure out the current ratings before buying them. Often, a
switch is advertised as 10 A but digging deeper, itll turn out that the rating is at 115 VAC. Seems
good enough since were only using 12 V, right? Nope. AC (alternating current) means that the
current alternates directions, so at some point the current is zero. When the switch is opened, any arc
that forms will disappear within one cyclenot so with DC (Direct Current). When a switch is
opened in a DC circuit, an arc forms, which can quickly destroy the switch. The faster the arc can be
extinguished, the better, and this is done by opening the contacts really quickly. Switches designed for
DC have stronger springs to open the contacts really fast. So, the goal is to find switches that are
designed for DC; the AC rating really means nothing. Realistically, this can be hard to do, since most
switches dont even list it. The best that can be done is to use switches that at least have a DC rating.
In general, use a relay when switching current higher than what the switches can handle, or around
10 A.
Wiring the car
The first thing to do is to make a list of all the electrical loads: lights, ECU, fan, fuel pumps, vent
fan, etc. As a very rough idea of what currents are involved, here is my setup:
Chapter 27: Table 1
Based on this, a suitable fuse box and wire harness were ordered from Painless Wiring
(www.painlesswiring.com).
I was fairly happy with the Painless harness. I wish theyd supplied a real
schematic of their fuse box, though; they conveniently assume everyone uses
domestic GM switches. This forces everyone into a one-fit solution, which
Kimini didnt fitI didnt want to use a huge GM ignition switch or a GM
headlight dimmer switch.
They label every wiresometimes extremely helpful, at other times
misleading. For example, they have generic signals like ACC1, which is fine
as long as you have a spare ACC1 to hook up! The point is, there is a fair number
of wires named for things I didnt have, wouldnt use, or would use for different
things. It would have gone faster if they had supplied the fuse box with a good
schematic and separate rolls of wire.
Complaints aside, if youre wiring an early American hot rod, youre all set,
since thats exactly what these kits are intended for.
What wire to use?
Definitely stranded wirenot solid! Copper work-hardens just like aluminum, getting stiffer and
brittle over time, and eventually cracks through. For insulation, stay far away from thermoplastic; the
stuff is terrible. Its very soft, and if a wire is run over an edge, over time the insulation flows away
from the edge and causes a short. When thermoplastic covered wire is soldered, the heat causes it to
shrink back up the wire a surprising distance. Oh, and it gives off toxic fumes when it melts or burns.
Use Tefzel insulation, cross-linking poly, or Teflon; anything is better than thermoplastic.
Once the harness arrived, it was time to start drawing schematics. It was pretty straightforward
since the electrical system as a whole is simple. Theres the lighting system and the ECU, and thats
about it. The tricky parts were the turn signals and hazard lights. The surprise was the turn-signal
switch; I had no idea they were so involved. Thats because the taillights have three functions with
only two filaments: stoplights, turn lights, and running lights.
The turn-signal switch has to do quite a bit of trickery to make it all work, so I bought one intended
for hot rods from speedwaymotors.com. Words cant quite describe my experience with this switch
well, maybe you can get the idea from the picture. From the build diary:
Every single component in that thing was defective in some way. I ended up completely
disassembling it, first to figure how it was supposed to work, then fixing the springs holding down
the switch contact. Turns out the springs werent long enough. Then the turn stalk wasnt indexed
right; in the center position some of the switch connections werent being made. This was in
addition to the uninsulated (bare!) wire inside its metal housing and the wires passing through the
back of it, through a hole with sharp edges and no grommet. Amazing, and it was brand-new. There
was a good chance of a fire if it was used as is.
You can see that I was pretty upset when this was taken. I didnt return it because I figured they
would just send another. (I did send a complaint, and while they refunded my money, they continue to
sell them.)
After receiving the stock Mini Mk1 taillight assemblies, I found that they used three separate
filaments. This made the circuitry much simpler to design, so the Speedway Motors turn-signal switch
was cheerfully dropped in the trash. The schematic was modified to suit a much simpler common
double-pole double-throw (DPDT) switch for the turn signal, placed on an aluminum extension so I
could reach it from the steering wheel.
I should have used a momentary switch; Ive forgotten to turn if off several
timesdriving along for miles, just like an old driver. I cant hear the flasher,
and theres no automatic cancel mechanism.
With the schematics done, wiring could finally begin. First was the dash, constructed on a carbon
panel, just because I had some and it looked nice. This carbon panel has some history. The panel here
was used on one of the Nissan GTP cars, making it the only part of Kimini to travel faster than 200
mph! While it looked nice, the fact that it was honeycomb meant that it could get easily crushed by
over-tightening the switches. Next time Id just use a black-wrinkle finished aluminum panel.
Since the switches couldnt be tightened too much on the carbon honeycomb,
and since switch-mounting hardware is notorious for loosening anyway, a blob of
silicon was put behind all switches to lock them in place.
The dash is designed to come off quicklyhave you ever removed a dash from a modern-day car?
It can take hoursno thanks. All connections to the dash feed through two circular connectors (from
Allied Electronics). One goes to the front of the car, for the lights and fuel pumps, and the other heads
to the back for everything else.
Heres why its a good idea to carefully lay out the instruments before mounting them. In this photo,
taken from eye level, all gauges are visible through the steering wheel. Early on, a cardboard template
was mounted in place of the dash. The edge of the steering wheel, between my eyes and the dash, was
traced onto the template. This marked the edge of the visible area for all the instruments, so theyd
always be visible.
With the dash done, the biggest job of all was started, going through each and every wire of the
Honda ECU harness. This is where the factory service manual isnt just a good idea, it is an absolute
necessity! After each wire was examined (and removed if possible), the ECU was mounted to another
carbon panel. Every wire was first brought out to terminal blocks and then on to its original
destination. This makes every signal easily accessible for troubleshooting. Wonder what the MAP
voltage is? No problemjust put a voltmeter on its terminal.
Grounds
Poor grounding causes more trouble than any other electrical problem; good grounds are extremely
important!
Do not share grounds between circuits, connecting dissimilar circuit grounds together and
expecting them to play nice. When high current from one circuit goes through the shared ground
wire, it affects the other circuit. This can be very bad news if the other circuit contains low-level
signals, like ECU sensors do.
Make sure that all grounding attachment points are clean. That means no paint, pinched
insulation, rust, dirt, or oil. I welded small bolts head-down to the chassis for easy ground
connections. Paint had to be removed before using them.
Ground wires must be at least as thick as the power lead supplying the circuit larger is better.
Think of wires as water pipes. If you have a big pipe feeding a circuit and a small one leaving it,
hows the current going to get out? The small wire will limit the overall current.
If the donor car had shielded cables on the ECU signals, be sure to reconnect them to ground. It
is a sure sign that these signals are very sensitive and need protection from electrical noise, such
as the ignition system. My Honda H22A1 factory service manual has pages of diagrams covering
nothing but ground wires.
If the donor car had engine-to-chassis grounds, be sure to reinstall these, too. Just because the
engine is metal doesnt mean its well grounded. Rubber engine mounts cause it to be
electrically isolated. This goes for the cylinder head, too. This is why its especially important
to have a big fat wire running from the starter motors case to chassis ground. The 300 A of
current running the starter motor want to get to ground badly. If the starter motor ground is left
off, that 300 A will try running through the other ground wires, all much smaller. The result will
be melted wires, a destroyed ECU, and maybe even a fire.
Do not count on carbon fiber to conduct electricity. It will, but not very well. On the other hand,
another issue is the unexpected consequences because its conductive. That is, some people think
its like fiberglassan electrical insulator. No, it has carbon in it, so it does indeed conduct
electricity. If a bare wire touches it, it will electrify the entire carbon panel. Of course, since its
only 12 V, its not enough to give you a shock, but enough current can flow through it to cause it
to get hot, melt, or catch on fire.
The best grounding system has grounds from all the subsystems returning to one point, with that
point being the only one that connects to the chassis. This is not practical in real life, so the steel
chassis is used for most grounds. This works fine in practice.
Ground wire rant
There are engine compartment wiring kits that consist of extra-heavy-duty
ground wires. The claims are that these increase horsepower and make you taller,
more handsome, and appealing to women. The kit claims to improve power by
increasing current flowit is also complete BS. Well, no, I take that back. It will
increase power, by about 0.1 hp. But since the wires weigh more than the stock
wiring, it actually slows the car down because of the additional weight.
Do not buy this product. As an electrical engine engineer, I can tell you that
your stock wire harness works just fine. If the aftermarket kit works at all, its
only because it inadvertently fixes a bad connection. A properly working stock
harness is fine.
This is one of those marketing campaigns that counts on the phrase, Adds up
to 10 hp. As I rant about elsewhere, this increase includes zero.
Heres the finished ECU panel in place, mounted on the passenger-side engine bulkhead, behind the
passenger seat. It seemed like the logical place, near the engine, yet away from the heat. It was close
enough so the stock harness wires could reach, and in fact, some were shortenedmore weight
saved.
With the dash and ECU taken care of, it was time to wire the rest of the car. The battery is in the
passenger-side foot well, with the Painless Wiring fuse box immediately above it. The coiled wire
bundles from the Painless kit were already attached to the fuse box, so now it was just a matter of
running them where they needed to go. As delivered, the harness wasnt intended for use with an EFI-
style car, but it didnt really matter. The fuse box was located so it would be easy to replace a fuse,
instead of the gyrations necessary in modern cars when trying to get at the hidden fuse box.
Later, the battery above was replaced with a smaller, lighter battery. Since I didnt have a big
stereo or anything else other than lights and the ECU, it didnt need a huge reserve, just enough to start
the car. The stainless box was both for my safety and to meet SCCA rules; it would prevent the
battery from becoming a projectile in the event of an accident.
The thick main battery cable was purchased from the local welding shop. You can also check your
local home improvement outlet; its probably cheaper than a car parts store.
The big mess here is from the testing of the taillight circuitry (Id sat everything on a plastic
trashbag so it wouldnt short out against the chassis). This test was necessary because it wasnt
obvious which wire controlled which filament; each taillight has two different brightness filaments. It
was exciting to see the taillights light up when the turn-signal switch was flipped or the brake pedal
was pressed.
With the taillight circuit figured out, it was time to wire the engine cover. Notice the battery inside
the shell, used for double-checking each circuit as it was added. Once the harness was known to be
good, split harness covering was added, which allowed adding a wire later on if I goofed. The
harness assembly was wrapped with electrical tape to keep out water, stones, and dust, and it was
attached to the composite using adhesive-backed wire-tie anchors. The regular adhesive would never
put up with the heat; it would dry out and fall off in no time. The adhesive was removed and the
anchors were roughened and epoxied to the composite.
When selecting taillights, theres a good reason to use the new LED type. Sure, they last a lot
longer (and are more expensive), but the real benefit is that they turn on faster. An incandescent lamp
takes about 1/10 of a second to light, while the LED is virtually instantaneous. That doesnt seem like
much, until you realize that at 60 mph, the SUV coming up behind you is traveling 88 ft./sec. Having
your brake lights come on 1/10 of a second sooner means that the Soccer Mom whos on the cell
phone will start slowing down nearly nine feet sooner (assuming she even notices). That could be the
difference between getting rear-ended and only having an embarrassed driver in the mirror.
Some people use a switch that is activated by brake fluid pressure to turn on
the brake lights. I didnt, because I felt it wasnt any simpler; it has to be plumbed
into the brake system. Later, I learned that they also turn on slowly (only after
theres a certain level of pressure in the line). For the same reason that its a
good idea to use LED lights, I didnt want any delay in telling Ms. Soccer Mom
to back off.
Since LED taillights sold specifically for the Mini cost $150(!), another solution was needed.
There are screw-in LED replacements for incandescent bulbs, but I wanted a light higher up, up
where SUV drivers were more likely to see it. I used a center-mounted brake light assembly from
Signal Dynamics (www.signaldynamics.com) and a Kahtec controller
(www.kahtec.com/smart_stop.htm) that blinks the LED lamp six times when the brakes are applied,
then holds it on steady. That seemed like a good thing to havea light up high, blinking out, saying,
Hey, Im down here!
Next was wiring the nose section. This was pretty straightforward, just connecting all the lights.
Theres the same battery again, being used to test each circuit as it is connected. Also scattered about
is shrink sleeving, a heat gun, crimp terminals, and a crimper.
In the same way as the engine cover was wired up, so was the nose. The wires were cut to length,
connected, tested, pushed into the split harness covering, and then taped up. The little black squares
are tie-wrap anchors epoxied to the shell. The plastic headlamp assemblies are nicefrom
minimania.com. Theyll never rust, and Ive since seen them being used in American hot rods.
The little dongle thingy at the bottom is a circular connector. It has to hang loose because when the
nose is installed, the connector has to be plugged in before the nose is fully in place, so it needs a bit
of extra length. A service loop is what we call it in the business.
A rant about fuses
Fuses dont go bad, yet I hear all the time, I popped a fuse so I need a new
one. No, a fuse opens because too much current ran through it. The hapless fuse
is waving its little arms, trying to tell you something: you have a problemfix it.
The blown fuse means it did its job, which is to protect the wiring from melting
or catching fire. Having the fuse blow buys you time to correct the problemthe
same problem that will blow the fuse again if you simply replace it. Check the
circuit! Either a wires shorted to ground, or theres simply too large a load for
the fuse to handle. Dont put in a larger fuse until youve accounted for the higher
current and determine that the wires can handle the additional load.
Headlights: 5 A each, figure 10 A per pair
Running lights: 2 A each, so add them up, probably around 10 A total
ECU: Hard to say, probably 510 A
Cooling fan: 2035 A, and more at startupuse a relay
Fuel pump: Depends, 220 Ause a relay if over 10 A
Coolant pump: 520 A
28: Shifter
In the Honda Prelude, the shift cables run from the stick shift on the center console, curving
smoothly through 90, to the actuators on the transmission, three to four feet total.
In a mid-engine application, the FWD assembly is located behind the driver. This places the
transmission shift actuators behind the driver, and more annoyingly, behind the engine and
transmission. The factory shifter wont work now, since the cables are intended to run out the front,
and of course, the cables are now way too short.
As always, there are several approaches. One is to use the factory shifter and get longer cables.
This isnt a bad idea, because it avoids sinking a lot of time into making a shifter from scratch. The
problem is figuring out how to bend the cables, exiting the front of the shifter, to head toward the
back. High-quality push-pull cables cannot be bent in a tight radius turn.
Another idea is to use bell cranks to change direction; this way the cables dont have to be tightly
bent. The trouble is that theres a 180 turn at the shifter, a 90 turn to head up behind the engine, and
another 90 turn to face the shift actuators. Using even one bell crank with each cable means that there
are now 12 additional pivot points that add their own play into the system. There will always be two
actuators, be they cables or rods, one selecting the gear row and the other the gear column.
Some people replace as much of the cable as possible with rigid rods or tubes. This absolutely gets
rid of any slop, though going around corners is still an issue. Bell cranks can be combined with shift
rods to completely eliminate cables, but be aware that the engine and transmission move around under
braking, acceleration, and cornering. If shift rods are rigidly mounted, which they will be, the
drivetrain will likely pull itself out of gear as it moves around in its engine mountsnot good.
Probably the best solution is to combine the two, using rods to get as close to the engine as
possible, then switching to cables for a couple of feet to handle drivetrain movement.
One problem with push-pull cables is that they introduce play, but not in the push-pull direction as
expected. There is an inner cable inside a heavy-duty outer shield; the problem is that the two arent
bonded together. Everything is fine as long as the cable is straight, but thats not why we buy push-
pull cableswe buy them to go around corners! Once bent, as one end of the cable is pushed, the
inner cable shifts sideways inside the outer sheath. Each 90 bend introduces about 1/8 of play.
Other push-pull cables appeared to be made similarly, so I went ahead and used what I had. Yes,
there is a little dead area in the shifter, but no more than in the factory shifter.
After I designed my shifter, I saw how one kit-car company dealt with shifting its mid-engine car.
They used the stock shifter, simply turning it around to point toward the back of the car, connecting it
to pushrods. Then, behind the engine, they used two bell cranks to correct the shifter motion (to get
left/right and up/down correct) and fed this motion to the stock shift cables. The advantage is that an
entirely stock shifter assembly can be used. The downside of bell cranks is that each one introduces
its own play into the system. But this approach is probably the best solutionpretty clever. So what
did I do?
I started with the factory Honda shifter; it had really nice bushings and a two-axis ball mount; there
was no point reinventing a really nice shift mechanism. The shifter assembly has two cables operated
by the shift lever. One cable is activated based on the front-to-back motion of the shift lever, while
the other moves with its left-and-right motion. Each cable does its own job at the transmission, one
lever selecting the gear row (1-3-5 or 2-4-Reverse) and the other selecting the gear column (1-2, 3-4
or 5-Reverse).
The two cable motions combine to select any gear in an x-y fashion. The trouble, of course, is
that the cables are set up to head toward the front of the car and not the back.
Simply turning the shifter around to point toward the engine inverts and reverses the shift pattern.
The task was to draw out all the combinations of the shift actuators, drawing all of the positions of the
two levers, and recording what gear it resulted in. That done, the motions needed to make it all
correct had to be figured out. Since the shift pattern was now upside-down and reversed, I wondered
if I could turn the shift assembly over and use it upside down. That looked promising, so the shifter
assembly shaft was cut off and welded back onto the bottom of the assemblypretty drastic-
sounding, but it worked fine. I used Cablecraft cables, which have worked very well and are
available from www.push-pull.com.
Pushrods can be a bit iffy because the rods need to dip down to get under the
header and oil pan, then back up to the bell cranks. This means they must be bent,
which means theyll flex, giving a less-than-perfect feel. Its always something.
In Staniforths High Speed, Low Cost, he tried a hydraulically activated
shifter. The stick-shift end was coupled to two small push-pull cylinders, with
the same setup at the transmission end. The four hoses running between could be
run without regard for going around corners. He said, after tons of effort to get it
to work at all, that it ended up working too well. Due to a lack of feel, it was too
hard to get into gear. This was the kiss of death during competition, the whole
mess later being replaced with pushrods and bell cranks.
If you want to see what a bell crank is and how it works, heres a great
working model: http://www.flying-pig.co.uk/mechanisms/pages/bellcrank.html.
Recommendations
I feel that the best approacha balance between function, time, and costis to use a complete
stock cable shifter assembly as is. Turn it around so that the cables point to the rear of the car. Doing
this inverts the shift pattern, so it will have to be inverted again at the transmission end of the cables.
Normally, the cables approach the transmission levers from the left if the transmission is on the right
side of the car. If the cables are removed from their mounts and moved to the right, the shift pattern
will be inverted again, making the stick shift pattern correct. In other words, instead of pushing and
pulling, they are now pulling and pushing.
I feel this is the fastest, most straightforward way of shifting a FWD drivetrain moved to the mid-
engine position. Of course, longer shift cables and a new transmission mount for the other end of the
push-pull cables will be needed, but usually there are existing bolts nearby that can be used to mount
an adapter bracket.
29: Steering
A stock Triumph Spitfire rack was used. It has a 9/16 splined steering shaft, but most domestic
aftermarket steering systems are based upon a 36 spline. Fortunately, adapters are available to
convert from one to the other.
I found that the splined shafts caused problemsnot so much the shafts themselves, but the U-
joints. It seems like everyone sells U-joints that are held on with a setscrew, which never really
worked for me. The manufacturers recommend drilling a slight recess so the setscrew has somewhere
to seat, but no matter what I tried, they always rocked slightly. Surprisingly, this was even true when
both the shaft and U-joints were purchased from the same manufacturer. What works much better are
clamp-on U-joints, which unfortunately are difficult to find, expensive, large, and heavy.
A local hot-rod shop owner said that most shops dont even use splined shafts anymore; they now
use double-D shafts. Well, thats niceafter Id spent money for a splined setup.
A quick-release steering coupler was also used, helpful for getting in and out of the car (though
normally I leave it attached). When its really handy is when I have to work on the pedals; getting the
wheel out of the way is a big help. Prices for these range from $30 to $200, and you get what you pay
for. I got mine from a circle-track parts supplier. While many makers claim their units dont introduce
play into the steering, mine did. Its not muchin fact, its less than the U-joint splinesbut its still
irritating. Supposedly, the $200 couplers are made differently and dont have any play at all, but I
didnt want to spend that much money to find out.
The splined coupler that comes with the quick-release is meant to be either
welded or pinned onto the steering shaft. On the other hand, U-joints, regardless
of type, should not be welded.
Most have needle bearings and grease as part of their assembly; welding will
cook the poor thing (if not accidentally welding the bearings, too, by passing
welding current through them). The right way is to drill through the U-joint and
steering shaft and use two press-in pins at 90 to each other, about 3/8 apart. Be
sure to think through the sequence of assembly so you dont end up with a steering
shaft assembly that cant be removed or disassembled! Several U-joint
connections will have to be clamped on to allow disassembly.
Play can be introduced not only by the splined U-joints and quick-disconnect coupler, but by the
rod-ends as well. Many vendors sell oversize rod-ends made specifically for use with
steering shafts. These special units are slightly oversize so they dont impede steering-shaft rotation.
The trouble is, theyre a bit too oversize, contributing their own share of side-to-side play. I had to
stick some shim stock between the shaft and the rod-ends to eliminate the play. Tanner
10
found a
Teflon pillow block that worked well for him. Doing it over again, Id give those a try.
Speaking of doing it over again, Id still use clamped U-joints; they work well, if you can find
them. I dont know if switching to the double-D type shaft and U-joints necessarily guarantees
elimination of all play, though. If I cant find any, I will put a touch of liquid thread-locker on the
splines to deal with it. Getting them off will be less pleasant, though.
Make sure to route the steering shaft so it is out of the way of your feet. With my setup, the shaft
comes in high to miss my feet, then turns and drops down to the rack, forward of the firewall. This
required two rod-ends to fully support the upper-shaft section. Note that there are two collars above
and below the top rod-end. These prevent the shaft from being pulled out of the U-joints and from
intruding into the passenger compartment in an accident.
I count on the fact that the steering-shaft segments form a Z to prevent the upper section from
injuring me in the event of an accident. Actually, Id like the steering wheel to be a bit further away.
The 6-bolt/3-bolt steering wheel adapter (from ioportracing.com) moves the steering wheel about
closer, which isnt welcome. Be sure to take this into account when calculating where to place the
steering wheel components.
The steering shaft has to pass through the foot well to get to the rack. I figured the easiest way was
through a sheet of rubber, so a piece was cut to size and screwed to an oversize hole, necessary to
fish the large U-joints through.
Regarding the steering wheel, its not just a steering wheel; its a very visual, tactile componenta
very personal choice. It needs to be small enough to miss your knees during shifting and braking,
when your feet come up off the pedals. Make sure that it will mate with the steering wheel quick-
release mechanism if you use one. I bought a Momo wheel.
30: Instrument Gauges
Autometer Ultralite gauges were used, and Im very happy with them. They worked much better
than the VDO gauges I used years ago (though I understand theyve changed the design recently).
I made the mistake of buying an ordinary air/fuel ratio metera complete waste of money. Thats
because of how air/fuel ratio is measured with a typical O
2
sensor. The sensors are very sensitive to
O
2
changes, but only around 13:1. Above and below this point theyre pretty useless. What this means
is that the typical LED A/F meter will show its lights bouncing up and down as the ECU continuously
rebalances the air/fuel ratio. In other words, the meter makes a great light show, but thats about it.
There are small dash gauges that are wideband, and while theyre much more expensive than the
cheap ones, theyre far more useful. Since the car would be driven on the street, it needed a
speedometer. Its about the last thing I look at, but the law says I need one. I wanted a large tach, yet
had minimal space, so a combination unit was found.
This unit (at left) is from SPA, available in several different rpm ranges. The LCD display can be
set to display many different parameters, and normally, I set it to display speed. Its a really cool unit
for many reasons. It uses a stepper motor for accuracy and is microprocessor-controlled. Everything
is adjustable, including tire circumference. It features a three-stage shift light and telltale rpm memory
and can measure 060 mph time and quarter-mile timevery cool.
The problem with a set of analog gauges is that, well, they arent looked at very often. In the heat
of battle on a track, there simply isnt time. During autocross events I have no idea if I run low on oil
pressuretheres simply no time to look.
While I should have an oil-pressure-activated light to indicate trouble, I
havent gotten around to it. A common method is to rotate all the gauges so that
when theyre at the proper value, all the needles point in the same direction.
Trouble is, I still wont look at them mid-corner.
Doing it over again, I would consider an electronic all-in-one unit. The downside is that theyre
expensive, but they have many advantages, the biggest being that limits can be applied to all the
readingsoil temp, oil pressure, voltage, and coolant tempand can be summed up into one big red
light. Now you dont have to look at anything other than the tachometer. If theres any problem with
any engine parameters, a red light will let you knowright now.
These all-in-one systems below are pretty nice and somewhat reasonably priced; no doubt there
will be better and cheaper units in the future. At upper-left is a unit from racepak.com; upper right,
etbinstruments.com; and below, two units from aimsports.comall expensive, but very nice all-in-
one solutions. (Images credit of the respective manufacturers.)
31: Battery
The size of battery required depends upon the expected continuous load from the battery. Since the
car was far simpler, both physically and electrically, than the original, battery capacity was
considered before choosing which battery to buy.
Most cars specify a 55 amp-hr battery, but lets review what its needed for. Once the engine is
running, the alternator supplies all the power needed. The battery is really only needed to operate the
ECU, starter motor, and fuel pumpswhen the engines not running. In my case, there was no big
stereo or rally lights or any expected need to run off the battery for long periods of time. The ECU
needs about 10 amps, the pumps maybe another 10 amps, and the starter motor a few hundred amps
while cranking so a surprisingly small battery will work just fine.
Researching various batteries showed quite a few candidates, but digging deeper through the
advertising hype revealed that most manufacturers put their own labels on the same battery.
That selected battery (right) was made by Odyssey (www.odysseybatteries.com/). They have
excellent on-line data regarding their different units, and based upon my current requirements, the PC-
680 was chosen.
At 14.5 lb., 600 cold-cranking amps, 17AH, and with a small 7 x 3 x 6.5 size, it could be
mounted nearly anywhere. Being a dry-cell battery means there is no acid to spill, or hydrogen to
vent, or a need to be mounted upright. Because of these features, it was mounted on its side, at the
front end of the passenger foot well inside a stainless box.
In over a year, the battery has never left me stranded and has plenty of reserve, never requiring
additional charging. I found one on-line for about $75.
Regarding overall system current consumption, consider swapping alternators if the project car
uses less power than the donor. For the Honda H22 engine, it turns out that a Honda D-series
alternator, weighing about four pounds less than the H-series, plugs right in. The heavy cast-iron
alternator mounting bracket can also be replaced at the same time, with a lighter fabricated piece.
32: Seats
I strongly recommend sitting in any seat youre considering buying. Theyre just too dependent
upon your own physical shape for every seat of size medium to magically fit perfectly. I bought
Cobra seats, not because I thought they were the best ones around, but because I found a place that had
them in stock, and they fit me. Recaro, Sparco, Cobrait doesnt really matter, as long as they fit you
and the car.
A well-fitting race seat is a surprisingly tight fit. The best ones require you to have to wedge
yourself into it sideways, then rotate back to facing forward. Its job is to hold you tightly so you dont
have to brace yourself using the steering wheel.
My brother ordered his Stalker kit car with the factory-supplied seats, figuring
theyd be fine. Unfortunately (but very predictably) they werent comfortable. He
ended up having padding added. Its a reason to order them locallythey can be
tried out first.
One small goof I made during component placement was not wearing my helmet while positioning
the seat. The headrest pushes my helmet forward slightlynot a big deal, but Im always aware of it.
It puts my head a bit closer to the steering wheel than Id prefer. Being such a small car, though, meant
that there simply isnt much room. With an 80 wheelbase and everything else being placed where it
had to go, the seat ended up in the only position where it fit.
When you design your car, wear your helmet when deciding seat inclination; Id like to have a bit
more seat inclination, but thats how it goes.
Instead of working inside the garage all day, when are you going to spend some time with me?
Cooper asked me this question many times when I had the garage side-door open in the
summertime. In the background is the composite shell under its cover.
33: Exhaust Header
Making the header was a very enjoyable side project, and heres how to make your own.
First, a header design is needed, and by design, I mean determining tubing diameters and lengths
before ordering parts. I found two places that design headers: headersbyed.com and
burnsstainless.com. In both cases, the customer fills out an extensive questionnaire involving many
engine-specific parameters. This is another reason its so important to have the factory service
manual on hand.
Given all the information, the above companies produce a report specifying all tubing diameters
and lengths. This gives enough information to order the primary tubing material, collector, and
collector flange.
Header tubing material
Header material ranges from mild steel to 304 stainless, 321 stainless, Inconel, and titanium (never
aluminum!). Mild steel is the cheapest, while the last two are very expensive, so pick your financial
tolerance level. Each has its own cost/benefits:
Chapter 33: Table 1
Based upon the design specs, get a rough idea of how much tubing will be needed. I ended up using
nothing but U-bendsno straight sections at all. What straight sections were needed came from the
cutoffs from the U-bends. I went for broke (literally) and used 321 stainless, ordering it from
Washbon in Newport Beach, CA, (949) 548-9783 (they dont have a Web site). Washbon had slightly
better prices than Burns Stainless, but if you have them do the header design, maybe theyll cut you a
deal.
For 304 stainless and mild steel, try store.racing-solutions.org or magnumforce.com for great
prices. The number of U-bends depends entirely upon your particular application, so you either have
to guess, or calculate how many are needed, then order a couple of spares.
Mock-up
Building a mock-up is a very important step that prevents wasting expensive material, if thats what
the finished product will be made from. A mock-up will pay for itself the first time you cut one of the
expensive tubes too short.
Rigidly fix the collector where it will mount, relative to the engine. If there are any obstructions,
make sure they are represented, too. I temporarily welded the collector to a bracket and bolted it to
the block so it would remain exactly where it had to be throughout fabrication.
Use 3/8 copper water pipe, the type used to plumb ice makers. Cut lengths to match the primary
lengths specified by the design specs, one for each cylinder. Cut a piece of wood to represent the
head flange, leaving the holes filled except for small center holes.
Routing the tubes
Since the copper tubes are now exactly the correct length, its impossible to screw up, right? Find
out from the tubing manufacturer what the radius is of its U-bends; in my case, they all had a
centerline radius of 6. In order to make the copper tube bends honest, make a tool to bend the
copper pipe around. This can be as simple as a wood form, with a diameter of 6 - (0.375/2) (for the
copper tube centerline radius). Using this to bend the copper tubing ensures that all bends are honest,
so they have the same radius as the final tubing; otherwise, its all a big waste of timemake sure to
get the bend radius correct!
Now, go to town bending the copper tubes with the tool and routing them as you see fit. Some
books say that for good gas flow, it is best if the header tubes come straight out of the head for a few
inches before turning, but they dont literally mean that. What they mean is, make the header tubes
come out of the head in the same direction as the ports for a few inches. Since some ports come out
at an angle to the head, its not necessarily the same thing.
After routing the tubing, cut paper discs of the same diameter as the primary header tubingone for
each cylinder. Cut a hole in the center of the discs and slip them over the copper pipes. Slide the
discs along each tube to ensure that at no point do they touch each other or any part of the engine,
engine mounts, alternator, or any other obstruction. At this point, it is very easy to fix problemstake
advantage of it!
When this is done, there is a decision to make: will the copper tubes be used as templates to cut the
final tubing? If the header will be mild steel, the answer is yes. But if the header will be made from
expensive tubing, you may wish to make it from mild steel first. I had two strikes against me: Id
never made a header before, and I was using expensive 321 stainless. I was a big chicken, unsure
how the stainless would cut on my little band saw, how it would weld, and unsure if I could keep the
cuts square to the tubing. To practice, I bought mild-steel bends to learn on first.
This worked out very well, even though it cost a bit more. Once the mild-steel
tubes were perfectly sized, they were used as templates for the 321 material.
With all the copper tubing positioned correctly, its time to transfer the bends over to the mild-steel
tubing.
Here you can see the first primary tube in place. A copper tube is removed, and the start and end of
each bend marked on the mild steel; these are the cut marks. Cutting the mild-steel tubes on the little
band saw went fine, and once all the segments were cut, they were connected with hose clamps
(thanks to Alan Brickey for this tip). Cuts that werent exactly square were touched up on either the
grinding wheel or sanding disc. Holes were punched in the hose clamps, so once the segments were
properly positioned, they could be tack-welded in place. (I highly recommend TIG-welding the
header.) Now, each segment is rotated (clocked) to make the primary tube assembly best replicate
the bent copper tube its replacing. When it is deemed as good as it can get, the segments are tack-
welded together, and the clamps are removed and transferred to the next assembly.
Each remaining primary tube assembly is made the same way: the copper tube is used to transfer
cut marks to the U-bends, the tubes cut and clamped together, and each segment clocked until the
assembly copies the copper tube. Always make sure each assembly still misses all obstructions, then
tack-weld them together. Shown here are all four mild-steel tubes in their final positions. Since the
final product would be 321 stainless, each tube assembly had index lines marked at each junction,
each segment was numbered, and then the tack-welded sections were broken apart. Since all the fit-up
work and effort were spent first on the mild-steel tubes, transferring the patterns to the 321 stainless
went much faster. It was a pleasant surprise how easy the tubing cut. Part of it may have been that the
stainless was only 0.050 thick, or maybe it was because a new band-saw blade was installed just for
this job; whatever the case, it cut very easily. As with the mild-steel parts, any cuts that werent
square were touched up on the sanding disc.
When welding stainless, be sure to either back-purge the tubes with argon or
use Solar-Flux paste, which is available from welding stores.
A company that had just released a new header for Hondas provided one of the
biggest piles of BS Id seen: It is so efficient at extracting exhaust that it actually
lowers the oil temperature of the engine. Pretty amazing claimmade much less
so after it was clear the exhaust existed out the side of the engine compartment.
They had rerouted the exhaust so the 600 pipes no longer ran under the oil pan.
Do you think that might be the real reason why the oil ran cooler?
And here they are, all four primaries assembled and polished. When this was taken, they werent
yet welded into the cylinder head flange or collector. The cylinder-head ends of the primaries were
flattened slightly to fit the elliptical exhaust ports and tacked in four places. Then, the tube ends were
hammered outward to the flange using a brass drift crude but effective.
Here, the first two primaries of 321 are tacked in place. Its now a relatively simple task of
replicating the mild-steel tube assemblies. After all the primaries are tacked together, they are
removed and fully welded. Each was then individually polished using Eastwoods buffing wheels and
polishing compound. Polishing them now is much easier than trying to do it once it is all welded
togetherit would be hard to get between the tubes.
The cylinder-head flange is welded on both the inside and outside to aid in heat transfer from the
tubes to the flange. It also ensures that the hot exhaust gas doesnt hit the end of the primary tubes
head-on. Welding both sides of the flange also minimized warping but didnt eliminate it. I had to
grind the flange on a large 20 sanding disc to get it flat, even after bending it as straight as possible
using a 10-ton press.
Here is the intersection of the four tubes, sealing the opening between them with a cone; its not
pretty but it got the job done. Looking at this picture now, it looks like I used mild steel instead of
stainless. Thats not good, since itll eventually rust through. Another way to plug up the intersection
is by stretching the common quadrant of each tube so they all meet at the center, but this way seemed
easier.
The collector is welded on last. Note that the areas in which the four primary tubes touch each
other have already been welded; this ensures a completely gas-tight seal, once the collector is slid on
and welded. People asked why I welded the collector on; its to eliminate exhaust fumes circulating
inside the engine compartment. I also think its quieter without the leaks and rattles.
Heres the finished product from a very rewarding project that didnt take too long and where
progress could be seen at every step. If I were doing it over again, Id use mild steel. I used 321 both
to prove to myself that I could do it and because it would stand up to the high heat from being
wrapped. Wrapping a header is widely regarded as a bad thing; it literally cooks the steel and leads
to cracking. I simply didnt have much choice. It has to be wrapped since the header is so buried in
the engine compartment. While it could be ceramic coated, I dont believe that the coating keeps the
surface temperature as cool as being wrapped, but I could be wrong. Another issue with the coating is
that if the header must be repaired, the coating has to be ground off, and once repaired itll rust unless
recoated. (Yes, most stainless steels do rust in the weld zones, due to the chromium being driven
out.) In any case, Im very happy with how it turned out, and it was a very fun project.
Mild steel: Lowest cost, easy to weld, but must be coated or it rusts.
304 stainless: Next least expensive, fairly easy to weld, does not need coating.
321 stainless:
About 3x as expensive as 304, but stands up to very high heat much better
(turbos)
Inconel/titanium:
Top-of-line and very expensive. Very thin material can be used, resulting
in a very light assembly. Used in the exhaust section of jet engines.
34: Engine Modifications
Few changes were made to the Honda drivetrain; there would be time for that later. While the
header certainly improved power, it was fabricated simply because there was no other way to get the
exhaust outI didnt want to run it under the pan and out the back. This also applied to the intake
tract; obviously, the stock Honda intake wouldnt just work in a mid-engine Mini!
There were a couple of parts, though, that were deemed absolute requirements: a limited-slip
differential and an aluminum flywheel. Then there was the fiasco surrounding aluminum crank pulleys.
Limited-slip differential
Adding an LSD to my old Datsun 1200 was the single best improvement I made to the car. Only
with the LSD could I get on the gas early coming out of turns, and only then could the power be put
down without smoking the inside rear tire. Being able to accelerate so much earlier improved my lap
times immediatelyall without changing anything else. For this reason, it was at the top of the list
when searching for drivetrains.
I had heard so many good things about the Quaife LSD (like a lifetime warranty) and decided to get
oneI havent been disappointed. It is virtually invisible in operation, with no noise or scrubbing
tiresit just works. Its like driving an open-diff car on track, with the simple exception that the
inside tire doesnt spin helplesslyIm a very happy customer.
A competitor to traditional LSDs is the controversial Phantom Grip. It claims
to work just as well as traditional LSDsat a fraction of the price. While some
people are happy with it, others are not. Unlike the Quaife, there are anecdotal
stories of the PG working greator destroying differentials. I asked myself, why
risk it? The so-called savings becomes a lot smaller, too, after labor is figured
in, so is it really worth the uncertainty? Not for me. Is it possible that it works
just fine, providing a function at a fraction of the expense of other units? I
suppose, but I guess I dont believe in something for nothing.
Aluminum flywheel
An aluminum flywheel provides a real, measurable increase in power, all without the engine
working any harder. How is this possible? It does so not by producing power, but by not absorbing as
much. With the heavy-stock flywheel, it takes horsepower to spin it up to speed along with the rest of
the drivetrain. Reduce this rotational inertia, and now that power can be applied toward speeding the
car up instead.
Why would Honda use such a heavy flywheel? It must be there for a good
reason.
It makes it easier to start off from a dead stop. It makes it easier to release the
clutch when engaging first gear. Can a driver change his driving style slightly to
use the lightweight flywheel? Absolutely.
I replaced my 20 lb. steel flywheel with an 8 lb. Fidanza unit. The power savings isnt the 12 lb.
difference in weight; its much more than that. Its where the mass centroid is on the flywheel that
matters, the difference being far greater due to it having to be spun up along with the crank. Just
replacing this one part can free up around 5% power in first gear (less so in higher gears). A power
increase of 5% is a huge amount, especially one that doesnt make the engine work any harder. A
secondary benefit is that it lightens up the car by 12 lb., nothing to be sneezed at. If it seems like I
really like it, I do; its one of the few changes you can do to a car that has virtually no negative side
effects but produces real power gains.
Aluminum crank pulley
The popular opinion on the street is that aftermarket lightweight crank pulleys cause bearing
failure. So what is this stock rubber-lined pulley for? What does it do? One theory says its to filter
out the torque pulsations from individual cylinders power strokes, with the assumption that at certain
speeds the crank will resonate (like a bell) and start flexing and vibrating, damaging bearings and
eventually breaking the crank. The rubber in the factory stock inertial dampener mechanically filters
out this resonance. The aftermarket part doesnt have this rubber ring, so it cant filter the pulses.
The other theory, perpetuated by the aftermarket lightweight crank pulley manufacturers, is that the
rubber is there only to filter out the instantaneous crank-speed variations from traveling through the
accessory belts. They claim its there only to filter out an unpleasant noise generated by the accessory
belts. They say if the noise doesnt bother you, theres a 610 hp gain to be had by reducing the
rotational inertia of the stock unit.
On the other hand, heres what AEM (an aftermarket manufacturer of harmonic dampeners) says:
Vehicle manufacturers have thoroughly investigated the use of torsional vibration dampeners,
and conclude that their use significantly reduces NVH (noise, vibration and harshness).
Furthermore, manufacturer and independent research shows that torsional vibration dampeners
are critical to engine life where long-term use is intended. Removing the torsional vibration
dampener from an engine can be highly detrimental to its long-term operation!
The first part agrees with the aftermarket manufacturers pointthat they are just for noise control.
And the last sentence indicates that the engine will wear out faster, but what part of the engine?
Instead of relying on the endless anecdotes about what they do, I contacted someone inside a major
automotive manufacturer and asked him. Why not ask those who know rather than endless
speculation? Heres his reply:
I talked to a factory engine builder who builds high performance V8s. He occasionally builds
engines without vibration dampers, but only for hill climbing and sprinting. He recommends that
the crank be replaced after no more than four seasons racing due to the fact that it will eventually
fail. The overall mileage covered is therefore very low over four seasons.
One of the most interesting facts is that <mfg name removed> will be using the production damper
on their new V6 racing engine. The engine is based around a production block and crank and will
be used for circuit racing. The mileages will be higher than sprinting but not so great. The
engineers think that the damper can be improved slightly (slightly reduced mass), but not deleted
altogether. The pulses through the crank from each cylinder are what are being damped as you
probably already know by now. A failure related to the damper will result in broken cranks or
possible pressure-plate damage, rather than bearing damage. I think any failed components are a
result of general abuse and not down to the vibration damper. I think that it is clear that any
vehicle that will cover moderate mileages should have a vibration damper to prevent the risk of
crank failure.
And there we have it. I have not installed, nor do I plan to install, an aluminum crank pulley.
Accusump
While the oil system doesnt directly improve performance, its job is to ensure that the engine
always receives a steady supply of oil. A dry-sump system is the right way to deal with this (and cuts
down on windage losses, which frees up power). I just couldnt afford the $2,000+ for a system.
An Accusump (accusump.com) is a much cheaper alternative. It is simply a cylinder with a free-
floating piston inside. Air trapped on one side acts as a spring, while engine oil is plumbed to the
other side. Oil pressure inside the engine pushes oil into the reservoir, pushing the cylinder and
compressing the air until a balance is reached. At this point, the Accusump is charged. If, during
hard cornering, the oil-pump pickup is uncovered, oil pressure will drop. The Accusump, originally
charged at a higher pressure, will start dumping its oil back into the engine, supplying all oil until oil
in the pan is picked up again or the Accusump becomes empty. The trick is the filling and emptying of
the Accusump.
Not wanting to fiddle with a manual valve, I bought the system with the electric solenoid valve,
wiring it into the ignition system so it opened whenever the ignition was hot. What was bothersome
was being in trafficcoming to a stop and watching the oil pressure slowly drop. The Accusump was
dumping its entire 23 qt. load into the enginethat couldnt be good.
I thought about switching the electric valve manually but figured Id forget to
turn it on when on track. Turns out, even Accusump themselves dont like the
electric valve for the trackread on.
Looking into how to fix this, I found comments on the Accusump Web site saying that the electric
valve setup was not intended for racing, but only for street drivinguh oh. I already didnt like how
it worked on the street, dumping its entire reservoir while sitting at a light. The second problem was
that they said the electric valve wasnt appropriate for racing. The reasons given were that a manual
valve is cheaper, simpler, more reliable, and allows higher flow. When I asked about this, they
explained that the electric valve had a small orifice, great for slowly charging the system, but it also
restricted how quickly oil could be supplied back to the engine. The good news is that Accusump
offers an upgrade for the existing electric valve systemcool.
Looking into what was supplied with the upgrade kit, though, confused me: they didnt offer a high-
flow version of the electric valve. How could this upgrade provide higher oil flow if the claimed
oil restriction (the solenoid) continued to be used? When I asked about this, I was told, I dont know
how it works, but it just does. Brother. So after looking into the system more completely, I concluded
that the upgrade did nothing and bought only the manual ball valve. The remote-actuated ball valve is
controlled by a push-pull cable routed through the center tunnel. Since the Accusump tank has to be
located near the engine (to minimize pressure loss), the control cable has to run near the header. Since
the shifter cables worked fine in the heat, I didnt think much about ita mistake. The Accusump-
supplied push-pull cable quickly melted and froze up. Nice. Something they knew would be located
near the engine failed in its intended use. The picture here shows a Cablecraft push-pull cable and
what was supplied by Accusump; guess which is which?
I now have a system that works, because I had to replace the Accusump cable that I paid good
money for. The valve is now opened from inside the car. It isnt quite a no-brainer, though, since I
have to remember to open the valve before heading out on track (a problem for me sometimes).
Whats really needed is a smart valve, one that reads oil pressure, engine
rpm, and maybe even oil temperature. Only then can the system really operate
right; anything else is a compromise. I could design and make one, but thats one
of those time-sucking side projects.
Lesson learned
I used factory engine mounts because of what I experienced on my Datsun
1200. Since I kept breaking the factory mounts in the 1200, I thought Id make
some solid onesthat didnt work so well.
I had just adjusted the valves on the Datsun, and part of that adjustment was to
rotate the engine with a large wrench on the front crank pulley. After completing
the job, I went about installing the solid mounts. That job done, I got in the car
and started it up. The very first thought in my head was, Crap, I forgot the
wrench on the front pulley! The vibration was so bad it blurred my vision.
Quickly shutting off the engine, I got out, expecting to remove the wrench Id
forgotten to removeonly it wasnt there. Nope, all that vibration was due to the
solid engine mounts.
Yeah, they may be great on a real race car, and right then I realized that I
wasnt prepared to take that stepthe stock mounts were immediately
reinstalled.
Fast-forward to now; I was unwilling to put up with that level of vibration in a
street car. I will admit, though, that they could be stiffened up slightly, perhaps
with inserts or by squirting some silicon-type goo into the stock mounts. I just
havent experimented with it yet.
35: Composite Work
Composite work was very much a love-hate relationship with me. While the shapes that can be
made are nearly limitless, getting there is a messy, smelly, dangerous, expensive, drawn-out, tedious,
and boring affair. But sometimes thats what it takes to get what you want.
Before going further, I feel compelled to mention the health aspects of working
with composite. About the time I finished working with the stuff, I developed
annoying (to say the least) asthma-like symptoms. Neither my doctor (Dr. Useless
I call him, but I digress) nor I know what caused it, but I suspect that it was either
the composite fumes, or more likely, the microballoons. Micro, as its called,
resembles cooking flour, but it actually consists of microscopic hollow glass
spheres. Micro is used to lighten an epoxy mix when used in a non-structural
application.
Whats bad about micro is that when its being removed from its bag, or when
its being sanded, it gets into the air. There are two ways it can mess up people
both bad. One is getting it in your eyesno doubt some of the spheres are
broken. Nothing like broken glass in your eyes to ruin your day. Worst, is getting
it in your lungs.
Cilia are tiny hair-like thingies in our lungs that sweep bad stuff up and out.
Microballoons are apparently too much for the cilia to deal with; being spheres,
theyre like ball bearings and just keep slipping back down instead of being
swept out. This means that the lungs are forever thinking theres something in
there they dont like but cant sweep out. A persistent congestion would be a
symptom (which I also have), where the lungs keep spinning their wheels
trying to push something out that wont leave.
Lastly, to remove epoxy from the hands, I know many people (me included)
sometimes use acetone. It has its own share of hazards: besides being very
flammable, it is absorbed through the skin and can go directly into the
bloodstream.
Yes, I used gloves, and yes, I used a real respirator (and not the weenie
disposable paint masks), but I didnt always use them occasionally getting a bit
lazy. Now, I get to contemplate if these lapses were the cause of some very long-
lasting health problems.
While I may never know the real cause of my symptoms, composite by-
products are a prime suspect. Use a proper mask! Use gloves! Always! (This is
another reason to have a separate workshop, away from the house.)
A year later, the symptoms still persist. You can either ignore this or learn
from it.
I know, personally, of two builders who were making great progress on their carsbuilding the
chassis, installing the running gear and electrical, finally getting them to the point that they did some
test drives around the block. Only after this did they start in on the composite workand virtually
disappearedbeginning endless months of sanding/filling/sanding/filling. It took them years, yet this
is what it takes to do it right. One builder eventually finished it himself, while the other builder burnt
himself out (or was on the edge of insanity), ending up handing it to a professional to finish it
properly. Each builder created a mold so multiple bodies could be made.
What do you care? You didnt have to mess with the shell.
Oh yes I did! The body was so rough that a layer of filler was needed to fix it
all. It is rather ironic that my lightweight carbon fiber shell is probably heavier
than a proper fiberglass version. Someone once said that I should have clear-
coated the carbon to show off the weave. Yes, it would have looked really nice,
if the body had been perfect right out of the mold, but it wasnt.
I guess that Im trying to instill fear for the Goddess of Compositeshe will draw you in with a
curvaceous shape and then cause you to spend an eternity sanding and filling, filling and sanding. Is it
hard? Not really, but it takes a really long time to do well; theres simply no way to sidestep it, since
the body is the first thing everyone sees. Every minor blemish will stand out, especially if the car is
painted a dark color.
The point is that unless you have the strength and patience to spend more than a year to work on the
body shell, I recommend buying an existing shell and building the chassis to fit; it will save literally
thousands of hours. Okaygloom and doom aside, there are many smaller parts that dont need the
perfection and detail of the shell. As an example of moldless construction, well make the air-filter
box.
The first step was to choose an appropriate air-filter element. Appropriate means one with
sufficient surface area thatll fit in the space available; in general, bigger is better. This is a K&N
filter, purchased from Kragen Auto Parts. The filter will be located on the right side of the car, behind
the header tank and ahead of the right-hand transmission mount. The aluminum frames hold the filter
one on top and one on the bottom. The frames will be bonded into the composite housing to provide a
rigid, accurate mate to the air-cleaner element. It has to be a tight fit to ensure that air doesnt leak
around the filter element. An added incentive is the fear that a small piece of sand could get past the
filter and roll into the engine!
Next step is to buy some foam blocks. It doesnt matter what kind of foam, as long as its rigid and
doesnt dissolve in the epoxy. Some people have even used the foam in rigid foam-wall insulation
from Home Depot or Lowes; I used urethane blocks from Aircraft Spruce. Whatevers used, test it
first to make sure the epoxy doesnt dissolve it. Here, the frame is pressed into a block of foamthe
lazy way of marking where to cut it. Then, a bunch more pieces were cut out and glued together.
Heres the assembly with the bottom piece of foam pushed into the aluminum housing; then more
foam was glued and stacked on top. Make sure to use glue that can be sanded; otherwise, keep the
glue away from any areas that will be shaped. Keeping the glue away makes it much easier to form,
since the foam is very soft compared to the glue.
The foam is then shaped into the appropriate shape. At this stage, it doesnt have to be exact, just
close. The only shaping tools used were a knife stolen from the kitchen and my fingers the foam is
that soft. I walked back and forth between the work table and the car about a million times to make
sure the fit was right, but thats what it tookwelcome to composite work!
Shown here is the lower air-box mock-up, in place to make sure it misses everything and points in
the right direction. Note that the outlet points to its destinationthe throttle body, at lower-left. The
stock rubber inlet hose will be used, since its already good at handling the vibration between the
engine and stationary air box.
Note that the throttle body inlet is taped over to keep foam (or anything else)
out. The foam particles get everywhere.
The finished form is wrapped with clear tape to keep the epoxy from sticking to the foam. (I tried it
one time without it, and it left a very rough inside surface and took a long time to clean up.) The
aluminum should be roughened up so that the epoxy sticks better (though I still added a couple of
rivets as insurance). This is the final form shape that will be covered in cloth.
The form is covered with fiberglass cloth and saturated with epoxy. Enough layers must be added
so that the overall form will be rigid. Its not a big deal if its found to be too flexible; additional
layers can always be added to problem areas.
After the epoxy has cured, the foam is dug out. It was fun digging the stuff out, watching the created
object slowly taking shape on the inside. After all the foam is removed, the tape can be peeled off the
inside surface and the edges of the composite trimmed to final size.
Heres the finished product. No, its not perfectno nice glossy external finish, no carbon finish,
just a functional component, and one that took about a day to make. Making this from aluminum would
have taken much longer, and the finished product probably wouldnt have looked as good. The shape
here perfectly fits the space available, doesnt weigh anything, and should last a long time.
Here is the air box, installed. This completed the downstream side of the air box, so the next thing
is to make the inlet side.
Using the same method as before, foam blocks are glued in the approximate shape of the cardboard
pattern, in the background. Note that the blocks fit inside an aluminum frame that will fasten to its
mate.
Here, the shape is approaching its final form. It has been shaped with the kitchen knife and my
fingers. The dotted line is the window opening, where air will enter the upper chamber.
The top half of the air box is set in position to doublecheck that its the correct shape. At this
point, it is extremely easy to change things; once laid up, though, while not impossible, changes
become much more laborious.
Deemed good enough, the upper plug is covered with several layers of carbon and set aside to
cure. Why carbon? No good reason other than looks. Im not impervious to bling.
Heres the air box from the other side, where it looks rather crude. Since the flat face will be cut
out, any effort to make it pretty is unwarranted. Again, many composite parts dont need to be perfect,
so dont waste time making them so. There are lots of other things to spend time on!
The two brothers meetthe fiberglass bottom and the carbon top. The aluminum frame was made
first, using the air-cleaner element as a template. Holes will be drilled in the edges of the frame for
screws.
Shown here, the air-box assembly is in place. The lower section straddles the large tube of the
chassis, mounted with two Rivnuts. Foam weatherstrip will be attached to the Lexan window to seal
the gap. As always, the actual parts are used due to laziness (I prefer to call it efficiency). Yes, the
whole mess could be modeled in CAD, but by the time all the dimensions would be entered, Id be
done. The point is, spend time efficiently, investing it where it gives the best return.
Note that the air box is mounted to the chassis with screws. Since the screws
are on the inside of the air box, it is very important that they never get looseits
a straight shot from there, down the intake manifold, and into the engine. The
screws were covered with silicon seal to keep them from moving.
Be aware that epoxy softens with temperature. If an epoxy composite structure
must stand up to high heat, epoxy intended for that temperature must be used;
otherwise, it will begin to soften. Soften, in this case, simply means that it will
no longer have 100% of its strength, not necessarily that itll collapse into a
sticky pile. Its not a big deal, as long as youre aware of it and design around it.
Since the air-cleaner housing isnt structural, it didnt matter if it softened up a bit
at high temperature. On the other hand, if youre building an all-composite
chassis, it can be very important.
36: Final Assembly
So here, finally, was a pile of newly painted parts, simply waiting to be bolted together into a car
at least that was the idea.
This point in the project was a monumental milestone. To have all the years of hard work done,
culminating in a pile of nice shiny parts, was pretty amazing. It was as if Id been climbing a huge
mountain for so many years that Id forgotten about ever reaching the summit. I greatly enjoyed the
climb the journeyfinally achieving the summit. Here, finally, were all the new parts laid out and
ready to go. In my mind, all the hard work was behind me. At this point, it had become a kit car,
something I could just bolt together, knowing everything would just fit almost. Of course, what I
was forgetting was that like a kit car, it never just goes together.
Thinking that the project was now reduced to a simple bolt it together and
itll be perfect scenario wasnt quite right. What I forgot were all the various
changes made along the way, changes having various implications to components
around them. None of the changes stopped the project cold, but it did mean I
wasnt off the hook for keeping the old brain working.
The first items installed in the new chassis were the coolant pipes; everything installed afterward
would make them more difficult to access. This was due to the aluminum bends that I welded on,
thinking it saved weight and complexity. While it did, time will tell how wise a decision it was. If a
weld cracks, it will be hard to repair.
Aluminum being what it is, welding wasnt such a good idea; using simple
straight sections and hoses to connect them would have been better. If a pipe
cracks, the offending tube assembly will likely have to be cut, to get it out of the
car, and replaced with a straight section. As another builder once said, Just
because you can do something doesnt mean you should.
While it was easy to still access, the bulkhead window was fit up. Best to get it sized and drilled
now instead of finding later that it wont fit with the engine in place. All of the Lexan windows fasten
to the chassis with #4 flat-head screws and Nylock nuts, for easy replacement. The holes are oversize
because Lexan expands a lot with temperature (like across this window!). In addition, no matter
how careful Lexan is cleaned, it does scratchscratch-resistant or not so they had to be easy to
remove.
Anxious to try my hand at riveting, a victim panel was chosenthe passenger foot well.
Structural Cherry Q rivets were used, and being aluminum, pulling them with the hand rivet tool
wasnt too badgreat fun actually. It was very satisfying to unwrap a brand-new painted panel, put it
in place, and see that all the rivet holes lined up exactly. Of course it wouldit was drilled at the
same timebut I always found it slightly wondrous that it really did fit.
About 600 rivets later, the stainless floor panels were in place. Here, Monel rivets were used
because I couldnt find stainless. Attempting to rivet these with the hand riveter was quite an
experience. They took so much force to pop that I was afraid the rivet tool would break, and
indeed, the handles bent. Then, they didnt exactly go pop; they went BANG, so hard that my
hands stung from the shock. Fortunately, a $29 pneumatic riveter from Harbor Freight worked okay,
though it wasnt too happy with the Monel rivets. It was always trying to disassemble itself due to the
tremendous shock. But it had no trouble pulling aluminum rivets.
The day the above picture was taken, my buddy Cecil dropped by to see how
things were going. Seeing that I was in the middle of riveting, he asked if he
could install one, just to see how they went in. After he did, he casually added,
You know what this means; it means I helped you build the car. In fact, I can
now claim I pretty much built it! I told him later I drilled out the rivet that he
installed, just to remove his participation. Hah!
As mentioned before, the main bulkhead contains the secret ingredient to make the car much more
pleasant to live with. Stainless paneling was on the back (engine) side, so there was no point in
painting it. On the forward side was aluminum, but the secret was insidehome wall insulation. It
did wonders for the car, keeping engine noise tolerable and the bulkhead cool to the touch, and
preventing the main bulkhead from drumming.
With the forward panel of the bulkhead in place, it was starting to take shape. The bulkhead panel
is in three sections because the angled door tubes prevented one large sheet from working.
Barely visible is a single hole at the top of the upper depression on the right. Its for the bolt that
holds the seatback. The bolt had to be installed from the engine side, providing an absolutely rigid
seat mount.
Paneling continued with the drivers side center-tunnel panels. At the outside bottom edge of the
chassis is the lip that the shell will sit on. I got an idea how strong structural rivets were when the car
later fell off the wood blocks and onto the floor. It fell right on a tool, putting a large dent in the
0.125 aluminum lip, but it didnt pull away from the chassis at all.
In this shot, the chassis looks a little like an aircraft fuselage.
The first components installed were the coolant pipes, followed by the fuel cell. As Ive said
before, it should have been installed from below, because if it ever has to come out, its going to be
very unpleasant work. The coolant pipes are in their proper locations, circling around where the
radiator exhaust duct will go. Rubber-lined Adel clamps are used everywhere to mount the aluminum
coolant lines (and just about everything else). Theyre fastened down to threaded studsscrews
welded head-down to the chassis tubes.
The steering rack slides in from the side. The rack mounts took a long time to position, to minimize
bumpsteer. The rack is one of the few parts not properly painted; instead, it was just shot with a spray
can since I didnt want to disassemble it. Its a full-length, unmodified Triumph Spitfire rack.
In goes the engine, and I hoped it wouldnt be removed for a long time. The roller stand is the type
of unit widely available for motorcycles and snowmobiles; this one came from Costco. The engine is
rolled in from the back, and then the left, right, and front engine mounts are attached.
With the engine installedand little elsethe entire chassis was precariously balanced on the
wood blocks under the chassis. I didnt realize how tipsy it was. From the build diary:
So I had the engine installed and had just pulled the engine dolly out. Unfortunately, with
nothing at the front end of the car, it was rather tail-heavy, but not quite enough to make it
obvious. About then my wife heard a tremendous crash. Coming into the garage she was greeted by
a long string of obscenities from her very pissed-off husband, looking at his chassis and engine
that had just fallen onto the floor. After making sure I wasnt pinned beneath it, she quickly
retreated to the house. Luckily, it crashed down onto a deformable structure some newly
painted brackets, which kept anything else from getting hurt. The carpet helped somewhat, too, so
there was only some scratched paintand a bruised ego.
Heres a bottom view of the engine tray just after riveting on the panels. From this angle its hard to
see, but the upper portion is angled upward (away from the camera), optimistically to become
something of a diffuser. The large slots in the panel are cutouts for the suspension arms. Doing it over
again, Id try to make the uprights such that the arms didnt cross the plane of the tray.
The stainless cable protector. The header will be immediately above this, running from lower-right
to upper-left, and all the control cables have to go below it without getting cooked. While it worked
okay, it was too small, some of the wires and cables getting nicely browned. A much larger heat
shield was added later immediately above this one.
Im considering using something better for insulation hereperhaps some sort of thermal insulating
tile.
The engine tray is finally in place, and I hope not to be removed for a long time. While it improves
airflow beneath the car, it makes access much more difficult. From this angle, its a bit easier to see
the panels upturnthe optimistically designed diffuser. Would it work?
Heres one of the rear-facing suspension brackets, attached to the bottom lateral tube running
across the base of the main roll-hoop.
While this worked fine, it made it difficult to fit the stainless panels around the brackets. Doing it
over again, Id make bolt-on suspension mounting brackets, so theyd attach after the paneling is
installed. In fact, the double row of rivets to the left wouldnt have been needed if bolt-on suspension
brackets had been used.
Heres the engine computer mounted on its sub-panel. This way, it could be assembled separately
on the kitchen table, in fact. Every wire was run to a terminal block, then on to its destination. This
makes it easy to debug electrical problems, since a voltmeter could monitor any signal very quickly.
Barely visible is a small toggle switch on the upper-right edge of the panel. It allows for blinking out
ECU error codes without messing with jumpers.
Heres the right-front suspension going together; the bright red looks really nice against the black
paneling.
Here, the Honda wheel bearings are pressed into the homemade rear uprights. They turned out
pretty well, although I went a bit overboard by placing the pivot points so far outboard. Both the
Honda ball joints and brake calipers are really close to the wheels, so close that small stones can get
jammed between the two, leaving scratches in the wheel. The large arm at lower-right is the toe link.
While the rear suspension is A-arm in nature, by necessity it had to miss the drivetrain. This meant
that the links were rotated so that the forward side of the A practically becomes a trailing link. The
shocks mount directly to the upright; I saw no reason to put bending loads on the suspension links.
Ive never seen anyone else do this; everybody else attaches them to the lower arm. Attaching them to
the upright avoids placing bending force on the A-arm.
Here, the suspension links arent so much half of an A-arm as they are lateral links. This is an
example of the car designing itself; the links pretty much have only one path to take. I got a bit carried
away with the bottom links, making them so long that they nearly meet in the center. Barely visible are
the bolt-on inner toe-link mounting plates. While I tried hard to get rid of all bumpsteer during design,
I wasnt certain if it would really be gone; at least bolt-on brackets could be changed if the design
proved wrong. Fortunately, it turned out fine, with just a bit of toe-in in both bump and droop. (This
ended up being an unwanted feature; see comments later, in the Driving Kimini chapter.)
There isnt much room here. One important human factors issue is making sure that there is enough
room for feet. My foot has to be able to move from the gas to the brake without hitting the steering
shaft; thats why the shaft comes in high, then drops down. A high-mounted rack, level with the upper
A-arms, would help, but that would raise the CG. Does it really matter, putting it down low? Eh,
every little bit helps. Think of it as a diet.
As more and more components were installed in this area, it became less and less pleasant to work
in there. The empty donut-shaped bracket, top center, is for a Dzus fastener, to hold down the nose
section. The small bracket below it will hold the electrical connector for the nose section. Its placed
there because whenever the nose is removed, the electrical connector must be easy to reach.
The radiator sits in rubber-lined aluminum cups. Visible in the background is a rubber boot on the
steering rack.
The rod-end is the front-lower A-arm mount, mounted on the forward-most tube of the chassis.
The upper half of the radiator and exit ductwork. Visible on the radiator end-tank is one of the
donut-shaped rubber radiator mountssimple and effective. The black panels dont photograph well,
but being assembled here is the steeply upturned radiator exit duct. There really wasnt much choice;
the Mini being so short means that the duct had to turn upward quickly. It exits into a low-pressure
area and also keeps heated air away from the fuel cell.
I have to admit something, so when it happens to you, youll know youre
carrying on the tradition. That is, with the car finally standing on its own
suspension, I, like all builders before, couldnt resist bouncing up and down on
it, with childlike awe and wonder that it really did move like a real car. It was
really cool watching the front rocker arms moving up and down!
With the car fairly complete, with all the major components on the car, it was finally corner-
weighed. While technically a car can be corner-weighed with bathroom scales and beams, using the
right tool saves a lot of time and maintains accuracy.
Corner-weighing
To maximize handling, the tires must be as evenly loaded as possible. Corner-weighing the car
means making adjustments to equalize the weight percentage on diagonally opposite corners of the
car. The goal is to get 50% of the total weight on the left-front plus right-rear, and 50% on the right-
front plus left-rear. This is where the adjustable collars on the way-too-expensive shock absorbers
pay for themselves. Regardless of the type of shocks used, there must be some way to adjust ride
height if corner weights are to be adjusted.
Its nearly impossible to achieve a perfect 25% weight on each tire; that is set during design by the
CG of the various parts. Setting corner weights does not move the CG of the car; it only adjusts how
much weight is on each diagonal set of tires versus the opposite set.
A common analogy is to consider the car as a four-legged table, where one short leg causes
problems. It isnt doing its share, making the two adjacent corners do extra work, and the leg on the
opposite side has little or no weight on it. Its pretty clear that a car set up this way isnt going to
handle well, never mind feeling strange. Because two of the tires are carrying more of the load, their
coefficient of friction is decreased. This means, as a whole, the car will not stick as well as it would
with four evenly loaded tires. Making things worse, the opposite tire has practically no load on it, so
its not doing its share. Taken as a whole, the car will not corner as well as one that has been
properly corner-weighed.
The trouble is that the table analogy doesnt work if the rigid table legs are flexible suspension
assemblies, which they are in real life. Think of the same four-legged table sitting on a mattress
(simulating the springs). Now, the same table with the short leg isnt obvious at all; the other legs
simply settle and take up the additional load. This is what happens on a real car, making it difficult to
know whats going on at each corner.
Because real corner scales cost around $1,100, its common sense to search for lower-priced
alternatives. One such device works by slightly lifting a tire off the ground and measuring its weight.
This works fine on a table because nothing deflects. But think about what happens in our table-on-
mattress example; its clear it cant give accurate values. The other three corners of the car will
assume a different stance while trying to measuring the fourth. If a wheel and tire must be lifted, say,
, to remove all weight, thats about 50 lb. that gets transferred to the other corners of the car,
messing up the very balance that were trying to measure.
Another popular approach is using four bathroom scales; unfortunately, most dont read high
enough unless the car is really light. Some people use beams to reduce the load so the scales see some
proportion of the total weight. This works fine as long as the tire is really centered on the scale;
otherwise, accuracy will be poor. On the other hand, if the idea is to balance the car, to get its CG
right at the center, thats different, and a simple balance beam can be used. But since the CG of cars is
virtually never exactly in the center, this is of little use. Again, corner-weighing a car does not move
the CG; it only varies the loads at each corner.
While real corner scales are downright expensive, unfortunately thats what it takes to do things
right. Electronic scales work so well because they dont deflect when a load is applied; if they did,
the weights seen by the other three scales would shift. As expensive as they are, and as infrequent as
theyll be used, it argues for taking the car someplace and paying someone to do it, or does it?
If the car is going through initial shakedown test drives, its likely that the ride height will be
adjusted several times. If it is done near home, it can always be driven to the shop and the corner
weights reset. But if springs or ride height is changed at the track, unless someone brought along the
scales itll have to be driven as is, so dont expect it to handle well. As painful as it is, investing in
corner scales could end up being cheaper in the long run than paying someone else to keep setting up
the car.
Does this really matter? Does it really have to be exact? Will the car be impossible to drive if its
not perfect? No, itll drive fine, at least on the street. But on the track itll be an unknown quantity.
Thats the tough part of car design and development: its fairly easy to achieve 80% of perfect but
much more expensive and difficult to get that last 20%. Thankfully, I was able to borrow corner
scales to set the car up the first time (thanks again, Bill!).
So, what did it weigh? What was the big scary number I was both afraid to know and curious to
find out?
Here it was, finally confronting my design demons, the numerous indiscretions of adding heavier-
than-necessary brackets and panels. Though this weight was with me in the car, it didnt include the
doors, shell, or all the other bits I was forgetting. They add roughly 180 lb., so full-up weight is
around 1800 lb. One of the most common questions is, What does it weigh? I say 1600 lb., since
driver weight and liquids are always conveniently ignored.
Continuing on with construction, the upper dash cover was finally installed, after a brake access
panel was added as an afterthought. The brake cylinders are buried under the shell, and the access
panel was needed if I ever wanted to get them out again. This was the last panel needed to completely
seal the passenger compartment at the front end.
I read a story of how race-car driver Bob Bondurant injured himself back in the 1960s. His race
car hit an embankment, which tore off the floor pan, allowing his legs to drop through. As his car
rolled over and over, his legs got crushed. For a while, doctors didnt think hed walk again. Helping
to prevent that scenario is this additional stainless panel. If the floor pan is ripped off, this panel will
remain, attached from the insideI would hope preventing similar injuries.
The passenger section of the shell was temporarily installed to guide installation of various panels.
This shot gives an idea of what was to come, how tightly everything fits together. Note the radiator
fan, placed on the forward side of the radiator. Doing this kept it out of the hot exhaust air and moved
the CG forward and down slightly. Also, fans are more efficient at pushing air than pulling it. The top
of the car is about 42 tall.
At this point, composite work began on the shell. Because it hadnt been vacuum-bagged, the shell
had a lot of bubbles. I worked on it for months until I was sick of it; messy, smelly, and unhealthy
isnt a good mix. After reaching my breaking point, I finally took it to an auto-body shop to be finished
and painted. While they worked on that, work continued on the chassis.
The WideView rearview mirror is installed. The idea is good; the glass is curved only
horizontally, not vertically like some mirrors. I lowered it later because the view to the rear was
obstructed by the upper roll-cage tubes. In hindsight, Id probably just use a Wink mirror. I dont like
how the curved mirror makes everything appear further away (costing me a traffic ticket, because I
couldnt see the cop).
The shifter boot, a generic part for Hondas obtained on eBay, is retained by an aluminum frame
inside the boot. While the boot gives the center console a more finished look, it also serves the
practical purpose of sealing off the shifter hole. A fair amount of air blows through that hole, warm at
best, but during a crash there could be flames or gasoline trying to get in. I dont pretend this is going
to stop flames, but it will buy me a little time to get out.
With all the components finally mounted in the front, its pretty cozy, everything barely missing
everything else. While great for being compact and keeping weight down, it makes accessing various
items more challenging, or said less eloquently, more of a pain in the butt. If I were doing it over, Id
try harder to ensure parts that would someday require removal could be accessed without removing
everything else first. Access isnt terrible, but theres certainly room for improvement.
A Halon fire extinguisher is the best type to use because it doesnt leave a residue. At the time, this
unit seemed perfectly adequate, until a buddys Lotus Esprit caught fire. He said, If Id had just one
more squirt, I could have put it out. Instead, he watched it burn to the ground. A larger unit could fit,
but laziness has prevented an upgrade. Note that the extinguisher is mounted on the floor for a low
CG, yet within reach. Some people say small extinguishers are completely uselessI disagree. The
most likely reason to need it is for a small electrical fire, not gasoline. If it can be put out before it
becomes a big fire, why not at least try?
The passenger compartment back from the paint shop! Psychologically, it was a huge step forward.
Having that perpetually unfinished side project hanging over my head for months was a real drag, kind
of like having a big term paper due for school yet not having the strength to finish it. So I paid
someone to finish itthe shell, not my term paper. As shown here, it is fresh off the truck, still with
the stiffening ribs attached.
Before the shell could be installed, it needed protection from the hot exhaust exiting through the
right side. One reason the exhaust exits here was so that it could have a static, finished appearance.
By that, I mean that it exits through a true hole in the shell, not through a slot with a big gap under it,
which would have been necessary had this part of the shell been removable. Stainless makes a great
heat shield since it is a much better insulator than aluminum.
The passenger compartment, now permanently attached to the chassisno going back now! Rivets
attached it to the chassis, and silicon seal helps keep out sand, dust, fumes, and noise.
With this done, the engine cover was next to go out to the body shop. Before that could happen, it
had to be carefully mocked up and drilled for Dzus fasteners. With that out of the way, work
continued on the passenger compartment, specifically the doors.
Ughthe doors! These were the least pleasant part of the entire project. Doors are probably the
largest incentive to build a car that doesnt have them! The problem with old Minis is that nothing is
consistent, including the doors. They arent identical, plus the composite shell was a bit vague in
places, due to it backing away from the mold. I dont know what happened, but stuff seems to change
size whenever it goes out of my sight. The engine tray somehow grew after being powder-coated
and had to be cut downit fit before! Then, one door, which fit the door opening at one point, no
longer did. Huh? Did the shell warp? Did the chassis warp during powder-coat baking? Did they
drop it? I wonder. Anyhow, one door rubbed at the bottom front corner, and the top front corner, and
the top rear cornerdownright impossible, I say, but thats what happened. One thing this project
taught me, however, is that you can waste time figuring out what happened, but in the end you just
have to deal with it. Not only did the window frame hit in opposite corners, it was also very bent
inward. When I tried bending it back out, the door made very scary noises, like it was about to break.
Since the door had to go back to the paint shop anyway, to fill the side mirror holes they forgot, out
came the saw. Yes, I cut that cursed window frame off, and it felt great. The frame had about
removed in order to miss the upper door flange.
Mk1 Mini door hinges are pretty coolodd, but cool. The doors are hinged externally, which
places the pivot point outside the bodywork. This works well and has a cool retro look; however,
they have a feature that I never figured out: the doors are curved, and the hinges mount flat against
them. Id expect that the hinge studs would extend through at 90 to the door surface, but they dont!
For some reason, the factory decided that the studs must be levelto the ground! Thats all fine, until
you try to tighten a nut on the stud; the nut is now at an angle to the surface its being tightened against
so it cant be tightened. Tapered washers were made to undo the factory cleverness (sigh).
And then there was the mischief with the weatherstripping. Lee said it didnt look like real Mini
weatherstripping to him, even though it came from Mini Mania. Anyhow, it stuck out far enough that it
was hard to get the door shut all the way, due to over-compression. Trimming it wasnt a good idea
because its of the hollow-bulb type. One more treat awaited me: apparently Austin changed its doors
many times during production, and my replacement Mk1 door (due to the first one literally falling
apart due to rust) might have been a later version; a portion of the weatherstrip completely missed one
section of the door. At the end of the day, it was made to fit okay.
Here are the external door hinges, with the beginnings of the 1/8 thick rubber padding used to
separate the body shell sections. The nose rests on the padding, and Dzus fasteners keep it from
moving around. A hard mount isnt needed for body panels that naturally want to move around some.
It also gives a little if something hits the nose.
Next was installing the windshield, which didnt go so well. The short story is, silicon skins over
very quicklywithin a few secondsand trying to trowel the silicon smooth was doomed to failure.
Silicon had to be used because, for some reason, the cutout in the shell didnt match the glass. While
there was a recessed areaseemingly for the glassit wasnt deep enough. Even more strange was
learning that Mk1 Minis dont have a recessed windshield flange.
This brought up the question of exactly what year Mini the mold was taken from. Lee said he
thought the donor shell for the mold was based upon a 1964 car, but even then, the early cars had no
flange at all, just a simple cutout. Not knowing this, I built the steel chassis to come right up to the
flange, so there wasnt room if it was cut back. Anyway, the windshield install looked badreally
badbad enough that I took it back out after this picture was taken. I set it aside so I could decide
how to do it better the next time.
I got some help from an auto glass shop; they painted a black trim border around the edge of the
windshield, which really improved the look. It was finished off with black silicon to fill the gap. It
came out okaynot great, but good enough. The secret was to smooth the silicon with a wet finger (or
ice cube). I wish the Mini gasket had worked out, though.
This picture shows the coolest part of the whole carthe door opener mechanism, visible here on
the left, inside the door. Its a stock Mk1 Mini pull-release, and the door latch is released by pulling
on it; its kind of like the rope you pull on to stop a streetcar. Inexpensive, light, simple, elegantI
get a kick out of it every time I use it.
About this time, the body shop finished the engine cover (whoo-hoo!), so Dzus fasteners were
installed to mount it to the body (shown here at the leading edge, and another row along the bottom-
rear edge). Hinges were considered, but that idea was discarded. The concern was that theyd be too
complicated and could allow air to lift the front edge of the engine cover at speed. The engine cover
really should tuck under the passenger section to make sure it doesnt lift, but that thought was lost
somewhere.
Roll-bar padding was added anyplace I thought I might strike in an accident. The correct stuff to
use is not hot-water pipe insulation, because its way too soft. Your head will crush it down to
nothing and then hit the steel cage directly. SFI-45.1 is the correct stuff to use. It is much more firm
than the cheap stuff and can save your life. Expect to pay about $4/ft. from circle-track suppliers.
Next, it was time to install scratch-resistant Lexan in the engine cover. Since even scratch-resistant
Lexan scratches (even with a paper towel), the windows are screwed in instead of riveted. The Lexan
was first marked and drilled with pilot holes, and then fit up to the shellClecos can be used for
things besides metal! With it fastened in place with the Clecos, the shell is match-drilled. Lexan
expands a lot with temperature, a 36 section growing about 0.25 when going from 32F to 200F,
which can be reached in the engine compartment. To give room for expansion, all the screw holes
were generously oversized, so the window could slide around a bit. The windows are fastened with
really cool #4 stainless Allen-socket, flat-head screws. From a couple of feet away, they look just
like rivets. Nylon washers and nylon lock nuts were used everywhere. The black trim here is none
other than simple black vinyl electrical tape. The right way would have been to paint it, but I didnt
have the proper paint at the time.
Soon after completion of the windshield, the nose showed up, toowhoo-hoo again! Lights, turn-
signals, Dzus fasteners, and the aluminum radiator exhaust grill were added, and hey, this thing is
starting to look like a car!
The aluminum mesh was from McMaster.com, as were many of the industrial-related components
in the car. Many times they were the cheapest source, since they cater to the industrial community
rather than car people.
Heres the proud father and child. When posted to Kimini.com, this photo received more comments
than any other. Only now was it apparent just how small the car wasalmost toy-like. People asked
how I could possibly fit; the truth is there was lots of room; after all, it was designed to fit me!
Moving on to the windows, heres a problem with the stock Mini window hardware. The lock
cant move any further outboard; Im holding it against the stock Mini chrome lock strip. Theres no
way the lock can rest against the window as it should. It turns out that even new Mini parts arent very
consistentthe chrome lock strip was bent wrong. The lower section of the assembly had to be cut
off so it could mount to the window.
Here is the lock after modification. Note how far it had to be moved outboard in order to work
properly. I wasnt very happy with the quality of these factory new parts.
Heres a neat tool for cutting large holes, available from Aircraft Spruce. It has a cutting tool
resembling an end mill, and for the Lexan windows it worked very well. The black knob is the center
point of the hole, and the cutter is slowly walked around it in a circle.
Heres why the large holes were added. The engine cover, while easy to remove, takes a while to
actually get off. A simple means of checking oil and water was needed, and these ports filled the role.
The oil dipstick is easily accessed from this side.
The coolant header tank cap can be reached from the other side. The ports may end up with covers
or maybe not. A fair amount of engine heat comes out of these, and not just at rest. For this reason,
they are being left uncovered.
More of the aluminum mesh is used to cover the lower engine compartment, attaching to the same
tube mounting the Dzus fastener plates. The lower vent serves two purposes: at speed, air trapped in
the engine compartment has a way out; at rest, it allows convection cooling by letting cool air in, to
replace the hot air that rises, exiting through the window vents.
For the rear window, I had once seen a really neat one on an old Ford Shelby GT-350R. It looked
like an ordinary rear window, except that the top edge curved inward. It helps vent hot air out the
back, and the idea was filed away for later.
Unfortunately, it isnt possible to bend scratch-resistant Lexan, and even if it were, due to the rear
window curvature, I would not be able to bend it in the vertical axis, too. Instead of having a long
vent at the top, it would have to be cut.
And so it was. The top of the window is tucked beneath the shell to give a better appearance.
Fairly large vents were cut to help cool the engine compartment. The holes up high help convection
cooling at rest, and at speed the low pressure behind the car draws hot air out the back. Plus, filling
the low-pressure area behind the car with hot engine-compartment air helps, in a small way, to lower
drag. Several people had suggested that the black border around the window be extended all the way
around the top, for a more finished look. I agree, but since all the screws were already installed, it
has to wait until later.
The fuel vent hose was added, going straight up out of the fuel cell, and then looping back down to
the bottom of the chassis. The idea is that if the car crashes, coming to rest upside down, the fuel will
not drain out, due to the loop. (Thanks to Alan Brickey for this tip.)
This gives an idea of how busy the finished front area is. Fuel lines, brake lines, fresh air blower,
coolant lines, electricalthey all come together in an uneasy truce, where everything is fine as long
as nothing needs work!
Safety wire on the front uprights. The cap on the upright contains the nut to the ball joint, serving as
the upper pivot. Since the cap must be removable, the three bolts keep it from falling out. The cap
would (probably) stay in place without the bolts, due to the spring force always pushing it downward,
but theyre still a good idea! Safety wire was used instead of Loctite, to be really, really sure they
dont back out.
Heres the first of the rubber sheets being laid into the wheel wells, this being the upside-down
nose section. Rocks kicked up by the tires hit the composite, potentially hard enough to leave a star
on the outside surface. The rubber absorbs the energy from the rock, plus keeps things quieter. If I
were doing it over again, truck bed paint would be considered, since its much easier to apply.
Heres the other portion of the rock deflector. A stainless tube was bent to follow the inside
contour of the front fenders. This worked well but tended to rub the tire sometimes. The right solution
is a composite inner shield molded into the nose sectionone of the millions of things still needing to
be done.
Speaking of wayward stones, for the nose of the car, transparent 3M protective film (from
stickercity.com) was used to protect the nose from bugs, sand, and small rocks.
At this point in the build, the car was nearly complete. I began receiving mail about driving the car,
which fell into two camps.
One groups philosophy was summed up by this advice: Drive it now. Dont worry about finishing
it theres always time for that later. If you wait until its perfect, itll never be driven.
But another group warned: Once you start driving, thats all youll want to do. Anything left
undone risks never being completed. People I know who started driving their project cars before they
were complete never did finish them. To this day they still drive an obviously uncompleted project
car that never made it to completion.
So work pushed on until it was really done, and the latter statements proved to be very true!
One thing I have a hard time remembering is to tighten lug nuts (this once caused me to lose a tire
on the Datsun). Using a torque wrench is a really good idea; I use it religiously whenever reinstalling
the wheels. I force myself to use it by not having a tire-iron anywhere in the garage. The torque
wrench ensures that each lug nut is the same tightness, preventing warped brake rotors or stripped
threads.
At this point the car was just about ready for the street, but before that could happen, insurance had
to be dealt with.
37: Insurance and Registration
Insurance
I had signed up with AAA, specifically for its towing service. I learned they had recently begun
covering hot rods, so I figured, Why not? Heres a portion of the build diary that captures the
moment:
So, Im checking with insurance companies to see whats available. Hearing
that AAA recently started covering classics and hot rods, I gave them a try. The
phone conversation went well; they had very reasonable rates, so we moved to
the next step, having pictures taken. I got a call later that day:
AAA: Sorry, we cant cover your car.
ME: Thats disappointing. Why not?
AAA: With that roll cage, it looks like a race car.
ME: Yes, but if I remove even one tube of the chassis, the car will collapse
the cage is the chassis. Dont you cover Jeeps with roll cages?
AAA: Then whats that blue tank on the floor?
ME: An oil accumulator, called an Accusump. Oh, wait, you think its a nitrous
bottle, dont you?
AAA: Yes.
ME: Well it isnt, and notice that its not connected to anything. Why didnt
you ask me?
AAA: And theres no passenger seat.
ME: I already explained thatthe car isnt finished. See the mounts for the
seat? Its up in the rafters, which you didnt ask about either.
AAA: Well, sorry, we cant cover you.
ME: Grrr.
They really got me steamedmaking their decisions based upon uneducated guesses. I dont see
how what I had built is any different from many 50s or 60s hot rods, which have been so extensively
modified that they have virtually no original parts in them. In fact, I would bet that I have more
original parts in mine than some American hot rods running around. I wish AAA had been up front
about their real motivation; if they feel its a race car, okay, say so, but dont start looking for
excuses. Good thing it wasnt redor had racing stripes. How do they insure 600 hp Cobras? How
can they insure a 1950 Chevy, where virtually nothing on the car but the frame rails is authentic?
I guess I have the street racer Honda kids to thank for AAAs paranoia. (Okay, and maybe my
racing seats, harness, and big fenders, but never mind.) So, if I dont have a roll cage, is it then easier
to insure? Does that mean its safer? Isnt that odd?the more safety stuff you have, the less
insurable it is. If I raise the car, making it less stable, and looking less like a race car, is that safer?
But a Jeep Wrangler, with a CG of around three-feet high, and a roll cage, is deemed no problem? I
guess my car simply doesnt fit their checkboxes, so theyre afraid of the thing. Well, so was I, but
that was beside the point.
That finished AAA, so the next stop was State Farm, since they covered our normal cars. That
discussion went okay, too, until they said I needed to get the car appraised. That really amused me;
how the heck could the appraiser figure what the car was worth? He could come up with anything
between $50 and $50,000. I really didnt want to bother, figuring hed come up with a low value, so
low that I wouldnt want to insure it with them in the event of a total loss. Anyway, I tried explaining
how my car was rather unique, but they just werent hearing me, saying, Well, just go find out what
cars like yours sell for. While there may be a few cars like it in England, I dont think thats what he
had in mind. He seemed to think there were hundreds of cars just like mine running around here in the
U.S. So that ruled out State Farm.
I planned to drive it less than 5000 miles a year, so Id probably just get liability and
comprehensive insuranceit just couldnt be that hard. There are lots of odd cars running around out
there, and mine actually looks like a real car. GeezI wondered if Id told them it was a kit car,
would they then say that it was okay? This seemed to argue for going to insurers that advertise that
they cover hot rods and kit cars.
Hagerty was my next stopa large, well-established insurer of hot rods and show cars. With
decent rates and reasonable mileage limits, they seemed like the obvious choicebut they werent.
This issue brought up the following build-diary entry:
Hagerty was my first choice, until someone pointed out the limitations. They
explicitly state that they do not allow driving to work or doing anything where
an ordinary car could have been used instead. So yes, perhaps my comment
about stopping to eat is a stretch, but it certainly makes me wonder where they
draw the line. If stopping to eat is okay, can I stop at the grocery store? In
other words, how long can the car be unattended before they stop coverage?
Heres a practical example. Say I take a pleasure drive (their term),
which they allow. I drive several hundred miles to a car show (but dont show
the car.) Is the car insured for the drive there? Since many car shows are both
Saturday and Sunday, I get a hotel room for Saturday night. Is the car covered
while sitting in the parking lot? Technically I took a pleasure drive to get
there, but now what? Since the car isnt in the show, I could have used a
regular car to get there. If they use their same logic of why you cant drive to
work (could have used a normal car, car sits unattended for hours), this seems
like a similar situation, so it probably is not covered. Yet, many car shows run
over the weekend, so what do people do? Guess I need to ask Hagerty. My
point is, what they cover and what they dont could potentially be interpreted
however they feel like, whenever they want. If the car vanishes overnight and I
put in a claim, I can see how it would be very easy for them to say it was not
covered. It could save them a lot of money, and they can hide behind the vague
rules. Of course 999 out of 1000 times nothing happens, but its that last case,
when you really need them will they be there?
What if I drove these same miles but put the car in the show, but overnight
had to park it at the motelis it covered then? Might be, might not; I just
wasnt willing to roll their dice. I know lots of people insure with them, but I
bet the majority are using their cars in situations where they arent covered
and dont realize it. As long as nothing happens, no problem.
This unknown scared me away from Hagerty.
Unfortunately, what I wanted didnt seem to be available. If all I wanted was liability coverage,
insurance companies are easily able to provide that. The problem was that I also wanted an agreed-
upon value in case of a total lossgetting stolen or burning to the ground. If it ever got stolen I would
be very, very unhappy. I wanted a check if that happened receiving zero compensation after 10
years of work would be a very hard hit. The received compensation would at least help dull the pain
and cover material costs, never mind the 4000 hours or so (no kidding) of labor. So, I did want it to
be worth something, but that meant signing up for comprehensive coverage, which suddenly made
everything more expensive. They seemed to think that if the car got a dent, it would be taken to a
repair shop. No I would fix it, but they cant deal with a self-sufficient owner, and it didnt matter
what I told them. Once again, I just didnt fit their little checkbox forms. This stuff wasnt impossible,
just really annoying.
I got in a discussion with a buddy about how insurance companies pick and choose their business.
The issue was my roll bar, deemed a high risk because Id likely drive in a more aggressive manner
than if I didnt have one. I countered with, If we get rid of seatbelts, that would make people slow
down, making the streets safer. Years ago, I read a comment from an automotive engineer, after
hed heard that driving speeds had increased after the introduction of airbags. His remark was, If we
replaced the airbag with an 8 steel spike, people would drive more carefully. I wonder if insurance
companies would back such a rule? People would slow down, for fear of being impaled; insurance
companies would have very few claims, and isnt that their goal, to maximize the income/claims
ratio?
This left Vintage Car Insurance, Voyager Classic, Grundy, and Condon & Skelly. After reading up
on these companies, I was most impressed with Condon & Skelly. The lack of any bad remarks from
customers led me to go with them.
They turned out to be very easy to deal with, and it was easy to insure the car as a kit car. Why?
The rules are much more flexible with a kit car, or any car composed as a combination of
components from many different makes. Oddly, a kit car with a roll bar didnt seem to matter.
Another big advantage was that they allowed an agreed-upon value to be assigned, meaning that I
could pick whatever value I desired. The only negative factor was the limited mileage, 2500 miles
per year. After thinking it through, I decided that this seemed okay since the car was going to be
driven only one or two days per week.
The decision was made, the forms were printed and filled out, a check was written, and the whole
mess was mailed off. About a week later I received the official paperworkKimini was insured!
Later on, insurance was changed from Condon & Skelly to Grundy. In four
months, the car had accumulated 1362 miles; it was pretty certain that 2500
annual miles wasnt going to be enough, and Grundy offered 5000. Grundy does
have the same fine print as the others: no driving anywhere a normal car could
be used instead.
(Note: When I renewed my insurance in late 2006, Grundy no longer had a
mileage limit.)
Not allowing the car to be driven to work bothered me less and less as time
went on. Traffic here in Southern California is bad and getting worse. My daily
commute is 17 miles, which takes about 40 minutes, so average speed is pretty
low. This average speed is misleading though, because theres one large blood
clot in the same area every morning, with heavy stop-and-go traffic. Driving
Kimini through it reminded me just how much I appreciate my truck, with its
automatic transmission!
Also, during the summer, it gets really warm (90F100F) and humid. As Ive
said elsewhere, Kimini is a car with no air conditioning and small windows that
dont open very far. While engine heat has been dealt with, the heat of the day
makes it very unpleasant. (Minis destined for Australia, where it gets hot like
here, were given roll-up windows.)
It made me realize that driving it to work, even once a week, wasnt quite as
much fun as expected. Ironically, it meant that the insurance rule I was
complaining so much about (not being allowed to drive to work) became
something of a moot point.
With insurance out of the way, that left one more hurdleregistration.
Registration
With the car based vaguely upon a 1961 Mini, registration was fairly straightforward.
Back when the donor car was purchased, it was immediately registered through AAA. The purpose
of doing that at the beginning of the project was to get a monkey off my back not having to worry
about whether it could be registered at all. It was also far easier to register a box of parts and a rusty
shell than to try to later register something that didnt look very much like an authentic Mini.
Once the donor was registered, it remained, in the eyes of the State, to be non-operational until
the car was done; no point paying full registration while it was just a pile of parts. Then, it was
simply a matter of reactivating the registration so it was legal to drive. This simple process was
rather nerve-racking.
Arriving at the DMV with all of the paperwork pink slip, proof of insurance, and registrationI
nervously waited my turn in line. Nearly a decade of work now hinged on this moment, and while I
was almost sure I had everything under control, I also knew that the government could have changed
the rules at any time.
Finally, my name was called, and as I stepped up to the window, I realized that my hands were
sweating. The lady went through all of the paperwork, casually mentioning that it had certainly been a
long time since Id bought it yes indeed. And then, she said what Id feared most, Oh wait, youre
going to need a smog check.
I just about freaked, seeing everything Id been working toward for so long being at risk. Sure, I
could trailer the car to events, but that was never the intent. I had gone to a lot of trouble so I could
drive it on the streetand now this.
Trying to keep my voice even, I asked, But its a 1961 Mini; it doesnt need a smog check, right?
To which she replied, Oh, youre right, I guess youre all done then. And I was thinking to myself,
Geez, dont do that to me!
So, with that disaster averted, the car was now finally ready for the street!
38: First Drive
So it had finally happened; after years of designing, collecting parts, building fixtures, cutting,
tacking, welding, drilling, screwing, riveting, soldering, crimping, bolting, bleeding, and cussing,
finally it was ready to drive.
Starting it up and waiting for it to come out of fast idle, I realized that my heart was pounding.
Why? I had done everything I could to ensure an uneventful test drive. But there is always the
unknown, the one screw or bolt that I forgot to tighten, the one little item that, at best, could cause me
to coast to the side of the road; at worst, this small part might cause a structural failure and send me
into oncoming traffic.
When I first pulled out of the driveway, I drove very tentatively down the street and back to warm
everything up, not driving further than I was willing to walk, or push! Then, it was time. The first real
drive was intimidating to say the least, coming out of our neighborhood street, to the real street, the
one with speeding cars and trucks. I thought better of just pulling out in case something brokelike if
the engine coughed and stalled. The brakes didnt inspire much confidence, either, since they were a
bit mushy. It was added incentive to wait for a good-size opening in traffic. After 10 years of work, I
could wait another minute or so. When I finally pulled out, heavy traffic prevented me from going
faster than about 45 mph, but that was nerve-wracking enough!
During the 45 mph drive, the speedometer indicated 90 mph. The SPA
tach/speedo counts both edges of the vehicle speed pulses, hence the 2:1 error.
After the short run down the street, when I turned back into the neighborhood, I found I had an
engine problem again. The engine acted like it was running on three cylinders, but only between
roughly 3000 and 4500 rpm. (Good thing I hadnt pulled out in front of that truck!) If I pushed through
that rpm region, it popped out the high side, running just fine. The stock Honda ECU said there were
problems with the EGR circuit and electrical load detector (ELD)no surprise, since I have neither.
Many people had assured me that neither would cause any drivability issues, and the car ran fine
earlier. My guess was that the ECU had switched modes once the coolant temperature came up. No
clue what the problem was, though.
Next morning, the plan was to diagnose the engine problem, only it had vanished, and now the
engine ran great! Uh ohwhen problems just go away, Im always suspicious of accepting gifts
from the Goddess of All Things Mechanical, wondering how Im going to pay for it later.
The engine problem never returneda year later it was still running strong.
Perhaps it was a stuck electrical or hydraulic valve, or perhaps a bad connection
on a fuel injectorperhaps I shouldnt question gifts too closely.
Anyway, so now that it was working right, how did it run? Pretty amazing!
The engine felt very free to rev, like it wasnt connected to a whole lot. At around 1600 lb., Kimini
didnt take much to get her moving. The steering was very light; no need for power steering, even at a
standstill. The two low-pressure fuel pumps put up quite a racket, though. Doing it over again, Id use
a single high-pressure pump inside a small sump inside the fuel cell. As it was, the panels that the
pumps were mounted to acted as sounding boards, making them really annoying. They were later
removed from the panels, left suspended by the short, stiff hoses; this is far from a proper solution, but
they are certainly quieter.
When I went out for another test drive, there was a fair bit of traffic, made worse by local
construction, so I didnt feel comfortable going too far. I was very insecure, so the plan was to
continue making laps to make sure nothing broke, leaked, fell off, melted, unscrewed, backed out,
shorted, opened, seized up, scraped, or rubbed.
The first test drive revealed the need for inner fendersthe car was a very
good rock collector! Later, rock-deflecting curtains were added, not fun to make,
but necessary to keep the car from filling up with sand and gravel.
Later, the car was taken out again. Traffic opened up so I finally got on it a bit in third gear. The
experience is hard to describe; it just kind of launched, kind of a rubber-band feeling. No spinning
tires, no straining engineit just went. Due to Kimini weighing half as much as the donor car, it was
like having a Honda Prelude with 400 hp and 300 ft. lb. of torque. Like I said before, it was like there
wasnt much attached to the rear tires, almost as if the rear end were on jack stands. I stepped on the
gas and the rpm climbed very quickly, because there was, indeed, very little load on the engine.
Then theres that VTEC thing I had heard kids talking about, and sure enough, the Goddess of
Acceleration grabbed hold of the car at 5500 rpmcompletely changing the character of the engine. It
was like suddenly finding another inch of gas pedal travel. The exhaust became much louder, the
intake more noisy, and off we went. A very addictive thing!
Besides the light weight, the engine is further helped by an 8 lb. Fidanza
aluminum flywheel replacing the 17 lb. steel unit, no power steering pump, smog
pump, air conditioning, or power brakes. The 13 wheels and tires also reduce
rotational inertia, and the intake and exhaust free up even more power.
As great as it ran, it still had cold-start issues. Something was going on with the coolant
temperature, but there was time for that later.
With the car checked out (as in, nothing dire happened), it was time. Filled up the tank, parked in
the driveway, went in, and told the wife, Its time. Helped her into the car and on with the belts
Kiminis very first passenger. We went out for a short drive, giving her the honor of the first ride;
after all, shed put up with this nonsense for 10 years. On return I asked her impressions:
Very nice. Smooth, noisy, yet comfortable; very, very stable. When it took off, oh my, that
engine is strong. You have to respect that engine.
Whenever I read this it makes me smile, and shes right about the engine.
Kimini makes her first public appearance at the Carlsbad, California, car meet on December 31,
2005. After 10 years of design and construction, it still hadnt quite sunk in that it was done.
39: First Car Show
Kiminis first car show was at Menghini Winery in Julian, California, on May 28, 2006. By this
time, Id put sufficient mileage on the car that I trusted itas much as you can trust a car built in a
garage. It was the longest drive yet, around 140 miles round-trip. Packed up the car with a chair, car
cleaner, hat, camera, paper towels, and some aerosol Windex for the windows. Heading out into San
Diegos east county in Kimini was very exciting, the first time Kimini was flying on her own, far
from cities, freeways, cell-phone access, and roadside assistance.
Partway into the trip, I started smelling Windexuh-oh. Something had pressed the nozzle on the
aerosol spray can and it was spraying Windex all around the passenger floorammonia really stinks.
Stopped at Dudleys Bakery, a local landmark that bakes really great breadin the middle of
nowhere. People on their way to the mountains and desert have kept them in business for decades.
Going up a particularly twisty section, a V8-powered Sunbeam Tiger came up behind me, heading to
the same show. I didnt want to do anything dumb, so I kept the speed down to about 35 mph but
maintained it through all the turns, just to see what he would do. The road was so twisty that he
couldnt use all his horsepower. That was funanother mark for light weight. He said, later, that my
brake lights were brokenno, I just wasnt using them!
The weather was perfect, with about 70 cars of all kinds at the show. I entered the Special Interest
class, knowing no one would accept the car anywhere else. The picture here shows just a small part
of the show, with Kimini at back left. I wasnt sure how hard-core these people were; it being my first
show, I didnt know if theyd call the Show Police if someone entered a car in the wrong class. Since
they were awarding the same prize for first and second place, and there were only three cars in my
class, it nearly guaranteed Id win somethingnearly.
At the end of the event, the prizes were awarded; first place was called, then secondand I wasnt
called. You know, until that very moment, I would have said winning didnt matter, and if the prizes
had been cheesy little plastic trophies, it wouldnt have been a big deal. But the prizes were really
cool wine decanters, and all of a sudden it did matterI both wanted one and had (past tense) the ego
to think I somehow deserved one. Apparently the thimble-headsI mean judgesthought otherwise,
giving it to a clearly undeserving dolt. But thats okay; Im grown up and mature and understand that
its all for fun. Just bringing the car to the event was enoughto know that the judges at least looked
at my car, deciding that 10 years of sweat and tears to produce a unique car meant nothing at all. And
finally, its okay that some mass-produced, boring, ugly car won instead. Thats all right; I mean, how
could boneheaded judges possibly be expected to notice the engine in the back? But thats okay; I
didnt want that dumb decanter anyway. I got over it.
Kidding aside, I do wonder why the car that won second place was even in
Special Interestit really was a mass-produced car from the 1950s. To me,
Special Interest means a much customized car that doesnt fit into any
production category, and that car was dead stock. Whatever.
The Special Interest class is a bit of an oddball; in the judges defense, how can they judge cars
that, by their very nature, dont fit in anywhere? If Im going to make a habit of going to car shows, I
should make up a display that explains the car; everyone kept thinking it was a kit. It wasnt clear at
all what Kimini was, and just walking by, some people in fact assumed that it was just a freakish Mini
with big flares. At car shows, all the cars are situated to have their front ends toward the crowds; it
wasnt until people walked around behind it did they notice the engine (which a number of people
didnt notice at all). Later, I took off the engine cover, but by that time the judges had come and gone.
Of course, then everyone was saying, Oh, a race car.what else would they think? I need
something to explain it all. Did I mention I got over it?
Later in the year I entered another car show. This time I wonyup, a cheesy
plastic trophy. (Am I fickle, or what?) I plan to enter the above show again, in a
never-ending quest to win one of those cool wine decanters, dang it!
After leaving the show, I thought it would be cool to check an item off the wish list of things I
wanted to do with Kimini. One was to take a picture of her in front of the famous Palomar
Observatory, so on the way home, I drove up Palomar Mountain. At about 5200 ft. altitude, power
output was less, but because of the stock ECU, the car ran flawlessly. On arrival, it was disappointing
to find that the parking lot is positioned exactly behind a couple of tall trees that all but blocked the
view. I asked a park ranger if he had a chainsawhe didnt see the humor. The observatory is
peeking over the trees in the background, looking like a cloudoh well, you get the idea.
When I parked it for this picture, I got surrounded by fellow visitors. The
people most interested had British accents; pretty funny, I thought. Everyone
wanted to know if it was a kit, and they got funny looks when told it was made
from scratch.
Coming back home on the freeway, traffic was light, and, well, why not, so I opened her up. A
speed of 120 mph came up very quickly and was completely uneventfulalmost oddly so. Other than
the engine noise, the speed simply wasnt apparent. Every other fast car Ive ridden in always started
changing character at higher speedsin a bad way. Theres always lots of wind noise and buffeting,
with one end of the car or the other getting a little squirrelly because it got light. Kimini didnt do
anything like that: no buffeting, no lightness at the front or rearnothing; it just hunkered down.
I ran out of room before achieving top speed, so thatll have to happen some other time. Since the
calculated top speed of 130 mph was only 10 mph faster, no nasty surprises were expected.
Cruise Night
Sorry for the picture quality. This was at the local hot-rod cruise that runs all during the summer
months. Hidden behind the crowd is little Kimini, unfamiliar with all the attention.
Kimini got a lot of attention; older people looked at it and smiled, some knowing it might once
have been a Mini. Younger people would laugh and point, then some of the guys noticed the engine
and stopped laughing.
Kimini was pretty much summed up by one teenage girl, who, upon seeing the car, exclaimed, Oh,
my, god, that is like, so cool! And there you have it.
A little bit of everything shows up at Cruise Night. Here I am behind a 200 mph Ford GT, stuck in
the same traffic jam as everyone else.
40: Driving Kimini
Theres a passionate love affair going on between the car and the driver. Its
rather like Im very proud of the lady Im married to, because she looks after me
so well. Theres no one else that I know that looks after me as she does. She
says, What do you want me to do? Well, I want you to, instead of oversteering
when we go in there, I want you to be a bit more compassionate, and when I ask
her, I say, Look, were now in troubleplease help me. Shell do what I want
her to do. With a car, you have a human being you can talk to, one that you create
and you can buildthis wonderful union.
Sir Stirling Moss
I am my worst critic, and cynical to boot, but the car came out pretty wellbetter than I ever
expected. Even after Id driven it for several months, it still hadnt sunk in what had been
accomplished. The near-stock Honda engine starts easily, idles smoothly, and is very tractable. Since
the car is so light, and with 13 wheels, its very easy to engage the clutch. With the insulated engine
bulkhead, it remains cool and quiet enough to carry on a conversation without shouting. Having the
passenger compartment completely sealed means no exhaust fumes, dust, or wind blowing in.
It is very much a multi-use car, driven hard one minute, yet cruising at low speed the next. The
engine runs at 3700 rpm at 60 mph, which has me reaching for another gear sometimes, but its a
small bother.
Driving the car is an occasion. That is, it takes a little while to get strapped in, but far from being
irritating, it heightens the excitement. This is enhanced by driving the car infrequently, yet if allowed
to drive it to work every day, it would, well, frankly, suck eggs; freeway traffic ruins the fun. I drive
17 miles to work, which takes roughly 40 minutes. That translates to an average speed of 26 mph; yup,
and driving a stick is no fun. The car was built to be many things, but a commuter vehicle it is not.
Its kind of like owning a thoroughbred race horse and only being allowed to
walk it. I cant imagine how frustrating it must be to own a 200+ mph super-car.
But when the weekend arrives its a different story; I find myself looking for excuses to drive it.
The light weight is the single biggest virtue of the car, literally transforming a nearly stock engine into
what seems like a racing engine; even on the freeway its great fun.
Driving a small sports car on the street is like dancing with elephants:
everything is fine as long as you dont get stepped on.
Street performance
After completing the car and collecting some preliminary performance figures, I was curious how
Kimini stacked up against other cars in history. I went through old Road & Track magazines and was
shocked to find that Kimini is as fast as an original 427 Cobra. Road & Track lists the Cobras 060
mph time as 4.6 seconds and its quarter-mile as 12.7 seconds virtually identical times to Kiminis.
If I really pushed it, I can reach 60 mph in second gear onlywithout shiftingin 3.9 seconds.
This is where the advantage of the mid-engine layout really shows itself. With
60% of the weight on the rear tires, traction is very good almost too good. That
is, the stock clutch wasnt happy when I tried some serious timed runs. The clutch
was slipping more than the tires, so a heavy-duty pressure plate will improve
times even further.
Okay, theres a bit of reality check to admit to regarding the Cobra comparison advancements in
tire technology have made an enormous improvement in traction over the years. If modern-day tires
were put on that same Cobra, Im sure it would leave me in the dust. Still, its great to honestly say
that Kimini is as fast as an original 427 Cobra (using 1960s tires, of course, which dont exist any
more, but never mind).
There arent a lot of cars that can beat Kimini in a straight line, let alone a twisty road. Even so,
Im careful what I go up against, because in Southern California at leastthere are people with a
lot of money. Theres some pretty fast hardware on the roads, and the owners can afford the potential
ticket better than I can. The cars Id mostly likely have trouble with would be the Corvette Z-06 and
various Porschesvery formidable adversaries. Its likely, though, that the driver would
underestimate Kimini and not really stomp on itonce. The far more prolific Subaru STi and
Mitsubishi Evo arent a problem, though the Nissan Skyline might be if its ever imported.
It wasnt so much the risk of losing, it was the rather dire threat of having the car confiscated if
caught racing (never mind being thrown in jail). Its not something to be taken lightly in Southern
California, where many kids have died while street racing; the laws have come down really hard on
the activity. Just lighting up the tires is considered exhibition of speed. Being middle-aged and
married also tends to put a damper on such antics.
Back when Kimini was being designed, there was the vague goal of building
something that could keep up with sportbikes going up the twisty road winding up
Palomar Mountain. As construction progressed (and I aged a decade), I began
having second thoughts about actually doing that. Or maybe Id just try to keep up
with a bike, perhaps even pushing a bit, to see what hed do. But what would I do
if he went down? How would I feel, having contributed to his accident? Yes, he
could have backed off if he felt he was in over his head, but the fact that I was
behind him certainly would make me feel partly responsible. How would I feel
about that?
It may be a moot point; the police have wised up to people driving up the
mountain in a hurry; they now patrol the road very closely.
It took many weeks to gain confidence, driving around the neighborhoodvery timidly at first. The
first concern was getting stranded, at best coasting to the side of the road, at worst having something
dire break, causing an accident. As Ive said before, I signed up for whatever happens, but I didnt
want anyone else to get hurt. Eventually, confidence grew to the point that it wasnt safe finding the
limits on the street any longer. Driving quickly in a straight line wasnt an issue; it was the turns. At
only 4.5 ground clearance, if I slid wide through a turn, bad things would happen hitting even a
standard 6 tall curb. The decision was made; it was time to take Kimini to her first autocross.
First autocross
April 1, 2006. It was nerve-racking, with my self-conceived notions of how the car would react,
plus thoughts of what other people might thinklike that mattered. After 10 years of design and
construction, it had finally come down to this, taking it out on track, to be judged by the Goddess of
Handling. Using equations to determine suspension settings got the car on the road, but how well
would it handle?
What became obvious was that the car leaned too much, though it shouldnt have been a surprise.
During design, the assumption was always that it would have anti-roll bars, so spring rates were
chosen assuming that the springs and bars would set the desired roll rate. During the years of
construction, though, I had time to reconsider that assumptiondid it really need bars? After Id
spent so much work building true independent suspension, it seemed a shame to undo all the effort by
cross-connecting the wheels with bars. So the car was run without, just to see what would happen. If
it did okay, great; if not, then bars would be added. It was, in essence, a free (lazy) way to see what
was what. Since the tires wore more on the outside edges, though, it was clear something had to be
done, so the anti-roll bar design was started. (Images credit Max Nealon.)
Design of anti-roll bars stopped, though, because I had also decided that the springs were too soft.
The ride on the street was very comfortable, a bit too comfortable, actually. Of course, as these things
go, nothing is ever black-and-white; its up to the designer/driver to determine whats best. For a
track car, lap times rule all, but for the street, ride quality is an issue; nobody wants to bite his tongue
off when the car hits a bump. I was willing to put up with a very firm ride on the street, though, in
order to do well at the track. If racing slicks were ever used, the nose would dive even more due to
harder deceleration, something anti-roll bars do not control. This argued for higher spring rates
okay.
Since the roll (lean) resistance of the car is a combination of spring and anti-roll bars, they have to
be handled separately. The soft ride was apparent right from the start, great for leisurely drives down
the coast, but not for hard-core cornering. Spring rates are determined by the desired cycles/minute,
and during design, Id read that 120 cpm was about right. The problem is, its a very subjective
quantity; what someone says is too stiff, another person (me) thinks is merely firm. So having decided
they were too soft, the rates were upped to 145 cpm at the front and 140 cpm at the rear, translating to
200 lb. and 300 lb. front and back.
Dive and squat can also be controlled with anti-dive, anti-squat, rising-rate
suspension, or progressive springs, but none of these were used, in order to keep
things simple and predictable. It was a conscious decision back at the design
stage, to stay away from tricks more clever than my ability to understand or
controlI dont believe in getting something for nothing. Smith didnt care for
anti-dive, either, so I followed his lead and did not design any in.
Dennis, of dpcars.net, let me in on a rule of thumb: for a street-driven sports car, the wheel rate
should be roughly half the sprung weight. On Kimini, the front corner weight was 330 lb. Subtracting
unsprung components (wheels, tires, brakes, etc.) brought this down to 270 lb. Dividing by two gave
135 lb., and dividing that by the spring installation ratio of 0.9 bumped the number up to 149 lb.,
virtually the same as the 150 lb. provided by the equations in Millikin
11
. The same rule of thumb says
that for a race car, however, spring rates will be equal to unsprung weight, or 300 lb. at the front. So,
the nebulous answer is somewhere between the two.
While this rule was spot-on at the front, it was not so at the rear, where it showed that my original
numbers had gone wrong somewhere. I had 185 lb. springs in the rear, yet the rule of thumb, and
Milliken, recommended 220 lb. Im not sure what happened. Anyhow, I made a judgment call,
deciding to increase both roughly 1/3, from sports car rate toward race car rate. The front springs
were increased from 150 to 200 lb., and the rears from 220 (what it should have been) to 300 lb. It
was rationalized that the springs are easy to change if I goofed, and going from 150 lb. to 200 lb. at
the front meant that a given bump would feel only 25% stiffer.
Yes, 25%, not 33%its a statistics thing. With the 150 lb. spring, a load of
150 lb. compresses the spring one inch. Changing the spring to 200 lb. means the
same 150 lb. force compresses it only of an inch, 25% less.
Raising the rear spring rate more than the front, though, meant that the car was more likely to
oversteer. Throw into the mix that the car had a fair bit of understeer on corner entry, and it got harder
to guess what was needed. Reading stories about mid-engine cars going off the road backward made
me a little scared of the thing; I would have to wait for the second autocross to see how the new setup
worked.
At the first autocross, Id like to say I got top time of day on the first time out, but it didnt work that
way. Between being 20 years out of practice, driving a car with unknown handling traits and old tires,
and fearing cone damage to the composite, I only did okay. Actually, Id sum it up by saying the car
did much better than the driver!
The concern about possible body damage was real; I did indeed overcook a turn and drifted right
through a line of cones. It wasnt all bad, though; it showed that the car could slide and still be
controllable. The rear weight distribution everyone said would be a problemwasnt. About the
damage, its hard to see here, but it turned out that only the body filler was cracked, and not the
carbon. Epoxy was injected into the crack and the area clamped until it cured.
Second try
At the next autocross, we did much better: the car, due to stiffer springs and increased camber;
and I, due to calming down, getting more time under my belt, and concentrating more on driving and
less on worrying. The car performed well enough that I decided anti-roll bars would not be needed,
so that was one less thing to deal with.
Im still coming to terms with the nuances of how Kimini handles. Coming from front-engine cars, I
needed some time to understand the new capabilities, and lots of track time to learn how to implement
them. Coming into turns at ordinary speed, on the brakes, turning in and getting back on the gas, it
showed understeer. Giving it throttle mid-corner increases the understeerunless more throttle is
added, but thats not smooth. To take full advantage of it being mid-engine required a different
approach.
Having now driven the car on the street and on track, I have some comments
about how I designed the rear suspension to toe inward in both bump and droop
I outsmarted myself.
At the autocross, I had noticed some oversteer on corner entry. At first I
thought it was due to trail-braking, but with more driving I found it was
happening earlier, as I got off the gas and before getting on the brakes. Having
thought it over, I think I now know what was happening.
Approaching the corner the car is under hard acceleration. Under acceleration
the rear suspension is in bump, so, true to the design goals, the wheels are toed-in
slightly. As I get off the gas the car assumes a more neutral stance, so the toe
returns to zero. The problem is the chassis views this not so much as going from
toe-in to zero, but in the direction of toe-out, so for a moment it unsettles the car.
I havent changed it yet, probably because Im getting used to it. What this
confirmed was the importance of designing predictable suspension. Having the
suspension do things for you isnt necessarily a help because its trying to
control the car instead of you. I was smart enough to stay away from anti-dive
and anti-squat, but this one got me. Next time Ill design for no rear toe-change,
instead just adding some static amount for stability. It wasnt a big deal but it was
an unexpected feature.
And before the anti-IRS crowd jumps all over this, it is no different than
purposely designing in a certain amount of toe-change at the front.
The quick way around the track took a while to learn: coming in slightly faster than usual and using
trail-braking, causing the rear to come out slightly. Getting on the gas then results in no understeer
in fact, with as little or as much oversteer as desired. I had to get used to coming in faster and getting
on the brakes harder. My old cars handleddifferently. A mid-engine car has the capability to be
faster, but only if its used on its own terms; it takes getting used to. Kimini is still waiting for me to
learn how to best use all of her capabilitiesIm still all over the place.
It was very humbling competing against lesser cars, ones driven by more
experienced drivers, like getting squarely beaten by a young lady in a modified
Honda CRX. I realized by building Kimini that I thought Id instantly beat
everyonelike I had bought my way in to automatically winninghah! Nope,
driving fast is very difficult to learn to do well. For me at least, it is a long
learning curve.
41: At the Track
After Kimini had been on the road for a year, I finally built enough confidence to take her to a track
event. The autocross events had shown that the design was sound, but what was still unknown was
how the car would do at higher speedon a real road course. The nearest tracks are about 150 miles
away, so first was the issue of how to get there. A trailer was designed and built (see the Trailer
chapter), so I didnt have to worry about how to get home if the car broke down. The two track
candidates were Buttonwillow, near Bakersfield, California, and Willow Springs, north of Palmdale,
California.
I hadnt been to Buttonwillow before, but videos showed that it is a pretty fast track. Taking the car
out on a high-speed road course made me uncomfortable, at least for the first time; my concern was
the possibility of highspeed oversteer. The other track, Willow Springs, consists of several tracks.
The large track is for high-speed events, but the smaller track, Streets of Willow Springs, looked
promising: a 1.8 mi. track with 1015 turns (depending on configuration), offering more corners and a
lower average speed. This seemed like the perfect baby step before tackling the big track.
Streets of Willow Springs
We loaded up the trailer and truck the day before and drove up that night so wed arrive at the
drivers meeting on time.
We meant me and my coworker, Craig. He had been one of the book
manuscript reviewers, and when I said I was going to a trackday event, he
volunteered to crew for mecool! He drove a Mazda Miata, taking it out on a
track for the first time. Though he still doesnt believe me, the guys a natural. I
was very impressed with how well he drove.
These events are not true road-racing events; theyre called High Performance Driving Events
(HPDE) or simply trackdays. Just about any car can enter, and nearly anything can show up:
Toyotas, Fords, Ferraris, Porsches, Hondas, Subaru STis, Evos, and the occasional home-built car.
The main difference between real road-racing events and trackdays is that trackday events arent
timed. To help keep things safe, passing is allowed only on the straights. This prevents most mishaps,
though it cant prevent a driver from doing himself in by dropping a wheel into the dirt and
overcorrectingsometimes rolling the car.
So, I approach these events carefully. There is no demand to go flat out right from the start, like at
an autocross. Since the events usually arent timed, the killer instinct to get that last second out of the
car right out of the gate is not necessary. Drivers have all day to learn the track, so I had no reason to
go out on track with my hair on fire.
Here we are after arriving, the car on its home-built trailer. The trip up was uneventful; any
concerns I had about towing were unfounded. We arrived just before the drivers meeting, around
7:30 a.m. This wouldnt have been so hard had we not arrived at the campground the previous night at
midnight, due to road construction. Camping right next to the railroad tracks would have also
provided for late-evening entertainment, had I not been so tired!
Since all preparation had already been done at home, the car was backed off the trailer and we
headed off to the drivers meeting. With about 40 drivers, it was shaping up to be a fun day.
The three-wheel vehicle is Dave Nortons Shrike. You can read all about it
at Dave Nortons site: http://www.maxmatic.com/shrike.htm.
The advanced group ran first, which gave me time (being in the intermediate group) to check out the
track and the competition. Finally, it was my turn, and nervous as always, I belted myself in and
headed out with our group. Since I had all day, there was no hurry to risk disaster by pushing things,
so I took my time coming up to speed. A few laps were run at low speed to warm up the oil, coolant,
and brakes. And then it was timeafter 10 yearsto see how my design decisions panned out.
In my experience with my Datsun 1200, when it was set up for autocross, it
needed a fairly stiff rear anti-roll bar to help it rotate into turns faster. This
advantage at the autocross, though, turned into a problem at track events. There,
the stiff rear end became a handful on the high-speed turns, with the car now
wanting to spin. It was solved by disconnecting the rear anti-roll bar at track
events. Would Kimini exhibit the same trait?
The catch is that Kimini doesnt have anti-roll barsfront or rear. If it didnt handle well, the only
chassis tuning would be to change springs or shock settings. Changing shocks is only good for
transient situations, which left springs as the only tuning adjustment. Since it handled pretty well at the
autocross, the concern was that my old friendoversteerwould arise again, due to the higher
speed. So it was with great trepidation that I slowly increased speed.
The Streets of Willow Springs is a mid-speed track. We ran a counterclockwise configuration,
running the outer-most portion of the track. This included using part of the skid pad as the left turn,
past start/finish. With an elevation change of about 100 ft. coming down the hill past start/finish, it
meant Id be reaching speeds of about 110 mph heading into the skid pad.
Gradually, I increased speed, and Kimini maintained her composure. The skid pad corner was
especially educational, showing that the handling was very neutral. Throughout design and
construction, Id been told by many people that an 80 wheelbase car, especially one with 41/59
weight distribution, would spin like a top. It simply wasnt true.
Finally, I had the car going as fast as my rusty 20-year-old driving skills would allow. On corner
entry, trail-braking could bring the back end out slightly, and then the attitude could be controlled
entirely with the gas pedal steering with the throttle is lots of fun!
The only drama was of my own doing. This extreme drift/spin I blame on Max, the photographer,
who requested a good photo opportunity. Not wanting to disappoint him, I came into the skid pad a
little too hot, late-braking to kick the back end out. Well, it worked better than planned; the tail went
out, and out, and out! This slide ended in me spinning the car, though with so much room, there was
never any danger. It did teach me where the edge of adhesion was, though!
The car sticks very well; as long as I dont get off the gas suddenly mid-corner, a slide can be
maintainedto a point. The more extreme the slip angle, the more power it takes to maintain the tail-
out stance. At some point, I run out of power, and around the tail goes. Even though it was unplanned,
it was very enlightening. (Image credit Max Nealon.)
After a couple of sessions by myself, I paid back some of my buddies who had helped out during
construction. Dave, owner of the three-wheeler; Max, the photographer; and Craig, manuscript
reviewer, were all given rides. The most excited was Max, who later offered this commentary:
Its tight and I am not small (6-2 and 215 lb.). I barely fit, but that was not going to worry me.
The five-point harness held me tightly. My legs where restricted by a diagonal tube that the dash
is connected to, and my feet had inches of movement only. Who cares? This is not a luxury barge;
this is a race car (Kurt maintains that Kimini is a toy; I disagree).
After flicking switches, the starter button is pressed and Kimini burst into life. There are a few
vibes through the cabin and you can actually feel the exhaust pulses. Snicking her into first, we
move off.
At this stage I knew I was in something special. From the passenger seat, I could feel the road
surface (not in a bad way). There was no crashing and banging, just information through the seat
of my pantsliterally.
We drove sedately onto pit lane and waited to be given the signal to join the track. When we
accelerated up the pit lane, I remember feeling slightly disappointed for a second or two. Kurt
changed into second prior to my expectation of the VTEC fury.
Then he nailed it.
All hell broke loose. The mechanical fury behind us defies description. The VTEC howl is almost
feral. Kimini squatted on her suspension (maybe an inch) and hurled us up the track. Imagine
being strapped into a jet fighter on full afterburnerssame sort of thing (I have been there).
Then he dabbed the brakes and turned into turn one on cold tiresslightly sideways,
controlled nicely, onto the power short straight and then there is a little kink: you turn right, then
left, in a short period, then rocket up to the right-hand bend.
As we accelerated into the kink, the acceleration was intenseI had heard about the 060 in
under four seconds, but boy oh boy, its brutal. Then the (slightly early) braking, and then power
through the kink, the car just changed directionas simple as that; incredible, almost unnatural.
The rest of the lap was amazing; the experience is beyond my ability to explainthe noise, the
smell, the G-forces, even the tastes! We had a number of slightly sideways momentsKurt was
trying hard, but not showing off. I had watched him on previous laps. He was trying to ignore my
presence and enjoy it.
The view out of the (tiny) screen is best described as Nintendo. As Kurt hit the loud pedal
(literally, as Kimini was running an open exhaust), there is an attitude change in the car and she
just hurls herself forward. Each gear change seems to have no real effect on the accelerationit
appears to be just as brutal in second as fourth. The view blurs in a warp drive fashion as he
gets into the power. VTEC is not seamless at all: the exhaust takes on a harsh (in a good way)
note, and the urgency with which Kimini drives forwards is increased almost exponentially.
But thats not the dominant factor. The car has an amazing ability to change direction; so
sublime are her abilities that I would hazard a guess that she is actually underpowered.
Sure, grip becomes a factorthe tires are old, and a mere 13but the chassis control is
beyond measure; more power, more grip, and more brakes would not require a single change in
the suspension.
We did several laps and had fun with a very well-driven Scooby (with over 400 hp, expensive
suspension, and massive Wilwoods). On the back straight, I saw 107 mph on the speedo (digital).
Under braking, the harness did its job; in the corners, I was thankful that the seats have amazing
bolsters. Generally we could have run quicker than the Scooby, but as Kurt was still not entirely
sure of the track, it was a good policy to follow a quick car and also an impressive demonstration
of Kiminis abilities.
Max is correct about the brakes, and they were later modified. The tire issue
will solve itself; once worn out, theyll be replaced with stickier rubber.
The end of the day was a good opportunity to examine tire wear. While I didnt bring a tire
temperature gauge, how the tires wore across the face was very informative.
The tires are constructed with the outer rain grooves shallower than the inner grooves. As the tire
wears, the outer grooves wear down first, hence the appearance of the outer half being bald. In the
bottom-left pictures, you can see small holes in the bald areas. They are depth indicators, which show
that the wear on the outer half of the tire is the same as on the inside areatheyre actually wearing
evenly.
After the event, I was very happy with both the handling and tire wear. I felt that the suspension
design had proven itself.
I really need to get a tire temperature gauge, though.
Laguna Seca
Its fitting that the last event in the book is at this track; this is the trip that I had thought about all
through construction. Back at the beginning of the project, I dreamed of someday camping at the track,
next to the car, much like the barnstormers did in the 20s and 30s.
Over the course of construction, though, I came to the realization that while driving the car to the
event would indeed be like the old days, I lived a long way from the track. In England, tracks might
be 50 miles from home. In my case, it was 400 milesso I towed it.
Unlike the Streets of Willow Springs, I went solo, not even with Cooper. I didnt think hed enjoy
the long drive, laying on top of all my race stuff behind the seats. I also didnt think hed enjoy being
tied up in the pits while I was out running the car.
I was reluctant to tow the car through Los Angeles traffic, but it worked out fine. To be at the track
early on January 2 meant towing the car up on January 1New Years Day. Since Id be passing
through Los Angeles early in the morning, traffic would be about as light as it gets. I debated which
way to go, up Highway 101 or the potentially faster, but longer, I-5. Either way, I planned to be on
Highway 101 at Paso Robles.
About 30 miles east of Paso Robles is a very lonely Y-intersection, where Route 46 and 41
converge. It was at this spot that actor James Dean, was killed while driving his Porsche Spyder to
Laguna Seca, where he was going to race. I wasnt going to stop, but I was ready for a break. I have
mixed emotions about the tribute, feeling that there are better-deserving souls, but okay.
Unbeknownst to me, the track event was hosted by the Golden Gate Lotus Club (in the on-line
registration, the organizing group wasnt listed). There were many diverse cars at the event, including
the: Honda S2000, Mitsubishi Evo, Subaru STi, 350Z, Corvette, and many Lotus cars. I considered
myself very lucky. I would finally get the chance to compare my design to the one car that best
matched Kiminithe Lotus Elise. On paper, Kimini should be faster, since shes 15% lighter, with
the same power. But the Elises werent running on 10-year-old rubber! I was curious how my design
would stack up against Colin Chapmans (in spirit) latest product. I hold him is very high regard, so I
looked forward to the showdown.
Next morning we all gathered for the drivers meeting and covered the usual material. This group
ran the event slightly differently than others. In a typical trackday, passing is only permitted in the
straights. This was true here, too, but only for the novice group. The intermediate and advanced
groups were allowed to pass in most, or all, cornersno thanks. The thought of balling up Kimini in
an effort to win (whatever that means) seemed kind of pointless. Laguna Seca is a faster course than
Streets of Willowpeople have died hereso I was in no hurry to come up to speed quickly.
Besides, I wasnt in a National Runoffsthis was for fun.
The organizers had strict rules about out-of-control drivers. For example, if you go off-course, you
have to sit out your next scheduled track time. If you go off-course a second time, you are politely
asked to leave. This makes sense, since it shows that the car, driver, or both are not in control. Taking
themselves out is one thing, but taking someone else out, too, is bad news.
One quirk of Laguna Seca is that, because its in a nice area of the country, there are many
expensive homes nearby. Therefore, homeowner complaints about noise (even though the track was
there before they moved in) are taken seriously. There is a strict 92 dB noise limit (which is very low
for a race track!). Drivers get two warnings; after that, theyre sent home. But because I had the
tunable SuperTrapp muffler, I hoped that it wouldnt be a concern.
Finally, it was my turn to head out on track. With 25 drivers in each group, it promised to be
entertaining. Even though I knew the track had changed since my last visit 20 years earlier, it still took
a while to get used to it. The changes were not to the track surface itself, but rather to the run-off
areas; theyre now much biggergood!
I took my time coming up to speed, both to get somewhat reacquainted with the track and to allow
all the fluids to come up to temperature. Then, I finally got on it, at least in the straights. Between
turns 5 and 6 is the dreaded sound meter. Sure enough, the very first time I came by at speed, I was
immediately flagged for a noise violation (95 dB).
When I tried removing some baffles from the SuperTrapp muffler, the internal nut on one of the
screws broke loose. This led to endless entertainment in the form of forever loosening one screw and
seeing that it would not come out; the only solution was to break the head off. At the end of the day I
discovered that the blasted thing was trying to disassemble itself.
I believe that one of the screws holding the end cap on came out (big surprise), and that exhaust
back-pressure pushed the entire end plate outward. Im very disappointed by this muffler; it makes
promises it cant keep, and I no longer recommend it.
What I should have done was better prepare for the noise problem. A couple
of elbow bends can redirect the exhaust away from the meter. I could have used a
180 U-bend to point the exhaust under the car (since my exhaust pointed straight
at the sound meter). Unfortunately, its not easy to make a suitable end cap for a
SuperTrapp without the adapter restricting the exhaust even more. I was lazy.
Muffler issues aside, Kimini did excellently. The brake upgrade made a night-and-day difference in
pedal firmness and especially by instilling much greater confidence. It was now much easier to brake
later and harder. I was late-braking far deeper into the turns than anyone else. Its hard to underscore
just how important a hard brake pedal is!
What also finally sunk in was just how fast Kimini can change direction. She really enjoys fast
transition corners, where the car must be flicked very quickly left/right/left.
There were some very fast cars there, including this supercharged Lotus Exige. I fully expected to
get my butt handed to me by it and various Porsches and Corvettes. What happened, though, was that
the really fast cars were in the advanced group, so I didnt run against them.
I assumed that they were faster but didnt have time to clock them.
Many trackday events no longer time cars, apparently for insurance reasons.
But I doubt any insurance company would cover a car at such an event. The fact
that the events arent timed is, I think, not enough to satisfy insurance companies
looking for a way out of paying out a claim!
The Exige turned out to be in my group, but it had problems. I was, however, able to run against
some Lotus Elises, to finally answer the question of how Kimini stacked up; Kimini kicked their butts
all of them!
Kimini handled so well that I wont change anything in the suspension; it was easy to place it
anywhere desired. I hope that installing new stickier tires doesnt unbalance the car.
Later, when I timed myself (while editing the video), I was disappointed; everyone in my group
was slow, including me. Reading that a well-driven BMW M3 can do a 1:36 was pretty humbling.
My brother said it looked like Id adapted a fast in, slow out method of cornering, and I have to
agree with him. It tells me I have a lot of learning to do.
Asked for my reasons for the slow lap times, I came up with the following: old tires, lifting past the
noise station, and most importantly, not getting on the gas early enough on corner exit.
I have to think hard about why Im doing that. I wonder if its a deep-set fear of power-oversteer,
of going off backward; the walls here are very hard concrete! It certainly means that I need to learn
how to drive better.
It was recommended to me that a GPS data logger would be far more helpful to
improving lap time than more horsepower.
Not a game
This Elise came down the corkscrew, turning right at turn 8A, at the bottom of the hill (the turn in
the background). All of us in the pits heard the car slide, and then it went quiet (as it slid across the
dirt), and then we heard a sickening Whump.
Thankfully, the driver was unharmed, but the car is probably totaled. Not only was it heavily
damaged at the front, but the back end was pushed over several inches, too. The Elise has a bonded
aluminum chassis, and if theres any serious damage to it, the entire chassis is typically written off.
I do not post pictures of others misfortune as entertainment. It is here to
underscore the fact that, while this isnt professional racing, things can and do
happen. It is up to each of us to answer the question, What if this happens to
me? Most (if not all) insurance companies will not insure a loss that occurs at a
track event. Can you deal with the financial responsibility of repairing the car
with out-of-pocket money?
This is probably the biggest reason that many high-horsepower cars at trackday
events are never pushed to their full potential. I think its why dumpy old cars
tend to have the fastest lap times; their owners can better deal with the
consequences of potentially writing the cars off.
Oh sure, some expensive high-end cars actually do get very good lap times, but
I think its due to the drivers being really good, and rich enough to afford
repairing the potential damage, or both!
This unfortunate event aside, Laguna Seca was all that I had hoped it would be, regarding how
Kimini handled. It was a happy ending to a decade-long development, and I hope I get the opportunity
to run the car there again.
42: Say Cheese
Kimini received a lot of attention once finished. While it had been on the forums for years, only
after completion did interest take off. Simply driving the car to the gas station for the first time was a
new experience. It took about 15 minutes to get out of there because people kept coming over and
asking questions. They all wanted to know if it was a kit, getting funny looks on their faces when I
explained that, no, it was scratch-built. Now that it was on the road, I ended up getting some big help
with the photography, which was unexpected and by accident.
Once the car was done, a reader of the Web site, Max Nealon, wrote that hed like to see it. Since I
was planning to drive it over to its first car get-together, I suggested we meet there. In an effort to give
him directions with Yahoo, I inadvertently sent our home address! As I was preparing to leave, a
stranger drove up and walked into the driveway. My first thought was, Who is he? I dont have time
for this.
After he introduced himself, all became clear. This accidental meeting proved to be the best thing
that happened. He followed me as we drove over to the car event, where he said hed taken a few
pictures on the way over. This picture, at right, is the very first shot of Kiminis first drive on the
street. (Image credit Max Nealon.)
In the picture above, no, the cars not dripping anything; its a crack in the
asphalt. And yes, the left taillight does look crooked, but its the sunlight from the
left throwing a shadow.
Once at the event it became clear that Max wasnt just an enthusiast with a camera; he made a
living with it. Max accompanied the car to several car gatherings, plus Kiminis first autocross and
first trackday; his enthusiastic generosity was very much appreciated, and he is given due credit for
his work. (Image credit Max Nealon.)
Magazines
People were taking note, and I received inquiries from various magazines, wondering if Id like to
get the car in their publications. Sure!
C16
The first article appeared in issue #116 of C16 Autostyle. Having never seen the publication
before, I wondered what I was getting myself into. It catered to teens and import kidsyou know,
spiky-haired, tattooed, size-44 jeans kinds of kids. When I received the magazine, though, I was very
impressed; it used very high-quality paper and ink, plus the photography was top-notch. The article
turned out better than I could have hoped for, and frankly, I was amazed (and it was a little odd) that it
was my car I was looking at. I kept teasing the photographer during the shoot, asking when the hot
Asian models were going to show up; he kept saying, Real soon. I think he thought I was kidding!
Grassroots Motorsports
A photo shoot and write-up were also done (coincidentally the same weekend) for Grassroots
Motorsports. I was happy to get the attention of a well-known publication, but what really floored me
came later, when they called and said they wanted to put it on the November 2006 cover!
I didnt know what to think. It was so much more than Id ever expected. Never
in a million years did I think Id see my car on the cover of a magazine. My wife
said, Now youre famous. No, not really, I said, knowing that the clock on
my fifteen minutes of fame was ticking downI know how it works. While I was
very taken aback by the attention, I knew that in a few months it would all be
gone, but it was fun while it lasted!
The catch was that they wanted a kick-ass action shot. It took a solid six hours of scouting
various locations, driving back and forth as the photographer worked to get The Shot. Perhaps not
surprising is that the sweeping shots that instill motion in magazines are done at very low speed. This
provided the big benefit, though, of not having to worry about being arrested for speeding. On the
other hand, Id imagine that trying to get The Shot for a drift magazine cover could be fairly risky if
it were done on the street.
MiniWorld and Mini Magazine
These caught me by surprise. I had submitted my pictures and questionnaire to MiniWorld more
than a year earlier. Hearing nothing, I figured the magazine had changed its mind. It wasnt until
readers told me about it did I realize the issue had come out. The Mini Magazine article was even
more surprising, both because I knew nothing about it beforehand, and because it came out the very
same month as MiniWorld! (Turns out the freelance writer who had handled the Grassroots
Motorsports article had submitted it.)
Working with magazines
With my vast experience of working with these magazines, here is what I learned:
They contact you; you do not contact them.
You do not get paid.
Time between initial contact and publication can be months, or even a year.
You fill out a very detailed questionnaire, often weeks before the shoot.
Often, the writer/photographer is a freelance, not working directly for the magazine.
A photo shoot will take between one hour and all day.
Many action shots are, in reality, taken at very low vehicle speed.
Typically, the photographer retains all rights to the photographs, at least the ones the magazine
does not use.
The relationship is all business, but mutually beneficial. From them, you get publicity; from you,
they look to increase readership. Clean up the car and dress appropriately.
Camera work
At some point, youll no doubt want to record driving the car, be it on a scenic road at low speed,
or pushing the limits on the track. This means using a camcorder, with mounts ranging from sticking it
to the side of your helmet with Velcro, all the way to a lipstick-cam. For normal recording, I welded
a hard-mount onto the chassis, between the seats. A Bogen pan-and-tilt head (from
bhphotovideo.com) bolts to the mount. It allows pointing the camera in any direction, even adjusting
the height somewhat, yet holds it stiffly enough to avoid a shaky picture.
Cranking it up a notch is a lipstick camera, so-named because of its small size. Chasecam.com
makes this nice unit, which has various mounts available. The camera can be placed just about
anywhere on the car, with suction cups, while the camcorder serves as the recording device. These
tiny cameras are also handy for unusual applications. For example, the camera can be mounted to
point at the shock to see how much suspension travel occurs (like when measuring aerodynamic left
and downforce). The nice thing is that a microphone is wired separately, so you can narrate while you
drive, regardless of what the camera is pointing at. In other words, you arent forced to listen to what
youre pointing the camera at. If you want to get one, make sure your camcorder has an external video
input. It can be powered from either the car battery or its own battery pack (but not the camcorder).
Years ago, on the Datsun, we installed a full-size camera tripod in the back of
the car, duct-taping it to the roll cage. That was back when no one seemed to
mind.
Camcorder rant
When filming from outside the car, use a tripod if possible. Theres nothing
more annoying than trying to watch a wiggling, shaky, seasick video. Also, be
aware that as the cameraman, whatever you say will be picked up and recorded
for all time. The car owner might not appreciate you calling him a dickhead as
hes driving his pride and joy!
And finally, dont try to make an MTV video. Many videos posted to the
Web have the camera moving this way and that, zooming in and out, tilting left
and right, and in general, trying to make people ill. Youre not making a music
video; if people want that theyll watch music videos. People watch your video
because they want to see how the car looks, handles, and sounds, not to listen to
heavy metal. If you must have music, end it when the engine starts and consider
offering a music-free version as well.
43: Post-Build Depression
After years of building, the car now done, the resulting depression really caught me off guard,
probably because I had never been exposed to itthat is, spending such a long time building
something that, once done, my head was filled by a strange emptiness, listlessness, and a feeling of
indirection and lack of focus. For years and years the goal was clear: finish this part, then finish that
part; buy these parts today so Ill have them by next weekend. Later, the goal was getting it on the
road, driving it, and testing it, andsuddenlythis was done, too.
For quite a while after completion, I would get home from work, fully meaning to get some task
done, only to spend the evening on-line, keeping the Web site up-to-date, visiting various forums,
checking e-mail, or writing this book. Worse, Id find myself checking the same sites over and over,
thinking, Maybe someones just posted something. (While I dont use text messaging or even use a
cell phone much, I see other people with this same bad habit, checking for messages every three
minutes.) Anyhow, this lasted for months.
Writing the book, while the car was being built and tested, was the only way to
make sure I recorded my thoughts while they were still fresh. Had I waited until
everything was done, (as if it ever is), a lot would have been lost or forgotten,
only to be remembered long after publicationOops, forgot that, sorry.
This brings up a hard question: Am I a driver or a builder? I think if I have
to ask the question, I already know the answer, but Im not sure what to think
about it. Saying that Im a builder means that Ill never be happy with the car, that
theres always some unfulfilled idea. I dont think I fit completely into that
category. To say that Im a builder means that Id be happy to see Kimini go to a
good home, so I can move on to the Next Big Thing; only, thats not really true,
eitherat least not right away.
A driver is someone who doesnt want to work on the car at allthats not
me. I find myself going into the garage to work on the car, only to end up backing
it out just for a minute and returning a couple of hours later, after having driven
it down the coast. So Id argue Im not a hardcore builder, yet I do indeed have
various ideas about what Id like to build next, so Im somewhere between the
two extremes.
As a build progresses (of any car, I suspect), there are always moments of,
Huh, if I did it this other way, that would have worked out better. Actually, in
the case of Kimini, there were various aspects of it Id like to avoid next time,
like the doors and windshield (which seems to argue for an Ariel Atom-type car,
or even a three-wheeler, doesnt it?). Im by no means unhappy with how the car
turned out; its just that Id do things differently next time.
Fortunately, after the car was completed, there were more tasks to be taken care of, such as
aerodynamics testing.
44: Lessons Learned
Most design-and-build books do something that bugs me: the author always seems to design and
build everything exactly right the first timeI just dont believe it. Its a shame, too, because I think
readers would benefit greatly by having the author sit down, get over himself, and quite honestly say,
I could have done this differently, or I could have built these better. This chapter attempts to do
just that, to review what could have been done better.
Engine heat
With the exhaust between the firewall and the front side of the engine, it is hard for the heat to get
out. This is made worse because the underbody panel cuts down on cooling. Adding cooling holes in
the engine tray would help, but since its a stressed panel I am reluctant to do so. The next time that
the under-tray is off the car, I may add some small cooling holes.
Header
The header is 321 stainless, so theres not too much worry about the header wrap causing it to
crack. People asked, Why not just use mild steel and coat it? The answer is that if it I were doing it
over again, however, yes, it would be made of mild steel and coated, for cost reasons. What I dont
know is what insulates better, header wrap or ceramic coating. I would also keep the collector higher
up off the floor, since its pretty close to the wires and cables passing beneath. The bottom cable at
bottom-left wasnt up to dealing with the header heat!
Muffler
I became unhappy with the Supertrapp muffler, especially how the baffle bolts seize up. While the
manufacturer stressed the importance of using anti-seize, it doesnt last. The supplied bolts extend
through the threaded mounts, directly into the interior of the muffler, exposing them to hot exhaust gas.
This dries out the anti-seize, which then drops off, making the bolts bind up when removed. I broke
three of the six, which had to be drilled out and the threads chased with a tap. They were replaced
with shorter bolts to prevent as much of the thread from being exposed to exhaust gas. The Supertrapp
promises the following qualities: compact, light, stainless, rebuildable, and tunable. Its tunability can
come in very handy when I exceed the noise limit at the trackremove a couple of baffles and Im
back in business. Ill be asked to leave, otherwise. When the car is taken to the road-race track, all
Ill have to do is remove the screws and Im effectively running an open exhaust, all without needing
to crawl under the caror that was the theory.
I finally lost my patience with the Supertrapp at the Laguna Seca trackday
event. Several more screws have seized, including one so tightly that it broke
loose the welded-on nutinside the muffler. The only way to get the baffles out
was to cut the screw off. And finally, the entire innards were pushed out,
including the entire end plate, partway out of the housing. Huh? Ive given this
muffler three chances and it has used them up. It makes promises it doesnt keep.
I later upgraded to a Walker Dynomax. Other possible choices are the Ultraflow and the Hooker
Aero Chamber. Regardless of the chosen muffler, the key is getting it to fit!
Accessibility
This is probably the number one issue and number one annoyance; if its hard to get at, its a pain to
deal with. If its a pain to deal with, maintenance will be put off. If maintenance is put off, problems
can happen and not be noticed, and then bad things can happen. Think carefully about where things are
in relation to everything else. For example, the front shocks are hard to remove because the lower
mounts are hard to access. Even worse is reinstalling them and getting the spacers back onto the bolts.
Thankfully, a spring change is a relatively fast process, since the lower shock
mounts dont have to be removed.
The engine under-tray is also very unpleasant; it is hard to reach the bolts along the bottom edge.
The inboard seatbelt anchors are also very difficult to access. If there is a next car, it will be
designed more with accessibility in mind and not for the last couple of tenths of performance.
The next time, Id also offset the fuel cell instead of putting it on the centerline of the car. There
was no good reason it had to go there, and shifting it to the right would have helped the center of
gravity: its weight off to the right would help offset my weight to the left. (The funny thing is, thats
why I put it on the centerline, thinking that since it was symmetrical, it would be good for the CG
oh well.) Another good reason is to give my big feet more space!
Composite layup
A huge amount of time and money was spent on the body shell, filling and sanding, all because the
layup wasnt vacuum-bagged. While its a pain, vacuum-bagged is well worth doing, since it prevents
hundreds of hours of filling and sanding later. If a carbon shell is not vacuum-bagged, it will likely
require spreading a thin layer of filler over the entire body to fix all the flaws. This makes the part
much heavier, negating any decrease in weight attained by using carbon in the first place (never mind
the extra expense). And finally, someone has to do the bodyworktime-intensive if you do it,
expensive if someone else does it. At the very least, use a gel-coat layer.
Ive had about a half dozen inquiries regarding selling shells. Yes, I say, Ill
sell one. Then we get to the issue of cost, and I never ever hear from them again.
A professional fiberglass layup (not chopper-gun) is around $4,000, and carbon
is around $5,000$6,000, not counting shipping.
Steering
My (subjective) recommended maximum range of values for a street-driven car is +/-30 for two
turns of the steering wheel, which turned out great. Its very quick and keeps my attention on the
freewaynot a problem once I got used to it. It has also helped me to not spin the car at the
autocross; since the steering is so fast, its very easy to catch a slide before it gets out of control.
Be aware that the wheelbase affects how sharp a turn the car makes for a given turn angle. The
quick-disconnect steering wheel coupler is very handy but introduces additional steering play no
doubt because I didnt pay $200 for the best coupler. The oversize rod-ends, intended for steering
shafts, are a little too oversize, introducing their own unnecessary play; I dont know if regular rod-
ends will fit. Finally, the steering U-joints that are held in place by only a set screw suck eggs
always loosening up. Im looking into either replacing them with the clamp-on type or switching to the
double-D steering shaft.
Fuel cell
The cell must be slid out the front of the car, after the radiator, ducting, and front bumper are
removedhad I made them easily removable! Next time, Id make the cell removable from below,
mounting it to a removable sub-frame. Next time Id also use an in-tank accumulator. While the
current setup works fine, the external accumulator, hoses, and pumps are more complex than needed,
more likely to fail or leak, plus the low-pressure pumps are noisy. Id just use one in-tank high-
pressure pump and be done with it. The only advantage with the high-pressure pump back near the
engine is that it minimizes the amount of high-pressure fuel circulating in the car.
Lastly, as said elsewhere, make sure to get a cell that can handle alcohol; more and more states are
adding it to pump gas. Not all fuel cells or foam is rated for it.
Seatbelts
Use 3 wide belts, in a five-point or six-point arrangement; narrower belts leave nasty bruises.
Having a five- or six-point harness is very important, not just for safety, but to
be able to feel what the car is doing. If youre sliding around, holding onto the
steering wheel to not fall out of the seat, youre compromising control of the car.
Because of our litigious society, aftermarket seatbelt manufacturers dealt with lawsuits by reducing
the stated life of race beltsto the edge of absurdity. They used to be rated for five years now its
two. Why? What changed? The questions Id like to ask are: (1) Did the manufacturing process
change to make them less reliable? and (2) Were people dying because their four-year-old belts
broke? I dont know for sure, but its oddly coincidental that its also a great way to increase
corporate income. Whats next? One year? Six months? One race? Where does it end? While many
amateur competitions dont worry about stated belt life, official events do, so dont get caught with
old belts. Oh, and the stated two-year life doesnt differentiate how theyre treated. It doesnt matter if
theyre kept in a cool, dark place or used every day and beaten on by the sun. One time, I accidentally
bought some brand-new belts that happened to be one year old. You guessed it, they were already
50% used up yet had never even been out of the box.
Each Simpson belt is tested to 3000 lb.; in the Earnhardt crash, theres no way
that they just failed. I made a point to buy Simpson belts for my car, as a show
of support. If you want to read a good report about the Earnhardt fiasco, check
out
http://en.wikipedia.org/wiki/Dale_Earnhardt:_Controversy_over_the_cause(s)_of_his_death.
The irony here is that seatbelts made for real racing are given a life of only two years, while the
belts on our Fords and Toyotas last the life of the car (and are used every day). Why such a disparity?
I had an interesting experience with the cam-type lock on a seatbelt harness.
On my old Datsun, I used the tried-and-true lever type of harness release. When it
was time to buy belts for Kimini, I considered switching to the cam type. They
seemed like an elegant design: just give the knob a twist, and the belts all pop off
at the same time. Before I went to the race shop, I read how one guy said hed
never use them because they can jam. Whatever, I thought, He probably had
some piece of junk belts.
Later, I went to the race shop and checked out the cam release on a new
harness. I slipped the belts into the cam housing one at a time, then turned it to get
a feel for how well it released, and it jammed son of a bitch! I couldnt
believe it, so I messed with it a while longer, and sure enough, about one time in
twenty, the cam release would jam; it simply wasnt possible to turn the release
mechanism. That scared me away from them on the spot. Thats why I will use
only the lever-release type.
Suspension
Shock ratios
For a given wheel movement, wed like as much shock travel as possible. Some installations result
in very little shock travel, which makes it difficult for the shock to give smooth damping control. But
nothing is written in stone, and many cars work just fine with poor shock ratios. At the front, the upper
rocker arm gives a 1:1 wheel-to-shock ratio, and at the back, its about 0.95:1. Just do the best you
can, but more is better.
Rising-rate suspension
I didnt use any fancy suspension tricksno progressive springs, no anti-dive, and no rising-rate
suspension; the goal was to have predictable handling.
Regarding predictable handling, see additional comments regarding my too
smart for my own good design of having the rear wheels have toe-in in both
bump and droop. I may have outsmarted myself a bit.
At the autocross, it became apparent my car leaned a lot (of course, a lot is very subjective).
This was of no surprise, since anti-roll bars were part of the design but hadnt been installed. I saw
no harm in trying the car out as it was, just to see how it would do. So yeah, it rolled more than Id
like, so anti-roll bars were needed. Of course, roll can also be reduced by increasing spring rateat
the cost of overall ride quality. Millikens formulas indicated that under 1 G deceleration the nose
will drop about 1.4, or 50% suspension bump travel. That seems a bit much, but the springs would
have to be stiffened quite a bit to make much of a difference.
As an example, with 150 lb. springs at the front, and the nose dropping 1.4, there is 150 * 1.4 =
210 lb. of weight transfer. Increasing spring rate to 210 lb. would decrease nose dive to 1.0, but now
the suspension is all-around 40% stiffer, quite a change in ride quality. Of course, nothing is written
in stone; its all up to the designer and driver to decide whats best. For a track car, lap time rules all.
For the street, though, its all up to what youre willing to put up with. You probably want a sports-
car-like ride but probably dont want to go all the way to race-car-hard suspension, since the street
has bumps, curbs, and driveways.
Suspension clearance
Make sure everything has room to move. At full bump, the wheels rub against the steel shell mounts
because I didnt give the wheels enough space. It wasnt a big problem, but it was annoying that it
wasnt done right the first time. (Stiffening up the springs had the side effect of fixing the issue.)
I do regret cutting things so closely on the uprights. The brake calipers are very close to the wheels
and get scarred by stones getting trapped between them and the calipers. Same with the top and
bottom bearingstoo close to the wheels. Next time Id give them around clearance.
Recently I saw a Ferrari with really big brakes. The calipers were very close
to the wheels, about as close as mine. I examined the wheels, and sure enough,
there were scratches left from small stones getting jammed between the two.
Now I dont feel so bad!
Caster: how much and why
Caster, like a caster-wheel chair, causes self-centering of the tires. It provides the self-return-to-
center feel of the steering wheel, and thats the catch: its a feel, which makes it a nebulous
personal preferencethere is no right amount.
The more weight on the front tires, the less caster is needed, and with the engine in the back, more
was neededthe question was, how much? There is no black-and-white answer; its whatever
amount you want, whatevers comfortable. I used an initial guess of 5, but doing it over again, Id
increase it to 79. As it is now, there is very little return-to-center effect. Does it matter? Not
really, but since its free to put in at design time, a bit more would have been better. The amount is
very subjective, and 79 would have been about right for me, for my specific car, with its specific
weight distribution. This is why I cant give magic values; its all subjective, and what I like, you may
not. So, the next time you hear someone saying it has to be this or that amount, just smile and nod.
Until you drive a car with it set to some particular value, the right amount is just an opinion.
Wishbone tubing
A lot of people say to use only DOM (Drawn-Over-Mandrel) seamless tubing. I would have, too,
had it been more available and less expensive; its about five times as expensive as electrically-
welded (EW) tubing. Overall, though, I couldnt find a compelling reason to spend the money, and
while its stronger, the EW tubing is strong enough, too; I just picked tubing sizes that worked. To be
fair, though, a nice feature of DOM is that, since the inside of the tubing is smooth, its easier to tap or
to slide threaded inserts into.
Passenger cooling
Having ventilation is a good idea. A defroster would be nice, though so far it hasnt been needed
(in Southern California). More than a couple of times I wished I had a heater, but its a convenience,
not a requirement.
Many people said, It must get really hot in there with the engine right behind
you. Yesand no. It gets warm in the summer but not for the reason everyone
thinks. Yes, the engine is right behind me, but the bulkhead does wonders for
keeping the heat and noise out of the passenger compartment. Its the combination
of having small windows, no air conditioning, and coolant pipes running forward
inside the center tunnel that makes it warm. On a long drive, enough heat comes
off the pipes to make the center tunnel unpleasantly warm.
The answer was to cover the pipes with insulating sleeving. At the moment Im
using cheap hot-water pipe insulation from a home improvement store, though Im
not happy with it because its flammable. Fiberglass or ceramic sleeve would be
much safer and wouldnt produce poisonous fumes in a fire.
Doors
While they should have just worked, they didnt. The shell has slight variations from side-to-
side, probably from the steel car the mold was taken from. The doors themselves had slight
variations, one of them so bad that I had to shorten the window frame by 1/4! The weatherstrip was
stiff enough that closing the door was an effort, and the sliding windows are pretty stiff (though these
have since loosened up). Yes, the doors offer protection from the elements, but they were such a pain;
if there is another car project, I may do without, and just step into the vehicleI disliked working
with them that much.
Side mirrors
I bought some nice-looking bullet-type mirrors they are nearly useless. The only mounting point
for mirrors on the Mk1 Mini door is far down the door, due to the window sliders. The result is that
the passenger-side mirror was only half visible. The driver-side mirror is okay, though even it is a bit
low.
Later, I learned that Mk1 Minis came from the factory with no side mirrors. For as much good as
they did, I should have left them off for a cleaner look, but now there are holes in the doors. One
possible solution is to use replacement mirrors that stick up higher and have large bases to cover the
old holes.
Magnetic numbers
For autocross events, I bought magnetic stick-on numbers. They look really niceif theyd only
stay on the car. Its not the manufacturers fault; the paint shop covered the doors with a thin layer of
body filler thats just thick enough to make the magnetic force a bit iffy. Sure enough, after one event,
someone came over and handed me my numbers, with tire prints across themnice. Not sure what the
solution is. Static-adhesive numbers? Strong magnets? Shoe polish?
Drivetrain movement
With rubber engine mounts, the drivetrain can move around a surprising amount. Just because
nothing gets too close to the engine at rest, it doesnt mean that nothing can hit it under hard
acceleration, cornering, or braking. Also be aware that because of this motion, CV joints that you
thought had plenty of clearance and free motion (not bottoming out) may not. (Credit goes to Kent
Collins for this important observation.)
Brake system
The brake system has its own chapter. Summed up: use hard-line everywhere possible, do not use
synthetic fluid, and try to use rigid calipers.
45: Aerodynamics
Given the brick-like shape of a Mini, why is this chapter even here?
It doesnt hurt to try to improve the car, even a little. During design and fabrication, there was
considerable freedom to try things, so why not see how good we can make something that, straight
from the factory, has a drag coefficient (Cd) of 0.50!
Before we get into this, heres a story about fast cars and The Big Picture:
Dennis Palatov, of dpcars.net, has had the opportunity to drive many different
cars on track. Detailed on his Web site is a comparison of two very different
cars, both instrumented with GPS equipment, on the same track, and under similar
conditions.
Car #1 is a 2006 BMW M5, 4000 lb. with 500 hp, giving 8 lb./hp, and totally
stock. Car #2 is a 2006 Ariel Atom, 1300 lb. with 245 hp, giving 5.3 lb./hp. The
Atom was on very sticky tires and track suspension, while the M5 was on
factory rubber.
Both cars were driven on Pacific Raceway in Washington, the catch being that
the M5 was owned by Dennis, while the Atom was not, meaning he took it easy
on the Atom because it wasnt his. So, which car was faster, the lightweight
fighter on semi-race tires, or the overweight highway cruiser?
The M5 was five seconds faster per lap! Dennis said that while the Atom
could eventually match the M5, it wouldnt be easy. That the M5 was so much
faster is even more amazing because it was on factory tires. Dennis discovered
that the equalizing factor was top speed, the M5 being 20 mph faster down the
straight. The Atom easily beat the M5 up to about 100 mph, but lousy
aerodynamics conspired to slow it down above that, while the M5 kept right on
accelerating.
So, when comparing cars, its important to look at the entire package. The
often-touted power-to-weight number is only part of the big picture. If a car has
poor aerodynamics, it can be beaten by something slower, even on a track
where handling should make all the difference.
Its not just power-to-weight; its also power-to-drag.
The numbers
McBeath
12
has a very helpful equation relating drag to horsepower:
Horsepower absorbed by drag = Cd * A * V
3
/ 146,000
where:
Cd = drag coefficient
A = frontal area of the car
V = speed in mph
For an Mk1 Mini, Cd is about 0.50. With the elimination of external body seams, it is certainly less
for Kimini, except that huge fenders were added. A guess is that it is now around 0.45. Frontal area
was measured at 20 sq. ft. The engine has a modified intake and exhaust, so figure maybe 165 whp
(wheel horsepower).
Plugging in the numbers, the equation says Kimini cannot go faster than about 135 mph, due solely
to aero drag. The irony is that Kimini is already limited to about that same speed, due to the small
tires, which effectively re-gear the transmission. A number of people suggested that I swap
transmissions to get a higher top speed, but the equation says that it wouldnt help, other than lowering
rpm at cruise.
Lets say that I increase horsepower from 165 whp to 200 whp. For all that effort, top speed
increases only 8 mph! Sure, acceleration is faster, but it was surprising to see how important drag is
at the top end. In the big picture, it doesnt really matter; this is more for curiosity than anything else.
Im not sure how often Id want to go 140+ mph in an Mk1 Mini anyway.
Above-body
Aero-wise, the worst offenders on the stock Mini are the external body seams: the vertical ones at
the leading edge of the doors, and especially the one across the top of the windshield. Fortunately,
this was corrected when the steel body was prepared as the buck for the mold; the seams were brazed
on the backside, imparting enough strength to allow them to be cut off. The remains were tapped down
and the depressions covered with body filler, resulting in a very smooth shell. Of course, the rather
huge fenders certainly dont help matters.
Below-body
As a clean-sheet design there was nothing keeping me from seeing if I could improve things.
Almost all cars have very bumpy undercarriages; the engine, transmission, suspension arms, brake
lines, bolt heads, transmission tunnel, rear axle, and engine compartment air all conspire to mess with
air flow under the car.
Being a street car meant not being able to lower it much below four inches, but a few tricks could
still be tried. Katz, New Directions in Race Car Aerodynamics, and McBeath, Competition Car
Downforce, were researched to come up with a chassis I hoped would improve the Minis dismal
aerodynamics. The answer was to keep radiator air out from under the car and to use a smooth floor
pan and a rear diffuser.
Cooling
Radiator
I thought if I picked a good radiator and an efficient duct design from an aero book
13
, there
wouldnt be any problemsit would just work. I could then take full credit for my brilliant aero
design, and, well, it didnt quite work out that way. While Id chosen a good radiator/duct layout,
thats not the whole system. For example, inlet ducting to the radiator is supposed to diverge at about
7. Maybe on a real race car but not a Minitheres just no room. Another issue is the very short
hood, with its high pressure area at the base of the windshield consuming much of the hood area,
made worse by a nearly upright windshield. This required forcing the air, coming out of the radiator,
to turn 90 upward, so it exited forward of the high-pressure air, near the leading edge of the hood.
Once I started driving, everything was fine; through spring, coolant temperature stayed at 180F
until the first time it was 90+F. On the freeway, coolant temperature rose to 210F30 warmer
than normalwhich wasnt good. This meant additional work was needed to improve the cooling
system, starting with the easy corrections first.
Since the body was finished, I didnt want to start cutting up the nose. Even if I did, there really
wasnt much that could be done. With the short hood, cutting the exit back further would offer little to
no improvement. Years ago, during design, there was a fallback plan in case of overheating: adding
ducts from the radiator exit out to the wheel wells. The big unknown was the pressure differential;
would the air flow as expected and in the expected direction? It was a bit tough to test, since after that
one hot day it cooled off.
As of early 2007, no duct modifications have been made. Its because the car
isnt much fun to drive in the heat and humidity, due to the lack of air conditioning
and its small windows. If the car isnt driven when its really hot out, it cant
overheat!
There is another aspect of the high coolant temperature, though: what exactly
constitutes too hot? I really dont know for sure, so it may not be fair to say
210F is too hot.
Radiator air flow testing
Curious about what exactly was going on with airflow through the radiator, yarn tufts were taped to
the hood to find out. There were two questions: (1) Was the bump at the front of the hood doing its job
of providing a low-pressure area above the exit grill? and (2) Was there any reverse flow, where air
was actually flowing down into the grill rather than up and out of it?
Here are two tufts on the windshield centerline. The interesting thing is the direction of airflow:
down and toward the front! Yes, indeed, there is a big bubble of high pressure air in front of the rather
upright Mini windshield, rotating downward against the glass.
Below, more tufts show decent flow up and out of the radiator exit duct. Note, though, that the two
tufts closest to the windshield, at the back edge of the duct, are bent forward and flat. Apparently this
is right at the front edge of the recalculating air bubble in front of the windshield. It isnt actually
going back into the duct, but rather into the area behind the duct (the grill is slightly larger than the
radiator exit duct, with the rear corners open to the area under the nose). This should probably be
blocked off. It indicates that there is low pressure under there, good for dumping radiator air into the
wheel wells.
Engine compartment and engine tray
Heat is the enemy here, so several tricks were tried to keep things under control. First, I hoped that
the diffuser at the back edge of the chassis would create some downforce, or at least create a low-
pressure area that draws hot air out from the engine compartment. The area above the diffuser was
opened up and screened in, the idea being that air will exit through this area, drawn out by the low
pressure below and behind the vent. It also lets air out that becomes trapped in the engine
compartment as it rushes in around the lower A-arm openings of the engine tray. (The loop in the
center of the panel is a tow hook, since theres nothing else to fasten to.)
At rest, the engine compartment would get really warm without the benefit of cooling air. Large
slots were cut at the top of the rear window, serving as an exit for hot air rising through the
compartment. It really works, too; putting my hand near the slots after a hard drive shows a lot of hot
air coming out. At speed, low pressure where the roof turns down draws hot air out through the vents.
If a wing is ever added, it will increase the low pressure and further increase the airflow. Lastly, the
vents serve as lift points (handles) when removing the engine compartment cover.
Lift/downforce
I planned to measure lift and downforce; the question was how, and as the various methods were
considered, they kept getting simpler and simpler. Finally, I realized that all I really needed to know
was how much the suspension moved due to airflow over the bodyforget about pressure
differential. That is, if the back of the car lifts , thats all I need to know; between that and spring
rate, lift could be easily found.
A camcorder, or better yet, a small lipstick camera, could be fastened in place, pointing at the
shocks as the car is driven down the road, with readings analyzed later. A scale could be taped to the
shock absorber shaft to measure displacement, so thats what was done.
Simple scales were attached to the suspension arms to see if the car had high-speed lift or
downforce. The front one stuck up through the grill, so I could watch it while I drove. For the rear, I
used the small lipstick camera and pointed it at my high-tech scale. The wooden shafts were attached
to the A-arms, moving up and down with them, to give a gross indication of what was going on. The
idea is that it doesnt matter what the scaling is, or even what the motion represents, but rather if its
moving up or down.
It didnt work at all.
With all the bumps in the road, it was impossible to see a trend. As Dennis pointed out, I was
trying to measure values that are buried in the noise. If I had thought it through, I would have never run
this test.
Im using 300 lb. springs at the rear of the car. Lets optimistically say that I achieve 100 lb.
downforce (very doubtful). The total force is 100 lb., or 50 lb. per tire. That means that each spring
compresses 50/300, or 0.16. I was trying to see this small signal while buried in at least one inch of
noise. It was an impossible task.
The right way to measure downforce is in a wind tunnelor just measure air pressure. The
problem is trying to calculate total force over the cars surface.
Regardless of the outcome, it doesnt dispel the strong feeling that Kimini sticks to the road. Ive
never driven any other car that felt so composed at high speed. A hint that something was done right is
that when passing a truck, Kimini is not pushed to the side by the wave of air coming off the truck.
Here are two pictures of the bottom of the car taken with the car on the trailer. At the top edge of
the top picture is the engine cover; the screened-in area is the engine compartment air vent, just above
the back edge of the diffuser. The bottom of the car is smooth all the way from the front. The square
hole is the engine oil drain.
46: The Trailer
At a car show, I was talking to my buddy Steve Coe about getting our cars to various events, and
regarding trailers, he said, Be careful what you wish for, which got me thinking.
When we first start racing, we typically modify our cars for the local autocross, drive to the event,
run the race, and drive homenice and simple. As our skills improve, we want to go a little faster,
eventually requiring a dedicated set of tires. If they are street-legal tires, we may install the tires at
home, drive to the event, run it, drive home, and swap back. If theyre racing slicks, we jam them in
the car, along with a jack, air tank, and lug wrench, drive to the event, and change the tires there. At
some point, however, it becomes clear that the car should be trailered. Everything is already
prepared, so by the time we arrive at the event, we are all ready to go; if something breaks, we can
still get home.
If I just had a trailer, Id be all set.
So having made the decision to get a trailer, the question is what kind to buy. There are open
trailers, single-axle, dual-axle, electric brakes, hydraulic brakes, and if its an enclosed trailer, there
are endless options on whats inside: shelving, lights, compressor, generator, bed, toiletthe skys
the limit.
If I just had a place to store the trailer, Id be all set.
Eventually a trailer is chosen and purchased, and now we have to figure out where to store it. If
you have a big yard, youre all set; the rest of us have to put it in the garage, next to the garage, or
worse, at a storage yard.
If I just had a tow vehicle, Id be all set.
Having a trailer means we need something to pull it with. This is easy if we already have a large
SUV or pickup truckor a real pain in the neck if we dont. So we save up and eventually get a tow
vehicle. Of course, now theres the issue of where the tow vehicle is stored if its normally not used
(in addition to the trailer). On the other hand, if it replaces your daily driver, its not an issue, though
getting 815 mph driving to work every day isnt so great.
Be careful what you wish for.
The first thing that happens when you go to connect it to the tow vehicle is that you find that the
tow-ball isnt the right size for the trailerrun to the parts store to get the right size ball. Once its
connected you find that some, or all, of the taillights dont work. Thats assuming, of course, that you
have a mating plug thats wired the same on the truckif its not corroded that isto fit the corroded
trailer plug. Most trailers have either electric or hydraulic brakes. If theyre electric, youll need a
controller.
Steve added that he never damaged his cars at the trackuntil he started bringing them on a trailer.
He suspects something psychological switches in his head, saying its okay to take bigger risks.
This whole exercise of loading and unloading all the gear reminds me of scuba
diving. From the time you decide to go to the time you return and clean up
consumes most of the day, all for about 30 minutes underwater. In the case of
autocrossing, its even worse: the day gone for three minutes of track time.
And after all that
I finally did get one, by building it. All the used trailers I looked at were in very sad shape: leaky,
rotting tires, bad wheel bearings, nonfunctional lights, no brakes, and rusty frames. Building a trailer
had its attractions: I could get exactly what I wanted for about the same price as a needlessly larger
and heavier used one. It could be made just large enough for the car, with everything else going in the
truck.
Okay, to be honest, the real reason is because I enjoy building things. Building
a trailer was an enjoyable change from the exactness of building the car. If youre
off , so what! Besides, it would help alleviate some of the post-build
depression.
Only, that turned out not to be true. When it came time to cut and weld the
tubing, I just couldnt bring myself to be sloppy. It ended up being built with the
same precision as the car.
So the first thing, much like with the car, was to make a list of requirements. The goals were to
keep the weight and size down so I could push it around by myself.
The tow vehicle
I didnt have a full-size turbo-diesel pickup truck, so it was a little irritating to read posts claiming
it was the only type of truck that would workhogwash. A truck, like everything else, is a
compromise. Is it your daily driver? How far do you drive to work? How often do you tow? What do
you tow? How much weight do you pull? How much truck can you afford?
Fortunately, I had built a small, light car, so the trailer could be kept light, too; a big truck wouldnt
be neededyet another perk of building small! I expected to use the trailer maybe six times a year,
not every day, so a compromise was possible. Since I already had a V6 Toyota Tacoma truck, I felt it
would be a suitable match for a 1500 lb. car and a 500 lb. trailer. So lets get on with designing and
building one.
The axle
The axle was the key to the whole project. Single-axle trailers can easily carry 3500 lb., so there
was no need for an expensive dual-axle setup (though they do look nice). A low trailer has the
advantages of better stability and easier loading and unloading. A traditional axle needs leaf springs,
which require clearance for movement, so alternatives were examined. An answer was found in an
axle design that uses rubber cartridges, coupled to a splined 4 drop-axle. This solved several issues
at once: it lowered the trailer by 4, provided adjustable ride height, and didnt require leaf springs
or shocks. This particular one is called a Flexi-ride rubber torsion axle and is available from
trailer parts suppliers; this one is from etrailerpart.com.
Brakes
Electric brakes were high on the list, too. While the towed car, plus trailer, would only weigh
about 2000 lb., it didnt seem like a good idea to count on the trucks brakes alone. Electric brakes do
add expensethe brake controller as well as the brakes themselvesbut their inclusion was deemed
too important to skip. My concern, being a novice at towing, was having the relatively heavy load try
to push the light rear end of the truck around and possibly jackknifing. Another big perk of trailer
brakes is that if the trailer ever starts fishtailing, the brake controller can be activated manually to
dampen the oscillation.
After towing the car 300 miles round-trip to the track, I completely recommend
electric brakes. Once calibrated correctly, I could hardly tell that a trailer was
back there during braking.
Taillights
LED taillights solve the annoyance of the bulbs burning out. Since they take much less current than
regular bulbs, theres no problem running them in parallel with the existing taillights; the blink rate
will not change noticeably. As with the car, an important safety benefit is that they light immediately.
Regular bulbs take about 1/10 sec. to light, and in that time Soccer-Mom-on-the-cell-phone has
traveled 8.8 ft. closer. The extra cost was deemed worthwhile.
Overall design
The goal was less is more, borrowing some tricks from the car design. For example, there was
no need to use big heavy rails in the side beams; a truss would weigh far less and cost less to build
and less to tow. No need to completely cover the bed of the trailer, either; leaving the center area
open allows access to the bottom of the car. Of course, that was of dubious worth, since Kimini had a
smooth under-tray, but since it kept the weight down it was incorporated anyway.
To keep weight down meant keeping the trailer as small as possible, so it was sized just big enough
for Kimini. To keep the trailer and car as stable as possible meant keeping it low. The 4 drop-axle
helped a lot, getting rid of the leaf springs and the otherwise required spring space. It allowed the tire
centerline to pass through and above the top surface of the trailer. The cross-tubes (parallel to the
axle) were placed under the trailer bed to keep them out of the way of the towed cars tires and to
keep the CG down. When the design was complete, the loaded trailer surface was less than the tire
diameter radius (13) above ground.
There are a couple of places on a trailer with pretty big forces. The highest force will be in the
side rails, with the weight of the car basically trying to bend the trailer over the axle. Thatll never
do, so the side rails have a second tube above them to form a truss, like the side profile of a bridge.
One big tip if you use the truss idea is to keep the top lower than the bottom of the doorso you dont
hit it. Of course, its likely that the trailer tire and/or fender will be in the way; its always something.
The other place of sizable forces is on the tongue, where the V of the tongue attaches to the bed of
the trailer. With 200 lb. flexing the tongue assembly, gussets were added to the bottom (on the tension
side) to keep the welds from cracking, just in case.
One trick I used when designing the trailer was picking tubing lengths that
divided into the standard 20 ft. tubing lengths sold by the steel yards. For
example, the trailer bed is 9.5 ft. long; the cross-tubes, 6.5 ft. long; and the short
cross-tubes, 18 long. The point is, if tubing length is flexible, try to pick lengths
so you dont end up with wasted pieces of material.
After taking a day to draw up plans, it was time to visit the steel yard. Visiting the steel yard again
reminded me of when I was buying tubing for the car years ago, and I was reminded (again) that it
doesnt matter what the plans call forif they dont stock it. Then, as now, it meant quickly
reviewing what was wanted versus what was available, and juggling tubing sizes to suit.
It doesnt look like an entire trailer, does it? But it is, minus the axle and flooring$187 worth of
steel.
The trick was to determine where to place the axle. A trailer rule of thumb is that 10% of the total
weight should be on the tow ball. Since Kimini weighs around 1650 lb. with gas, and the trailer
around 500 lb., that means around 215 lb. should be on the tow ball. But that still didnt answer the
question, where does it go? While I had calculated where it should go, it was critical enough to hold
off bolting in the axle until it could be double-checked. This way, when the trailers really done, then
the axle position could be placed accurately. Place a bathroom scale under the trailers jack and some
wood beams under the back edge of the trailer to keep it from rocking back. Block the tires so it
wont roll, then carefully drive the car onto the trailer and see what the scale reads. If its too high,
hammer the axle forward, and if too low, hammer it back.
A few people suggested aluminum diamond plate, and yes, it looks really nice. It had a
correspondingly nice price, too: around $500. Even surplus fiberglass grating was around $4/sq. ft.,
so that was out. After looking at many different types of decking material, nothing beat wood for the
price. It was surprising to find that with wood, you dont get more for your money with a larger piece.
For example, a 2x12 is not three times more expensive than a 2x4; its more like four times. I think
thats because a 2x12 has to come from a big tree while a 2x4 can come from a much smaller, younger
treea much smaller investment for the growers. Price-wise, I was unable to find anything that beat
wood.
Along with the axle came this box of partsno instructions, no diagrams, no schematics, nothing.
For example, theres a roll of colored wire, which I know is for the lights, but they dont bother
explaining what color should do what. Then theres the big roll of black wire. Thats apparently for
the brakes, though since its a four-conductor cable, Im not sure what the point is. The packing also
left something to be desired, what with the heavy hitch thrown in loose along with the LED taillights. I
call it Billy Bobs Trailer-rama Grab Bag OTrailer Parts.
Thank goodness I had experience building a car it would be needed! Its not like anything was a
real problem; its just amusing that this is business as usual in the trailer world.
Construction was slow but straightforward. I enjoyed building it, probably because I hadnt built
one before. After it was all welded together, there was, as with the car, the question of how to paint
it. Again, as with the car, I set out thinking I was going to paint it myself, then at the last second I
decided that powder-coating was best.
Since the axle contained rubber inserts, though, it couldnt be powder-coated, since once the
powder is applied, the part must go through a 400 oven. As I mentioned elsewhere, the axle was
painted with POR-15, then top-coated. After dealing with the axle, I was glad I sent the trailer itself
out for powder-coat. If you are unsure how to paint the car or trailer, be sure to include all the costs
of POR-15. The total cost isnt that much less than powder-coat. This trailer, including pickup, media
blasting, powder-coat, and delivery, was $350 well worth it in my mind. After the trailer was
painted and the wood decking added, the lights and brakes were wired up. With the trailer done, the
electric brake controller and receiver hitch were added to the truck, which took a day. It was then
towed in for DMV inspection and registration.
As for ramps, after checking the price of steel, I determined that it was actually cheaper to buy pre-
made aluminum ones. I would have spent more on the steel, paint, and wood; besides, the aluminum
ones are much lighter.
Doing it over again, Id more seriously consider buying a trailer. It was fun to
build one, but it took a lot of timetime better spent preparing the car for the
track or working on the book! Still, the trailer is exactly what I wanted; its light
and easy to move around, and it cost no more than a used one, which, no doubt,
would have needed a lot of fixing up. Regardless if the trailer is built or
purchased, dont forget the additional expense of a tow hitch, brake controller,
and ramps.
Epilogue
So what now?
With the car complete, theres always the question of what might be next. My thoughts are a bit
scattered (even more than usual), and it depends somewhat on the path Kimini chooses for herself.
What this means is: things happen, opportunities present themselves, and things come up.
For now, however, Im simply enjoying the vacation from the decade of buildinghopping in the
car and driving, be it cruising, a car show, drag race, autocross, or trackday. Its wonderful having
such a versatile car thats so adept at so many diverse venues. It may sound odd, and a little corny, but
I feel honored to own something like this, all the pieces coming together into something greater than
the sum of the parts.
The ultimate goal of the project, right from the start, was to take it to Laguna Seca Raceway in
Central California. Driving Kimini on that trackcoming down through the corkscrew was the
crowning achievement of the project, the materialization of a dream that kept the project rolling,
through the good times and bad. Making that dream complete, though, involves doing something I
havent quite come to terms withdriving the car to the track, and back.
Why would anyone do this? Because thats what they did in the old days and it would be quite
the odyssey. Yes, it is far more sensible to trailer the car, but the trip, much like building the car, has
little to do with whats sensible and more to do with an experience that will be remembered forever.
After that, who knows? As far as possible modifications go, theres always the pull toward the
dark side: increasing power through supercharging, turbocharging, or perhaps a high-rpm screamer.
It would be nice to routinely do single-gear pulls from a dead stop to 60 mph, and that means a high-
rpm engine. That depends how cheaply I could find a second engine, whether stock, rebuildable, or
already built. Or instead, theres always the possibility of switching to Hondas newer K20/22/24,
but theres no rush. Im not sure Id gain much power for the amount of work and money involved. Of
course, since the Prelude engine had only 12,000 miles on it when installed, it should last a really
long time, assuming that I treat it right. Then theres the alternative to keeping the car.
Ive said, only half kidding, that Im waiting for Jay Leno (who owns one of just about everything)
to buy Kimini. Yes, thats right, I might consider selling her. Blasphemy, you say? Is it because I dont
like the car? No, not at all, but during the decade of designing and building, there was always the
thought of what might be next. Having learned a tremendous amount during that time, there is the lure
of how that knowledge might be applied to a blank-sheet design. If I want to start another project,
though, two things hold me up: not being rich, and having little space (unlike Jay). Selling Kimini
would free up both.
The thought of selling the car is a difficult one, because Kimini works so well. The car is reliable,
is easy to work on, starts easily, and is easy to drive, just like a real car. Theres little preparation
for anything. Want to go to breakfast? Just go. Autocross? Check the tires and go. Cruising? Go. A
low-maintenance vehicleas it should be.
Id like to drive the car for several more years, to build a full set of memories and experiences, but
sometimes things come up; if someone offers the right price, who knows? If the car gets damaged,
maybe Ill modify it into something different. If the engine breaks, who knows what might take its
place?
Thats the beauty of the future, not knowing exactly whats coming, good or bad, but Im very
thankful I got to experience designing and building Kiminiits more than I ever hoped for.
The crystal ball
Heres my take on the future of home-built light sports cars.
With the popularity of Ron Champions Locost book and related Web sites, the Super-7 will
remain a very popular choice for builders. But with the diminishing supply of front-engine, rear-drive
drivetrains, this will change. Builders are continuing on with the old design, now forced to use heavy
pickup truck or V6 sedan components, while others are starting to adapt FWD engines, using the
engine in the front and adapting rear-drive transmissions and axles. Even so, the days of the
traditional Super-7 are numbered.
I find it humorous how much trouble builders go to, forcing a FWD engine into
a Super-7. The drivetrain cries out to be installed mid-engine, yet builders insist
on removing the transmission, cobbling up an adapter, finding a front-
engine/rear-drive transmission to use, and having to typically use a pickup truck
rear axle. So much work to recreate a dated design, but to each his own.
Whats about to happen is the transition to using FWD power plants mounted mid-engine. While
Locost builders portray themselves as freethinkers and designers, the truth is that they arent pushing
the envelope. Instead, they stick with the familiar, the tried-and-true (and free) Locost chassis plans
floating around the Web. But they arent taking the next step.
What will emerge is a car something like a Locost, but mid-engine. The Ariel Atom has certainly
stirred the imagination, but the difficulty of bending the tubes, and with virtually no weather
protection, I feel that designers will stay with the basic Locost layout and design, moving the cab
forward slightly to make room for the drivetrain behind the driver. The continued availability of
Locost cowls, fenders, and nose cones will make this a logical advance. It avoids tons of time dealing
with composite work and greatly speeds the assembly time.
The future is right in front of us; if someone doesnt take the next step, maybe I will. Then again, as
of late 2006, the price of gasoline is showing no sign of slowing its upward spiral. Perhaps well go
straight to sportbike-powered three-wheelers that get 50 mpg, yet accelerate to 60 mph in 4 seconds.
That would be pretty cool, so either way, the future looks great for building cars from scratch.
Ill close with an anecdote that I thought of many times during construction. Its about a guy at the
harbor, whom well call Bob, watching a fisherman, whom well call Tom, unloading fish from
his boat:
BOB: You caught a lot of fish today.
TOM: Yup.
BOB: You know, you could take people fishing.
TOM: Yeah, then what?
BOB: You could make more money and buy a bigger boat.
TOM: And then what?
BOB: Then you could take out even more people.
TOM: Then what?
BOB: So you could make even more money.
TOM: Then what?
BOB: So you could retire.
TOM: And then what?
BOB: So you could do whatever you want.
TOM: ButI am.
So into the sunset we go, off on the next adventure. Its been a wonderful journey.
Best of luck to you in whatever you decide to build, and I sincerely hope that you get to experience
the same incredible adventure that I did.
Sincerely,
Kurt Bilinski
April 2007
FAQ
These are common questions received during construction, and later, at car events:
Whats with the name, Kimini, and how is it pronounced?
My wifes name, Kim, plus Mini pronounced Kim-many. (Yes, I know its not pronounced
correctly!) As an aside, having a unique name is magic on the Internet, because its easy to search for.
When did you start the project?
Thinking about it started in 1995; design started in 1996; and construction in 1998, with the car on
the road by the end of 2005.
Is it a kit?
No, scratch-built. The carbon fiber shell was made by a friend, and this is the only one in
existence. I started with nothing but the shell; the only authentic Mini parts are the doors, windshield,
and lights.
So you put your own shell on an old Mini?
No, there are no original Mini parts other than the doors, glass, and lights. I fabricated the chassis
from scratch. (Many people have a hard time understanding it wasnt a Mini underneath.)
And it is street-legal?
Yes, registered as a 1961 Mini.
What cars influenced the design?
The sparse, raw aluminum look of 1960s race cars, especially the original Ford GT-40, and the
utter simplicity and functionality of the Lotus Super-7.
How long did it take to build?
About 35004000 hours total, but a lot more, if time spent thinking about it is counted.
Did you have a schedule?
No, schedules are for work; the car project was for fun.
Fine for you; I dont have time to build a car.
Sell the TV, or at least cancel cable service; television is a time-sucking anesthetic to creativity.
Use the money saved from the $200$500 annual cable bill to fund the project. There are now 1520
free hours a week to work on the project.
Whyd you use the Honda H22A1; I would have used a
It was the engine to have back in 1996, when the car was designed. There was no Honda K20, or
Hayabusa bike engine, or. When you build your car, you can use anything you want.
How much did it cost?
About $25K, including tools. Of course, tools can be used for other things, so how much of the tool
cost should be counted? Do I include labor? Good Heavens, no! If, on the other hand, we look at it as
entertainment, as in fun/$, it is cheap. How much do people pay for entertainment these days? Spread
over the full build, it cost around $7/day. Give up Starbucks and bring your lunch to work, and youve
paid for the car!
How fast does it go?
The most common question, with the answer invariably a disappointment; people seem to think top
speed equals performance, but to answer the question, around 130 mph.
Whats the 060 time?
Starting in second, smoking the clutch and hitting fuel cutoff results in 3.9 secs. (because I dont
have to shift). Shifting at a more reasonable rpm gives 4.6 secs.
What does it do the quarter-mile in?
My tach/speedo has a built-in quarter-mile timer, which, during a single attempt on a deserted road,
gave 12.7 secs. Since it takes two shifts (2-3-4), theres more time to be had because I shift slowly. It
could probably get down to 12.4 or so but thats speculation. Yes, I know, no time slip = no street
cred. On the other hand, the tach/speedo starts the timer as soon as the car moves, unlike real drag
racing, where the timer starts when the wheel rolls a foot or so in the timing box. The point is, the
actual time will be slightly faster.
Whats it weigh?
Roughly 1600 lb. dry; too heavy in my mind, but oh well.
How many miles per year?
Dont know yet, but based on what its done already, roughly 4000 miles annually.
Do you sell a kit or plans?
No, its too much work for the few people who would actually put down money. Besides, with this
book you can design and build your own. I would consider selling shells, though.
Whats it get for mileage?
I cant fill it up in the traditional sense, letting the gas station filler valve automatically cut off,
because it would overfill the fuel cell. Because of this, its hard to know how much fuel has been
added. A guess is 2025 mpg average.
When did you work on it?
Every Sunday afternoon for roughly 67 hours, plus holidays as time allowed. The plan was to
work on it a few hours each night after work, but it just didnt work out. For me it took too long to get
my head into the project. About that time the evening is already gone. While it was worked on only
one day, it was constantly in mind the whole rest of the week, setting goals for the next Sunday.
It must be really hot inside when you drive it.
All through the spring I wore a sweatshirt because it was cold. The center tunnel does get pretty
warm due to the coolant lines, as does the rear bulkhead window, but its not bad. Driving in the
middle of summer is a different story; its a small car with no air conditioning and small windows.
When its 95 and humid, its no fun, but thats true of any car.
It must be really noisy to drive.
Its about as noisy as Ill put up with. Its easy to talk to passengers, but its noisy under throttle. I
may look into some carpeting, because the steel floor no doubt bounces the sound around. I wear
earplugs at the track.
How do you fit?
Thats the beauty of building a car from scratch: you start with the driver and design the car to fit,
not the other way around. It fits like a custom glove because it is a custom glove.
Do you sell body shells?
The short answer is yes. The longer answer is, once material cost, labor, and shipping are added
in, the expense scares everyone away. A carbon shell would be around $6,000 shipped (2007).
Whats your workshop like?
One half of a two-car garagetoo small, but okay.
What do you listen to when building?
Prairie Home Companion , Reggae, World, 30s, 40s, 50s, and 80s music. To better hear the
music is an incentive to place the noisy air compressor outside. Heavy metal is bad when a steady
hand is required.
What do you do for a living?
HW/SW engineer, BSEE. Considered becoming an ME, but as much as I like mechanical stuff, I
enjoy the electrical side more. The advantage is that I can balance the HW/SW stuff at work with a
fun mechanical project at home, unencumbered by the usual work-related project limitations.
How much welding rod did the entire project consume?
Three pounds.
Whats your favorite beer?
Guinness.
What other car stuff have you done?
Built and raced a 1971 Datsun 1200 in the 80s. Raced means autocross and HPDE (I have no
interest in wheel-to-wheel competition). The Datsun taught the importance of light weight, beating
cars with much more horsepower because it could outhandle them. Helped my brother with his Datsun
510, 240Z, and Mazda RX-3.
What else are you into?
Astronomy, telescope making, gardening, beer making, Koi. Future projects might include a three-
wheeler, gas turbine, Van De Graff generator, flying saucer, and time machine.
Does your wife read your Web site?
Shed be bored out of her mind. In general, I dont talk to people about cars because I know most
people are completely uninterested. That is, unless they start the conversation, then you cant shut me
up.
How the heck are you still married?
Since I worked on the car only one afternoon a week, it was tolerated. My wife put up with a lot,
but at least knew where I was and what I was doing. Her gripe was about the money spent.
What are your two all-time favorite cars?
McLaren F1 and Talbot Lago T150C, two opposite extremes of the automotive world, yet fantastic
examples of creativity and engineering excellence. (Images credit McLaren and Pebble Beach
Concourse.)
Appendix A: Reference Books
Race Car Vehicle Dynamics, Milliken and Milliken.
Expensive, but worth it if you want to really get into serious design; the design book when it comes
to suspension. I put many of the equations in this book into Excel to figure things like center of gravity,
weight distribution, roll centers, roll couples, etc.
Prepare to Win
Tune to Win
Engineer to Win
Nuts, Bolts, Fasteners, and Plumbing
These four books were written by Carroll Smith, widely considered the expert in race car
fabrication techniques. These are absolute required reading for anyone building a car, or aircraft for
that matter.
Theory of Wing Sections, Abbot, Von Doenhoff
A very hardcore book on selecting the proper wing shape for a given task. I have yet to make use of
this text; its good but very technical.
Race Car Aerodynamics, Katz
Covers all aspects of airflow over a car and answered all my aerodynamic questions. A must-read
if you plan on driving over 60 mph.
Competition Car Downforce, McBeath
A must-have in addition to the Katz book above. They complement each other nicely.
Chassis Engineering, Adams
A good book, up there with the Staniforth books for simple, straightforward explanations of
suspension.
High Speed, Low Cost, Staniforth
Truly a classic and back in print through Amazon.uk. The book takes you back to the early years,
when a fellow could single-handedly design, build, and competitively race his own sports cargreat
reading of a simpler time. The book details how to design and build an open-wheel formula car based
on a Mini drivetrain. Theres no reason why you couldnt use this book as a guide and use a modern
drivetrain like a Honda or Toyota FWD assembly.
Race and Rally Car Source Book, Staniforth
Dont be misled by the title. This is a terrific sports-car design book and the one I probably got the
most information from. Very similar to his later book, Competition Car Suspension . Either one of
these is excellent.
Competition Car Suspension, Staniforth
If I were forced to buy only one book, this would be it.
Build Your Own Sports Car for as Little as 250 and race it!, Champion
An easy-to-read, clearly written book on how to build a Super-7 type car from scratch; it even
includes the drawings. Many people are building this car, called the Locost, and because of the
unusual name, its easy to use Google to find dozens under construction on the Web. The Web site
is/was http://www.championmotorcars.com/, where they sell/sold plans and kits (they appear to have
gone out of business, but who knows). BTW, the title is very optimistic when it claims it can be done
for 250!
(If you really want to build the above, check out the McSorley.net Web site. His drawings and
dimensions are far more accurate than those in the above text.)
Build Your Own Sports Car on a Budget, Chris Gibbs
The Champion above book has recently been replaced by this one.
How to Build a Cheap Sports Car, Keith Tanner
A different take on the Locost, using a manufactured chassis and a Mazda Miata as the donor.
Racing and Sports Car Chassis Design, Costin and Phipps
A somewhat dated book, but the basics always apply. Costin, by the way, was the Cos in
Cosworth.
How to Make Your Car Handle, Puhn
Dated but still very useful.
Metal Fabrication Handbook, Fournier
Sheet Metal Handbook, Fournier
Ultimate Sheet Metal Fabrication, Remus
For metal fabrication, nothing beats these three books. While the Carroll Smith books deal with
general race-car construction practice, these three books go into excellent detail about how to get
metal to do what you want. I learned a lot from these.
Race Car Fabrication and Preparation, Smith
Dated but very valid useful information for any type of race car. Deals mostly with heavily
modified road-racing Camaros and Firebirds, but it still applies.
Welders Handbook, Finch
Performance Welding, Finch
Formula Car Technology, Holmes and Alexander
Driving Ambition, Nye, Dennis, Murray
The story of how the McLaren F1 super-car was designed and built. Gordon Murray is my hero.
The Unfair Advantage, Donohue, Valkenburgh
Back in print, and it blew me away. Mark Donahue lived through the most exciting time of racing
development, and each page is like looking over his shoulder as he and his peers figured out
suspension design and aerodynamics for the first time. You have to read every single pagenot a
book to be skimmed.
One Good Run, Hanna (Amazon.uk has it as of this writing)
While the movie, The Worlds Fastest Indian, was wonderful, some creative license changed
the events and rearranged some details. Read how it really wasBurt Monro was an amazing
character.
Dwarf Car Technology, Smith
A crossover book from the roundy-round boys, as in circle-track racing. Just because they only
turn left doesnt mean they dont know what theyre doing.
A Strength of Materials and Statics and Dynamics textbook, available on-line or through
university bookstores.
Standard Aircraft Handbook, Edited by Larry Reithmaier
Acceptable Methods, Techniques, and Practices.
Aircraft Inspection and Repair, FAA
Standard Aircraft Workers Manual, Edited by Larry Reithmaier, Fletcher Aircraft
Moldless Composite Construction Homebuilt Sandwitch Aircraft Construction
Composite Aircraft Design
Light Airplane Construction for Amateur Builders, L. Pazmany
These books are all available through Aircraft Spruce (www.aircraftspruce.com). If its good
enough for aircraft design, its good enough for us.
And dont forget the factory service manual for your drivetrain. Not the weenie book sold in
bookstores or auto parts stores, but the real $100 factory manualits worth it!
Appendix B: Parts and Vendors
Race-car parts suppliers
Often we can borrow parts from circle-track suppliers. Since they have a much larger customer
base, their prices can be very competitive. Dont forget to check them first for things like shocks and
coil-over springs.
Summitracing.com
colemanracing.com
lefthanderchassis.com
stockcarproducts.com
speedwaymotors.com
Regular race shops
Impactracing.com (safety gear)
Ioportracing.com (various)
pegasusautoracing.com
Truechoice.com (Koni shocks)
Fuelsafe.com (fuel cells)
Jazproducts.com (fuel cells)
aa-mfg.com (lots and lots of brackets in case you dont want to make them)
racerpartswholesale.com (more for the formula-type cars, so check circle-track suppliers above
first)
Racerwholesale.com (safety gear, but check the circle-track sites above first)
Wilwood.com (all things brakes)
Tiltonracing.com (all things brakes)
Chassisshop.com (everything you need for your dragster, though many parts can be used by us)
Specialty suppliers
aircraftspruce.com (aviation parts supplier but many parts perfect for our projects)
aircraft-tool.com (tools made specifically for aircraft repair and very useful for car building, too)
ustool.com (tons of tools perfect for sheet metal fabrication; dont let your wife catch you with
this!)
burnsstainless.com (tons of stainless tubing for all your exhaust needs; also aluminum tubing)
headersbyed.com (header parts plus header design service)
Autometer.com (gauges)
Eastwood.com
Minimania.com (Mini parts, what few I bought)
Jd2.com (tubing notcher)
Coastfab.com (when you need exactly the right bolt or piece of hardware; also sell exhaust parts)
Magnumforce.com (cheapest place I found for mandrel mild-steel bends)
Push-pull.com (where I bought my push-pull cables, made by Cablecraft)
Mossmotors.com (a good on-line source for British car parts)
Artmorrison.com (hot-rod parts supplier; again, many parts will work for us, too)
summersbrothersracing.com (custom axles and other drivetrain parts; good prices)
Bluecoilspring.com (great prices on coil-over springs)
Acestainless.com (everything stainless: tubing, plate, elbow, tees, bends)
grafityp.co.uk/graficast%20series.htm (self-adhesive film, used in place of paint?)
Specialty hardware
kitplaneparts.com (structural rivets and aircraft parts)
wicksaircraft.com (rivets and aircraft parts)
ddaircraft.com (structural rivets)
bonaero.com (structural rivets)
crawfordproducts.com (rivets)
hansonrivet.com (rivets)
Industrial parts
McMaster.com (virtually everything industrial: hardware, chemicals, industrial supplies, etc.)
Harborfreight.com (cheap Chinese-made tools but quite suitable for home shops)
Electrical parts
Digikey.com (virtually everything electronic you could ever need)
Alliedelec.com (good source of military circular connectors)
Appendix C: Web Resources
Useful Internet sites and forums
locostbuilders.co.uk
Home builders of the Locost Super-7 clone, plus other scratch-built cars. Based in the UK, but
people are building these all over the world.
locostusa.com
Similar to the above site, but focused more on using U.S. domestically available drivetrains and
suspension parts. Check out the forums, under Discussion Forums.
grassrootsmotorsports.com
A domestic magazine and forum catering mostly to autocrossing. There are some good ads in it for
hard-to-find performance parts.
pgmfi.org
People here figured out how the Honda ECUs work and how to adjust the fuel maps. Extremely
valuable if you plan to heavily modify your Honda drivetrain.
honda-tech.com
A big Honda site covering everything Honda: appearance, modifications, drag racing, autocross,
road racing, hybrid cars, turbocharging, and supercharging. A great resource while I was building the
car, when things werent quite working right!
honda-tech.com/zerothread?id=972109
Once your car is done youll probably want to drive it on track. Where to go? Who to call? I
started this thread on the Honda-tech forum, and it was popular enough that they made it a sticky. It
lists many U.S. car events at real road-race tracks (HPDE/time trial events, not real road racing).
SCCA.org
Where to start if you want to go real racing. I found the roll-cage specifications here.
metalmeet.com
Real men building everything from metalnone of this sissy composite stuff. Be sure to check out
the forum on scratch-built cars. Also a great resource on techniques and finding where to get
metalworking tools.
racingcomposites.net
A small forum focusing exclusively on race-car composite fabrication techniques and tools of the
trade.
dsr.racer.net:
If youre considering using a sportbike engine, this is a great forum, since D-sports racers already
use them and have all the technology figured out.
beesandgoats.com/boostfaq/g2icturbo.html
Planning to build your own turbocharger? A great on-line resource for what it takes. Focuses on the
Acura Integra but all the data will apply toward any engine.
dpcars.net
One man designs and builds an 800 lb., 4WD, Hayabusa-based, V8-powered super-car and spares
no expense to do it right. There is a lot of design data plus blogs on the many interesting sports cars
hes owned over the years.
kimini.com
And of course, my own site, detailing the entire build.
Appendix D: Spreadsheet
Using a spreadsheet was a big help during design. Its most useful output was the center-of-gravity
calculations and finding spring rates, but it can also find:
Total vehicle weight
Power/weight ratio
CG longitudinal location
CG height
Front-to-rear weight distribution
Front and rear unsprung weight
Front and rear natural suspension frequency
Wheel rates
Shock/spring compression
Spring free length
Required roll-bar resistance
How much the car will lean in a corner
How much the front end will dive under heavy braking
How much the car will squat under acceleration
How fast (theoretically) you can corner before turning over
Some of the equations I figured out, but most came from Milliken
14
. It didnt fit well into this book,
and even if it did, it wouldnt be helpful without the underlying equations. But since you bought this
book for design guidance (or at least entertainment!), I posted the spreadsheet on my Web site.
Supplying it this way also means you dont have to type in all the equations. The secret address is
www.kimini.com/Misc/VTECCAR.xls (note address is case-sensitive).
Another big plus (for me) is that I was always forgetting where I got parts from, so theres a
column for that. Another column can be for hard-to-remember facts such as brake caliper piston
diameters, necessary when calculating the necessary master cylinder diameters.
At the end of this section is a small portion of the spreadsheet for illustrative purposes. Please
refer to it while referencing the descriptions below.
Columns:
Appendix D: table 1
Rows:
Most are self-explanatory.
Green represents cells that need values entered and red cells display calculated results.
None of this is an excuse for not understanding the equations involved; all it does is speed up the
calculations and make it easy to try what if scenarios. I really suggest reading the Milliken book
Race Car Vehicle Dynamics, which is where many of the embedded equations are from. But as the
Disclaimer below says, I havent checked that the equations are error-free.
One feature not complete at the time of writing is the equations for calculating anti-roll bar
thickness and arm lengths.
Component: The part in question
Weight,
total:
The total weight of the part or of a sub-assembly (e.g., assembled upright)
Sub-weight: Component weights (e.g., brake caliper or brake rotor)
$: Cost (if you really want to know)
Source: Where you got the part from
X-location:
The horizontal center of gravity for the part, measured from an arbitrary datum
line
18 forward of the front axle (it makes no difference where this point is)
X-moment: The calculated moment about the datum
Y-location: The vertical center of gravity for this part, measured from ground level
Y-moment: The calculated moment about the ground reference
Big Disclaimer:
This spreadsheet is a work in progress and is not guaranteed to be errorfree, nor has it been
fully tested. Do not blindly accept the values produced; think about themDo they make sense?
Do they seem about right? Some equations, such as the overturning moments and anti-roll bar
calculations, have definitely not been tested. Use with the understanding that you need to be
vigilant about potential errors!
About the Author
Kurt Bilinski grew up in Southern California in the late 1970s, receiving a BSEE degree from San
Diego State University. Always working with his hands, he became interested in cars in his teens.
Not being able to afford a nice sports car, he decided to build one from scratch. He attributes this
ability to his parents, who built their house and raised kids who always seem to be busy building
things.
His other diverse interests include gardening, astronomy, beer-making, paintball, and Koi ponds.
Hes currently considering an ominous home remodel, and after thatanother car project.
He lives in Southern California with his wife, Kim, their faithful dog, Cooper, and a pond full of
Koi.
Kurt can be reached through his Web site: http://www.kimini.com/
1
Allan Staniforth. Competition Car Suspension: Design, Construction, Tuning , 3
rd
Ed.
(Newberry Park, CA: Haynes Publishing, 1997), 63.
2
William F. and Douglas L. Milliken. Race Car Vehicle Dynamics , 2
nd
Ed. (Warrendale, PA:
SEA International, 1995).
3
Allan Staniforth. High Speed, Low Cost , 3
rd
Ed. (London: Robert Macknay, 2003), 125127.
4
Carroll Smith. Tune to Win (Fallbrook, CA: Aero Publishing, 1978), 3839.
5
Smith, Tune to Win, 2930.
6
Allan Staniforth. Race and Rally Car Source Book: The Guide to Building or Modifying a
Competition Car, 3
rd
Ed. (Newberry Park, CA: Haynes Publishing, 1992), 63.
7
Keith Tanner. How to Build a Cheap Sports Car (St. Paul: MBI Publishing, 2005), 2022.
8
Carroll Smith. Engineer to Win (Osceola, WI: Motorbooks International, 1984), 214 217.
9
Joseph Katz. New Directions in Race Car Aerodynamics: Designing for Speed (Cambridge, MA:
Robert Bentley, 1995), 206, figure 6-39.
10
Tanner, How to Build a Cheap Sports Car, 6163. 202
11
Milliken and Milliken, Race Car Vehicle Dynamics.
12
Simon McBeath. Competition Car Downforce: A Practical Handbook, 3
rd
Ed. (Sumerset, UK:
Haynes Publishing, 2003), 2526.
13
Katz, New Directions in Race Car Aerodynamics, 216, figure 6-52.
14
Milliken and Milliken, Race Car Vehicle Dynamics.