The following document has been prepared to assist in the building and

troubleshooting of the PI4RAZ QRP PSK Transceiver.

The documentation that follows is transcribed from an article that was

presented in the April 2013 edition of RAZZies

I would like to express my thanks to Frank Waarsenburg PA3CNO for granting DX Kits permission to
produce PCBs and kits for the transceiver and thus keep the project alive.

Andy Hunter de G6LBQ

What did I start. I've asked myself that many times the past weeks! It all looked so simple, take a
proven existing design, build it yourself and present it as a project. In all innocence that
"provendesign" was the first erratic assumption.

We ordered
red the components for 3 prototypes according to the initial design of KD1JV, we had some
PCBs made which as it transpired contained numerous errors and at this stage time all alarm bells
should have gone off!

Wee started with enthousiasm building the transceiver,

transceiver, counting on a Plug & Play experience. That
turned out to be a Plug & Pray experience,
experience the receiver section worked right away but the
transmitter was a mess.

Much de bugging was done and our experiences described in the december 2012 edition of the
RAZzies publication.. From the mail correspondence with Steve KD1JV (the original designer) it
appeared that he never connected his design to an antenna. He never got passed the dummy load
on the test bench.

Moving on and after all our modifications

modification we had a perfect working rig with an output of 3W.

The PCB was finalized, with all modifications on it. The technical crew checked the PCB many, many
times to prevent errors in the final kit. The interest for the project exceeded our wildest
expectations, we expected between 10 to 15 builders but ended up with 44! 43 amateurs payed in
the end and the components were ordered from Mouser, Conrad and Kits&Parts.

The PCB design was sent to makepcb.com who offered a good price with- accordingtothesite-good
accordingtothesite
specifications. We included a conductivity test with the order to make sure the PCBs contained no
errors and ensure quality and re produceability.

We ordered parts for 44 kits, even though at the moment of ordering only 42 payments were
received. The idea behind this was that we would have sufficient spares when somebody would ruin
something. Henny PA3HK volunteered
volunteer to build one kit (number44) from scratch to verify the
manual and get some extra building tips which would make reproduction easier. That build did not
immediately generate signals of potential problems; Henny even made some QSOs with the kit.

Then, the first club meeting with the measuring equipment arrived. When powering up the first kit,
the IC of the power converter went up in smoke, though
though that circuit is not even powered when the
power of the kit is turned on. The receiver appeared to work, but the transmitter stayed ON. The
second kit had that same problem. Set number three started towork after soldering a couple of
forgotten jumper wires,
res, but the output power did not exceed a couple of hundreds of milliWatts.
And kit number 4 also stayed in TX mode. That really was a bad score. What was going on here? Jan
PD0NLRs kit was taken care of by Henny PA3HK, and Wim PD0JNGs kit temporarily moved mov to the
shack of Frank PA3CNO.
The aforementioned two builds were used for extensive research to accertain a solution for the
problems the kits were suffering from: staying in TXmode, insufficient power and a lot of"grass" on
the output signal. We examineded more kits that did not function and gained a pretty good idea what
the problems with most rigs were.
were

We noticed a lot of wiring problems with most kits. Forgotten jumper wires, but also many
incorrectly wired potentiometers and spot switches. May
May be we did not clearly explain the
connection of the potentiometers in the manual, so we will have a look at that again

Both the Audio In and the Audio Out connections of the transceiver are connected to the center
terminal of the potentiometers! The upper
upper potentiometer is the RX Level potentiometer, and the
lower one is theTXLevel potentiometer. Theconnector labeled as "toMicIN" is, looking from the
transceiver's point of view, theoutput. Thus the connector called "to PhoneOUT" is, from the
transceiver's point of view, theinput. That sounds a bit confusing, but the diagram clears it all.
To make things clear, below is a top view of the audio connector, with an overview of where you can
find the ground, ring and tip terminals. That looks as follows:

The SPOT switch was also often incorrectly wired. The reason is because the fourth PCB connection
appeared to be connected to ground, which makes it only suitable for connection to the center
terminal of the switch. Themodified instruction for connecting
connecting the spot switch looks as follows:
The wires of the fourth and fifth hole from the right have to be exchanged. Also the wires to the
main PCB have to be exchanged! Otherwise your SPOT will be permanently ON and you won't see
much on your waterfall.

Mind d that C32 is shown the wrong way on the silkscreen(thecomponent layout): this capacitor has to
be mounted the other way around with it's plus and minus terminals. It won't explode if you mount
it the wrong way, but it should be mounted the other way around.
around. These were the real important
flaws of the building instruction manual. We did not receive any other comments.

The most important moment when you have finished building your kit is testing whether everything
works. When writing the manual we presumed that not everybody has an oscilloscope. Though
adjusting the kit without an oscilloscope is possible, it really helps if you have one. Then you see
what really happens and deduct what to do from what you see. The following troubleshooting
instructions assume
ume the availability of such device. Don't you own one, ask an amateur at your local
radio club for help.

We have not found any problems in the receiver so far. They all worked the first time, without
exception. Most problems were with the transmitters. The
The following tips may help you to find the
problems what may occur in your kit.

The Kit Stays In TX mode

That means that the Mute line is pulled low for some reason. The most common cause is a defective
FETQ1, at theinput of thereceiver. When the source-gate has a lowimpedance, it pulls the Mute-line
low which causes the transmitting circuit to be permanently powered. Other possible
ossible causes are a
defective Q3 or a defective Q5, in order of likelyness.

Spot remains ON
Usually due to a wrong wired SPOT switch,
switch, see also the previously written instructions about wiring
the SPOT switch. Becareful: If your power LED turns RED (if you ordered the Battery Monitor Option)
when you switch the SPOT to ON, something is incorrectly wired and you have a short circuit!
Immediately
mediately turn off the power and check the SPOT switch wiring.

NO Output
If the power does not increase when increasing theTX-level,
theTX level, then you really need an oscilloscope. Of
course you need a PC, tablet or smartphone connected to the kit, equipped with a PSK
P program for
generating the well known PSK31 audio signals. That signal should look as shown on the picture on
the next page:
This is a nice DSB-signal.
signal. This is what you will find at pin2 of the LM324, adjustable with the TX-level
TX
potentiometer. After that, it will pass through the AM-demodulator
AM demodulator U4a, causing the following signal
at pin1:

After U4a we reach the input of the low pass filter, which has to remove any remains of the carrier.
That is the connection point of R28, D14, R19 and C43. The input
input of the low pass filter looks as
follows:

After filtering out the carrier only the envelope of the original PSK31-signal
PSK31 signal remains. You can find
that signal at the output ofU4b, pin7 of the LM324:
From here the signal splits in two directions: one direction
direction is the clamp circuit consisting of C59, D13
and R14 which lead to the squarer U4c, and the second direction is to U4d which in turn applies
signal to powerfet Q10. The clamp circuit together with U4c generates a pulse close to the lowest
level of thee envelope signal, triggering flip-flop
flip U5a. This flip-flop
flop takes care of the 14MHz phase
shift of the signal that passes the 74HC86. The squarer output looks like this:

This is the outout of U4c, pin8 of the LM324. If you look at pin1 of the 4013, then you should see the
flip-flop
flop trigger at the rising edge of the squarer signal. That looks as follows:

If every thing is OK so far, then most of the driving ciruits are OK. Adjusting transformers T2 and T3
is very critical. If they are misaligned you will
will only see rubbish at the output of the final amplifier, if
you see anyoutput at all. The best place to measure the alignment of those two transformers,
connect the oscilloscope to the junction of the collector of Q6, R11 and L4. Adjust T2 and T3 for
maximum
imum output. That should be at least 7Vtt. If you don't get at least 7Vtt, you have insufficient
drive. There is a cure for that: Increase C6 from 2p2 to10pF (or put 10pF in parallel, preferably at the
bottom side of thePCB, that is not critical). Now you should have between 7-10Vtt.
10Vtt. If you still have
less, and thelower part of the signal does not reach zero but floats about 2-4V
2 4V beyond ground level,
then most likely your 74HC86 has died. Check at first the voltage at pin14: that should be between
4.5 and 5V. If not, either Q2 does not work or because of a defect, the load of the 74HC86 on the TX
voltage is too heavy. If your 74HC86 is malfunctioning, you may also see the output signal on pins 6
and 8 float about 2-4Volt
4Volt above ground. How to find faulty BS170s
BS170s we will describe later on.
74HC86's don't die easily, but we have nevertheless seen this twicenow. Are all signals OK, than you
should see the following signal at the drains of the BS170s:

The easiest place to measure that is at the cathode of zener

zener diode D10 which is right beside the
BS170s. It is very important that the signal shows an almost straight bottom line. If not, you loose
alot ofpower, and the FETs turn hot. The signal might look as follows:

This is not good. The most common cause is one or more defective BS170s, which causes the output
to decrease. Measure the signal at pin 6 or 8 of the 74HC86(U3). That should be close to 5Vtt. If that
is less (just over4Vtt) then you have a faulty FET. If it is MUCH less (ca.2Vtt) then you won't see
output at all, and most likely the 74HC86 has died. The only way to find which BS170 is at fault, is to
de-solder them one by one and check whether the output of the 74HC86 or the final output
increases. This is the most common problem with the PA. Under normal conditions you should be
able to obtain 30Vtt at theoutput (2,25W). That can be improved even more, as we will see later on.
What we found, was that the transceiver on both the oscilloscope and the spectrum analyzer
showed a restless picture. A lot ofrubbish. We really puzzled for days what could cause that. The
only big difference between the prototype and the final kit was actually thePCB. The prototype PCB
was made at theTwente University, and the kit PCBs were made in China. As far as we can see it,
there is a difference. The PCB tracks are thin, which causes probably apretty high inductance and/or
resistance. In an attempt to improve the situation, we soldered some extra wires across the tracks,
to give especially the powerlines somemore"body". And it did make a difference. Not in the output
power, but spurious goes down and the signal looks a lot quieter. The next picture shows which
extraconnections were made with solid wire, to decrease resistance. Note that we alsoals made some
extra ground wires, to shorten the path the current has to take when going from one component to
another. This mod made theoutput signal look a lot better on the oscilloscope and spectrum
analyzer.
Experience showed us that there are some more improvements to make. Besides the extra wires to
reduce PCB inductance and resistance, and increasing C6 for a cleaner drive, we found some other
issues which mayy improve the performance a bit. These modifications are not mandatory: the kit will
work fine without them. They just make things nicer!

Extra diode in series with the TX LED

One bad habit of the kit is that after using the SPOT switch, the TX LED remains faintly
fain lit. That is
because after using the SPOT switch, the output of the flip-flop
flip flop remains high. That voltage feeds
theTX rail through the protection diodes of the 74HC86, which causes theTX rail to stay at just 2V.
The Forward voltage of a red LED is about 1.6V, so it will light up faintly. You can put a 1N4148 diode
(a normal silicon diode) in series with Rvs on the front pcb (cathode facing the LED). That adds 0.7V
to the forward voltage of the LED so it will stay off at RX.

More power
It is possible to increase thepower of the kit to at least 3W by modifying the final amplifier. De-
De
solder the cathode of D12 (not the side of the dash) from the PCB. Solder a 100nF capacitor with one
lead in the now freed hole, and solder the other lead to the free lead of the diode. From this
junction, solder a 15k resistor to the +5V (i.e. the emitter of Q2, connected to R2). Increase R22 from
1k to 10k. If your FETs are OK, you will measure 1,5V at that point. If not, that is a clear indication
something is wrong with your BS170s.
Shortening the hang-time
The hang-time after transmit was considerably longer than with the prototype. To shorten the time,
you may put a 82k (or100k) resistor in parallel with R9, or replace R9 with a 68k resistor. That makes
the hang-time less irritating long.

Moving C62
Some times the transmitter does not like the shielded wire to the Watt meter and responds with
spurios. You can improve that by removing C62 from the front pcb and solder it directly to the
antenna connector, in series with the wire to the Wattmeter. Put a wire link at the place of C62 on
the front pcb.

Insufficient Audio Drive

In some cases the audio is not enough to drive the transmitter to its full power. What we found is
that somePSK programs do not use both audio channels. In the design, we assumed that both
channels areused. The result is that the two 47 Ohm resistors that combine both audio channels, will
function as an attenuator. You can check that with a headphone or PCspeaker set. If only one
channel is used, remove the resistor off the unused channel. That makes a factor 2 differencen drive.
If that is still not enough, or if you do have a stereo signal, you may increase R28 in the input circuit
from 10k to 22k. That will double the gain of the Op-amp and should give you enough signal to drive
the transmitter to full power.

IMPORTANT
If for any reason you need to remove a component from the PCB where possible snip the
component leads prior to desoldering, the DX Kits PCB has hole plating and it is generaly much easier
to remove parts by snipping the body away and removing the component legs one at a time.