MULTIPLE INPUT, SINGLE OUTPUT

DC-DC CONVERSION STAGE

FOR DC HOUSE

A Thesis

presented to

The Faculty of California Polytechnic State University,

San Luis Obispo

In Partial Fulfillment

of the Requirements for the Degree

Masters of Science in Electrical Engineering

By

Jason William Baltiérrez

June 2019
 © 2019

Jason William Baltiérrez

ALL RIGHTS RESERVED

ii
 COMMITTEE MEMBERSHIP

TITLE: Multiple Input, Single Output DC-DC

Conversion Stage for DC House

AUTHOR: Jason William Baltiérrez

DATE SUBMITTED: June 2019

COMMITTEE CHAIR: Dr. Taufik,

Electrical Engineering Department

COMMITTEE MEMBER: Dr. Majid Poshtan

Electrical Engineering Department

COMMITTEE MEMBER: Prof. Ali Dehghan Banadaki

Electrical Engineering Department

iii
 ABSTRACT

Multiple Input, Single Output DC-DC Conversion Stage for DC House

Jason William Baltiérrez

In this thesis project, a proposed architecture for the multiple input, single

output conversion stage for the DC House was designed, simulated, and tested.

This architecture allows for multiple different input sources to be used to create a

single higher power output source. The design uses a DC-DC boost converter with

a parallelable output which has been demonstrated to allow increased total output

power as a function of the number of input sources available. The parallelable

output has been shown to distribute load amongst the input sources relatively

closely to optimize the system. This approach is also desirable since it allows for

flexibility in multiple configurations it can be used in. The design was tested using

hardware and data results show the performance met and exceeded the needs of

the DC House project. Data was taken for configuration with 1, 2, 3, and 4 input

sources providing greater than 600W of total output power at an efficiency of

greater than 92%. This architecture demonstrates the possibility of expanding the

total available power for a single output in proportion to the number of available

input sources.

iv
 TABLE OF CONTENTS

Page

LIST OF TABLES…………………………………………………………………..…..vi
LIST OF FIGURES……………………………………………………………….……vii
CHAPTER

1. INTRODUCTION ........................................................................................... 1
2. BACKGROUND ............................................................................................. 4
3. REQUIREMENTS .......................................................................................... 8
3.1 Input Requirements ............................................................................ 9
3.2 Output Requirements ......................................................................... 9
4. DESIGN ....................................................................................................... 13
4.1 Design and Considerations. ............................................................. 13
4.2 Calculations, Simulation, and Design ............................................... 19
4.3 Circuit Board Layout ......................................................................... 23
5. HARDWARE TEST AND RESULTS ............................................................ 26
5.1 Initial Testing, Challenges, and Modification .................................... 26
5.2 Results ............................................................................................. 33
5.2.1 Single Power Supply Performance....................................... 33
5.2.2 Parallel Power Supply Testing ............................................. 38
6. CONCLUSION............................................................................................. 49
BIBLIOGRAPHY ................................................................................................. 52
APPENDICES
A. Simulation Schematic......................................................................... 53
B. Layout Schematic............................................................................... 54
C. List of materials .................................................................................. 55
D. Top Side of Layout ............................................................................. 58
E. Bottom Side of Layout ........................................................................ 59

v
 LIST OF TABLES

Table Page

3-1: Summary of MISO Conversion Stage Requirements .................................. 12

4-1: Summary of Design Equations .................................................................... 21
5-1 Power Supply Regulation ............................................................................. 34
5-2: List of Test Equipment Used ....................................................................... 38

vi
 LIST OF FIGURES

Figure Page

3-1: Block System Block Diagram. ....................................................................... 8

4-1: Feedback Connected to Output after Or-ing Diode ..................................... 14
4-2: Feedback Before Or-ing Diode .................................................................... 15
4-3 Voltage Droop Demonstration Circuit ........................................................... 16
4-4: Voltage Droop Simulation ............................................................................ 17
4-5: Boost Converter Diagram ............................................................................ 19
4-6: LTC3814 Boost Controller Functional Diagrams [10] .................................. 20
4-7: 4 Parallel Converters Simulation Traces. .................................................... 23
4-8: Signal Ground and Power Ground .............................................................. 25
5-1: Configuration for Generating Bias ............................................................... 27
5-2 Configuration for Using an External Bias ...................................................... 27
5-3 Locations of Cut Traces for Modification ...................................................... 28
5-4: Modifications to traces on PCB ................................................................... 29
5-5: Completed Modifications for Applying External Bias Voltage ...................... 30
5-6: Completed Converter with Modifications ..................................................... 30
5-7: First Power on Observed Input (Ch. 1) and Output (Ch. 2) ......................... 31
5-8: First Successful Turn on into Regulation. .................................................... 32
5-9: Power Supply 1 Ripple at Full Load ............................................................ 35
5-10: Power Supply 1 Efficiency ......................................................................... 35
5-11: Power Supply 2 Efficiency ......................................................................... 36
5-12: Power Supply 3 Efficiency ......................................................................... 36
5-13: Power Supply 4 Efficiency ......................................................................... 37
5-14: Power Supply 1 and 2 Load Share Difference........................................... 39
5-15: Load Difference PS 1 & 2 .......................................................................... 40
5-16: Load Difference PS 2 & 3 .......................................................................... 41
5-17: Load Difference PS 2 & 4 .......................................................................... 41
5-18: Worst Case Load Sharing for PS 2, 3, & 4 ................................................ 42
5-19: Load current of PS 2, 3, & 4 operating in Parallel ...................................... 43
5-20: Custom Load Cables ................................................................................. 44
5-21: Block Diagram of Four Paralleled Converters ........................................... 45
5-22: Four Converters Paralleled Testing Setup................................................. 45
5-23: Four Paralleled Converters Shared Load Current ..................................... 47
5-24: Worst Case Load Difference of 4 Paralleled Converters. .......................... 47
5-25 Ripple at 12.8A Total Load Current ............................................................ 48

vii
1. Introduction

It is part of human nature to be curious and learn. This is what sets us apart

from the rest of the living organism on Earth, our ability to learn and comprehend.

Due to this ability to learn and comprehend we have become the dominant species

on earth. Humans use the power of the mind (some more than others) to advance

the civilization we have created to new heights. Humans have been able to create

tools and machines to create technologies which improve our ability to advance

our society. This understanding is creating growth at an exponential rate and is

shown through the development of hand tools, animal assisted farming equipment,

water mills, wind to drive sailboats, steam power, fossil fuels, and electrical power.

In the current era of technological advancement, the forefront driving new

technologies is electricity and electronics which drives computing power [1].

The understanding of electricity to potentially do work for humankind began

with the understanding and characterization of its properties. In the early 19th

century, Michael Faraday contributed to this characterization and was able to

demonstrate a simple induction motor. In the late 19th century, contributions from

Nikola Tesla and Thomas Edison helped to bring electrical power to the public by

developing the first commercially available electrical machines. This was the

beginning of electrical power conversion.

When the power of electricity first became available, there was no

infrastructure to support it. In the United States, Nikola Tesla was emphasizing the

use of AC voltage as the primary source of energy transmission. Westinghouse

Electric Corporation licensed Nikola Tesla’s AC machine designs to setup the first

1
long distance transmission lines using high voltage AC between Buffalo and New

York [2]. On the other hand, Thomas Edison was promoting his designs which

were based on DC voltage distribution. Edison was able to setup DC distribution

in New York, but there were limitations to what his machines were able to provide

power to. Edison’s distribution system initially generated 110V DC but quickly

changed to 220V DC so that the line losses could be reduced and cost of copper

for distribution could be minimized. Although this was implemented, the

serviceable area for electricity distribution was limited to about a single square mile

as demonstrated in the Pearl Street Station in New York. Anything requiring higher

loads was limited to being in close proximity of the DC distribution center.

Inevitably, the AC transmission method gained increased popularity and currently

is the main form of distribution across the world [3].

Although DC power was not used for mainstream distribution, DC

technology was still advancing in the early 20th century with the development of

various devices such as the metal tank rectifier, grid-controlled vacuum-tube

rectifier, ignitron, phanotron, and thyratron. These devices are relatively large and

it wasn’t until the mid 20th century that a new type of technology was created which

revolutionized the way DC power could be converted. That technology was the

creation of silicon semiconductors and this is the basis for DC electronics as we

know today.

With the development of silicon semiconductor devices came many benefits

to DC power conversion; the size of DC converters reduced, weight of devices

decreased, and the efficiency of conversion increased. Also, new forms of control

2
pushed the capabilities of the converters to greater performance. Once these

devices became commercially available, they soon made their way into consumer

electronics. Power electronics has come to be a mature technology in the last 40

years and can now be found in nearly all machines [4].

The widespread acceptance and integration of power electronics in our daily

lives is apparent. Nearly everything needs a charger or needs to be plugged in.

With such a high global population, more and more energy is required to keep

society moving. Sustainable energy is imperative to ensure earth resources are

not depleted. This understanding has led to the steady expansion of the use of

clean energy across the United States and the world. With this growth there is the

opportunity for technologies that support it to advance and grow. In areas where

there is not yet any infrastructure to support mass distribution there is the

opportunity to develop a new standard which the developed world has not seen.

3
2. Background

The United States is a society in which energy is readily available. Energy

availability is a necessity to keep this society going and any interruptions of this

power could cause issues which is why there is so much investment into the

infrastructure. Although this is the case for the United States, there are other parts

of the world in which reliable energy is not available. These underdeveloped parts

of the world need energy if they are to become integrated into the more advanced

societies otherwise the socioeconomic gap will be increased further. The biggest

roadblock there is to introduce widespread energy to all is the infrastructure to

support it. This includes the equipment for generating energy and also the

infrastructure for distribution. There are extremely large startup costs to layout the

foundation for delivering energy, especially in rural areas. These costs are

something an underdeveloped country cannot support without incurring massive

debt. To quickly equip these underdeveloped countries with energy, the startup

costs need to be greatly reduced. If the energy is only generated local to the area

of need, distribution is greatly reduced.

According to the article “Sustainable energy for developing countries”, there

are approximately 1.6 billion people across the world who do not have access to

basic electrical services [5]. The rural poor in developing countries make up the

main population of those without access to electricity. The biggest obstacle for

those countries is the financial burden of trying to implement generation and

distribution infrastructure. Minimal financial resources from the government also

competes with other societal needs such as healthcare. There are minimal needs

4
required for energy that could greatly improve the quality of life such as the ability

to read at night, pump required drinking water, and listen to radio broadcasts [5].

Focusing on the rural areas for energy distribution, the biggest obstacle to

energy availability is the costs [5]. Off-grid applications make the most sense for

rural areas since energy generation and distribution can be dispersed local to the

area of need without incurring additional costs for long distance distribution.

Renewable energy technologies have greatly improved and can provide energy at

a lower cost than other options when compared to connecting to the main grid. To

combat the fact that renewable energy generation can vary throughout the day

depending on the available sources, multiple sources could be used to increase

the chance of consistent generation. Furthermore, DC generation is common in

renewable energy generation. Maintaining the energy as a DC source instead of

creating an AC source will be more efficient and less complex which will reduce

costs and increase reliability. The proposed solution to the worldwide issue is to

use multiple renewable energy sources to generate DC electricity for local

distribution in rural areas of the world at a minimal cost. This is the basis for the

DC House project.

The DC House project will use multiple forms of renewable energy including

solar power, hydroelectric power, wind power, and human powered generation [6].

Each of these sources will contribute to the power for the DC house. The purpose

of this thesis is to create a method of combining those energy sources for a single

distribution source to the house.

5
 Prior to beginning this project other students have proposed solutions for

this main converter, a Multiple-Input Single-Output (MISO) DC-DC converter.

Those students proposed using flux-additivity in a multi-winding transformer to

combine the energy of the various sources. For these projects, a Flyback converter

[7] was used in one case and a Full-bridge converter [8] was used in another case.

Under both projects, there were technical issues encountered and the final design

was limited in performance mainly in the ability to deliver the desired output power.

For this project, the proposed solution will use an alternate method of

conversion than was previously used. The goal is to reduce the risk of creating a

converter that only works under certain circumstances. There may be different

configurations that this design will be used in so there should be flexibility in the

design to accommodate for those different configurations. The end goal is to also

minimize costs as much as possible so reducing unnecessary complexities will aid

in this effort. This will include minimizing the use of custom components by using

off the shelf components readily available to the public.

This project will be to design a single converter which can be paralleled with

other copies of the same design. By doing so, it reduces the number of unique

components which are used which will reduce the costs. The expectation is that a

different number of sources can be used and each one will have its own converter.

For any number of converters that are used, they will all have the capability to be

paralleled at the output to increase the overall output power capability. Since the

design will support energy generation in rural areas which are underdeveloped,

the design should be robust and require minimal maintenance. Further details of

6
the requirements will be covered in Chapter 3 and the details of the design

approach will be covered in Chapter 4. Results of hardware test of the proposed

converter will be presented in Chapter 5. Lastly, Chapter 6 summarizes the thesis

and discusses further improvements of the proposed converter.

7
3. Requirements

As stated in Chapter 2, the basis of the project is to take multiple input

sources of renewable energy generation and combine them into a single source

for the DC House, Multiple Input Single Output (MISO) conversion stage. Each of

the input sources should be capable of providing the same amount of maximum

power. When all sources are available, the single load (source for the DC House)

shall be capable of delivering the max rated power of the combined sources. The

output should be a clean source of power providing minimal noise so as to not

create issues for the loads. Figure 3-1 shows a block diagram of the system with

the item highlighted as the main focus of this design, a conversion stage taking in

multiple inputs and producing a single output utilizing a single design, which allows

for growth by being parallelable to provide additional power.

Figure 3-1: Block System Block Diagram.

8
 3.1. Input Requirements

The input source will be generated from one or multiple sources up to 4

sources. Note, these sources are generated using various types of sustainable

resources which include harvesting solar energy using photovoltaic solar panels,

energy from naturally flowing water using a turbine, harvesting naturally flowing

winds using a wind turbine, and a human powered generator. These various forms

of generation will be conditioned and will become the sources which provide the

input power for the MISO conversion stage.

Each input will have the same requirements. The nominal input voltage of

a single source will be 24 Volts. Understanding that there is tolerance on the input

voltage regulation of the sources, the parallelable converters should be capable of

operating when the input is 24 Volts +/- 2 Volts. It is also understood that each of

the input sources should be capable of supplying a maximum of 200 Watts. This

will need to be considered when determining the efficiency of the parallelable

converter. When considering worst case conditions, the maximum input voltage is

26 Volts which will be used when considering the voltage rating required for the

selected components. For maximum current considerations, one must take into

account the minimum operational input voltage which is 22 Volts. From this the

maximum average input current is calculated which is about 9.1 Amps.

3.2. Output Requirements

The final purpose of the load is to provide a single power source for the DC

House. The intended output voltage for this project is a single 48Volt DC output

which is the input for the various devices that will plug into the DC House. So that

9
downstream loads can be optimized and sized correctly, the source should have

some level of output regulation tolerance. The combined line and load regulations

of the output should be +/-3% of nominal output voltage. To achieve higher

tolerance on output regulation, different circuitry and more precise parts would

need to be added. This would increase cost and complexity which is undesirable

for the end use of this project. This results in an output voltage range of

approximately 46.5 Volts to 49.5 Volts.

To maintain a clean source for the DC house, ripple needs to be minimized

to not cause issues with the end load. The goal is to achieve a maximum peak to

peak ripple of 2% of the nominal output voltage. This should be measured when

the output is fully loaded and the input is at its nominal of 24 Volts. This indicates

a peak to peak ripple of approximately 1 Volt. Voltage ripple on the output of the

converter is important for the load it is feeding, depending on the input

requirements of the load, various issues can occur if the ripple is too large.

Excessive ripple can cause undesirable AC losses, a source of noise which could

couple into other circuitry, and excess voltage stress on components.

As previously stated, the maximum input power supplied by each of the

sources is 200 Watts. The goal is to be able to provide a maximum power of 600

Watts total when all input sources are available. As a side note, this would indicate

that the maximum output current is 12.5 Amps total which should be considered

when sizing components. Ultimately, the actual maximum output current would be

dependent on the output loading and the efficiency of the converters. If there are

a maximum of 4 sources to be used, this would indicate that there is a maximum

10
input power of 800 Watts. From this the efficiency of the system can be calculated.

With 600 Watts as the maximum output power and 800 Watts as the maximum

input power, efficiency of the system should be 75% at a minimum. As a stretch

goal, the system should strive to achieve higher efficiency without incurring

significant costs for higher performance components. Higher efficiency improves

the performance in multiple ways. Increased efficiency means less wasted power

which also means less losses. Less losses means components do not have to

dissipate that power so they don’t heat up as much, which would stress out the

components more. Reduced stress on components increases the reliability of the

system so that there are reduced number of failures. In all, this increased

robustness essentially produces a system with less failures, which means

decreased costs over the lifetime of the system.

Another performance characteristic that should be considered is that all

sources might not be available at any given time. So as the number of sources

decreases, the maximum output power will also decrease. Since each input is

capable of providing 200 Watts and the minimum efficiency is 75%, the total

maximum output power capable is limited to 150 Watts multiplied by the number

of input sources available, up to four sources. The variability of configurations

brings up another point; all the sources are not guaranteed to be available at all

times while the system is running. So, any combination of sources could be

available and the system should be able to operate under any of those conditions.

In other words, for any combination of available sources shown in Figure 3-1, the

system should operate at scale to the number of sources available.

11
 Table 3-1 summarizes the preceding requirements for the multiple-input

single-output DC-DC conversion stage consisting of parallelable converters.

Table 3-1: Summary of MISO Conversion Stage Requirements

Requirements Value

Nominal Input Voltage 24 V

Input Voltage Tolerance +/- 2 V

Max Number of Input Sources 4

Nominal Output Voltage 48 V

Regulation Tolerance
+/-3%
(Line and Load)

Maximum Output Voltage Ripple 2%

Maximum Output Power 600 W

Minimum Efficiency at Full Load >75%

12
4. Design

4.1. Design and Considerations

The fundamental topology for this design is a boost converter since the

voltage is being stepped up from 24 V to 48V. The use of a boost converter

eliminates the use of a custom transformer and has the potential to have high

efficiency. For current sharing, the converters will equip Or-ing diodes to protect

one output from damaging the other output when not powered up. Also, each

source would have an independent converter which would allow the system

approach to have n+1 redundancy and expandable. Not only could it

accommodate different sources of energy production like wind, solar, hydro, and

human power, but if there was an abundance of one resource, there could be

multiple instances of that source added to the system. The additional sources

could potentially increase the total power capabilities of this conversion stage.

Removal of one source should not affect the operation of the others aside from the

power capabilities

For this approach, Or-ing diodes will allow the outputs to be tied together

and prevent damage caused by back-feeding voltage. One of the limitations of

using this approach is the fact that to current share, the output voltages need to be

well regulated with a high level of accuracy. I will be using high precision parts to

feedback the voltage, but I will also be using an adjustable divider to fine tune the

amount of current sharing while the converters are operating in parallel [9].

Since individual converters are being used, each converter has its own

sense voltage. The design will allow each source to be independent of one

13
another. One case in which this is useful is if one source is available but then

another source becomes available while the first source is already regulating. As

a real-world example, in the early morning the hydroelectric source could be

available and providing power. The sun then rises, and then the solar power source

is also available to provide power. The idea is that the second source can come

on without interrupting power being provided by the first source. If the sources were

both sensing at the point of load, the second source would have trouble bringing

up its own output since it would already sense the first sources output voltage

which is already in regulation. A simplified schematic of this issue is shown in

Figure 4-1.

Figure 4-1: Feedback Connected to Output after Or-ing Diode

14
 To overcome this obstacle, the sensing is performed prior to the Or-ing

diode. This way the first converter can be in regulation and the Or-ing diode will

block the second converter from sensing the output voltage that is already up. A

simplified schematic of this updated configuration is shown in Figure 4-2. It is also

important to point out that when fine tuning the output voltage, it will have to be

performed under the load range to compensate as best as possible for the voltage

drop associated with the Or-ing diode.

Figure 4-2: Feedback Before Or-ing Diode

For the configuration discussed in the previous paragraph, since the second

converter is not providing any current through the load, no current is flowing

through the Or-ing diode of that converter. Since current is flowing through the first
15
converters Or-ing diode there is a voltage drop associated with it. Since the forward

voltage drop across the diode is proportional to the current flowing through the

diode there is no voltage drop through the second converter since it is reverse

biased as it starts to come up. As the voltage approaches regulation and the

voltage on the anode is higher than the cathode, current flow will begin and

increase through the diode thus increasing the voltage drop. Note that there are

also other voltage drops associated with the various connections and the

transmission line. So as the second converter begins to provide current, it reduces

the stress on the first converter and theoretically the first converter will provide less

current/power. See Figure 4-3 and Figure 4-4 which demonstrates the voltage

droop method for current sharing as two sources are applied at different times.

Figure 4-3 Voltage Droop Demonstration Circuit

Figure 4-3 is the voltage droop simulation circuit which simulates 2 closely

regulating voltage sources coming on at different times. Figure 4-4 shows the

traces of each input source, current through each diode, and the total output

voltage and current. The traces show that the first source (V1) turns on and
16
provides full current to the load R1. Some time later, the second source (V2) turns

on and begins to provide current until the sources reach regulation. Once both

sources are in regulation, the sources share current as expected.

Figure 4-4: Voltage Droop Simulation

This approach of using the parasitic and, potentially intentional, line losses

is called the voltage droop method for current sharing. Normally line drops are

considered undesirable in a power converter since they contribute to losses. In this

case they are beneficial to a certain extent because this inherently provides current

sharing [9].

17
 To increase the converters robustness and prevent it from operating in

conditions it was not intended for, some additional circuitry was added. There is

an under-voltage lockout circuit that prevents the converter from switching until the

allowable input voltage is reached. This is to prevent any peak currents from

flowing when the input voltage is low. There is also a circuit that monitors the bulk

capacitors near the output to ensure the voltage has been charged to the good

input voltage range. Once both parameters have been fulfilled, the circuitry allows

the converter to begin switching by releasing the Run/Soft Start pin. In both circuits,

hysteresis was also added to prevent the circuit from getting stuck in a ‘hiccup’

mode. This is when the circuit is enabled and begins to draw large currents which

causes a voltage drop at the sensed voltage which falls below the minimum enable

voltage which shuts off the converter. The input then rises again since there is no

current draw and passes the enable threshold which starts the process over again.

Hysteresis adjusts the enable threshold to a lower threshold once the upper

threshold is passed. This allows the input to fall slightly then continue to rise as

long as it doesn’t reach the lower threshold.

Since the point of regulation is prior to an Or-ing diode, there is increased

variability of the regulation voltage after the Or-ing diode due to variances in the

voltage drop of that diode. This applies to both variability from component-to-

component and over the load range since the voltage drop across a diode changes

as a function of load. In order to compensate for the amount of variability there can

be of the regulation voltage, a potentiometer was used in the output voltage

18
setpoint resistor divider. This allows the output voltage to be fine-tuned to allow the

most balanced current sharing between converters.

4.2. Calculations, Simulation, and Design

The boost converter is the fundamental topology used for this project. A

simplified diagram is shown in Figure 4-5.

Figure 4-5: Boost Converter Diagram

When designing a boost converter, the critical components of the design

are the inductor, switches, and the controller. Normally, a diode would be used in

place of M1, shown in Figure 4-5, but in this case using a MOSFET decreases the

losses and increases the efficiency.

To effectively control the timing of both switches, a controller was chosen

which specifically has this functionality. If a standard PWM controller was used and

the single driver was used to drive the MOSFET’s inversely of each other, you

could potentially short the output if both switches were conducting at the same

19
time. This is due to the fact that MOSFET’s have parasitic elements and there are

limitations to gate drivers which do not allow the MOSFET’s to turn on and off

instantaneously. Therefore, there needs to be a small amount of dead time (delay)

to allow one switch to turn off and the other to turn on. The controller that was

chosen to regulate this converter is the LTC3814. A functional block diagram of

this controller is shown in Figure 4-6 [10].

Figure 4-6: LTC3814 Boost Controller Functional Diagrams [10]

20
 The MOSFET which replaces the diode is known as a synchronous

MOSFET. Since the synchronous MOSFET could be conducting through the body

diode during the dead time before it is commanded to turn on, the LTC3814

datasheet recommends using a diode in parallel with the MOSFET. Current will

flow through this alternate diode instead of the body diode only if the diode is

chosen correctly by having a lower forward voltage drop. This provides the benefit

of lower losses during that transition and also distributes losses amongst other

components which reduces stress on those components.

When designing the Boost converter, worst case conditions are considered.

The highest currents occur when the input voltage is at a minimum. For this design,

the minimum operational input voltage used in the design equations was 20V

which provides additional margin since the required minimum operational input

voltage is 22V. Table 4-1 shows a summary of the design equations used for the

fundamental power components. These equations were placed in an Excel sheet

to allow easy manipulation of variables. These equations were taken from both the

LTC3814 [10] datasheet and book “Power Electronics” by Muhammad H. Rashid

[4]. There were also component values specific to the LTC3814 which were

considered for choosing components. This included items like Soft Start

capacitance for controlling the ramp on rate of the output which.

Table 4-1: Summary of Design Equations

Variable Equation Value

Duty Cycle (𝑉𝑖𝑛𝑚𝑖𝑛𝑖𝑚𝑢𝑚 ∗ 𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑐𝑦)
1− .6875
(Dmax) 𝑉𝑜𝑢𝑡
Inductor Average Current 𝐼𝑜𝑢𝑡
10 A
(Iavg) (1 − 𝐷_𝑚𝑎𝑥)
21
 Inductor Current Ripple 𝐼𝑜𝑢𝑡
𝑃𝑒𝑟𝑐𝑒𝑛𝑡𝑟𝑖𝑝𝑝𝑙𝑒 ∗ 2A
(Iripple) (1 − 𝐷𝑚𝑎𝑥 )
(𝑉𝑖𝑛𝑚𝑖𝑛𝑖𝑚𝑢𝑚 ∗ (𝑉𝑜𝑢𝑡 − 𝑉𝑖𝑛𝑚𝑖𝑛𝑖𝑚𝑢𝑚))
Minimum Inductance (Lmin) 18.7µH
(𝐼𝑅𝑖𝑝𝑝𝑙𝑒 ∗ 𝑓𝑠𝑤𝑖𝑡𝑐ℎ𝑖𝑛𝑔 ∗ 𝑉𝑜𝑢𝑡)

Inductor Peak Current 𝐼𝑅𝑖𝑝𝑝𝑙𝑒 𝐼𝑜𝑢𝑡

+ 11A
(Ipeak) 2 (1 − 𝐷𝑚𝑎𝑥 )

Once the fundamental component values were chosen, the circuit was

simulated using LTspice®. The simulation program has a model for the controller

being used so the simulations greatly reduce the risk of the design. Initially, a single

converter was simulated to ensure the fundamental design was operating as

planned. The circuit was simulated under various conditions including various

loading conditions and input voltage conditions. Once the fundamental design was

operating as planned, additional copies of the circuit were inserted, and the design

was simulated while paralleling converters. The simulated schematic is shown in

Appendix A. Figure 4-7 shows various traces of the simulation. For this simulation,

the input sources were set to different levels and also turned on at different

times/ramp on rates. The output voltage traces, both shared and individual

voltages before the Or-ing diode, were monitored to validate all converters were in

operation. Each of the converters output voltage setpoints were varied slightly and

the individual output currents were monitored. Once all the individual converters

were in regulation, the load current was increased up to full load of 12.8A. It can

be seen that the individual converters share load current during the simulation.

22
 Figure 4-7: 4 Parallel Converters Simulation Traces.

4.3. Circuit Board Layout

As previously stated in Section 4.2, the design was initially simulated using

LTspice® to gain confidence in the fundamental performance of the design. Real

components found mostly on https://www.digikey.com/, were put in a list linked to

the reference designators on the simulation schematic and would eventually be

purchased. Prior to purchasing, the simulation components were modelled as

closely as possible to the actual components that were to be used. This was to

verify that the component parasitic elements would not make a significant

difference in the performance of the converters. This schematic was then imported

into the printed circuit board (PCB) creation software, EAGLE. The final schematic
23
used in layout is shown in Appendix B. Some components did not use standard

packages from the EAGLE component library, so those component pad and layout

information had to be created manually based on the datasheet. A final list of

material purchased is shown in Appendix C.

The layout of the printed circuit board was done in the free version of

EAGLE. The overall flow of the power supply was generally laid out to support the

direct current flow and minimize impedance to create an efficient path. The PCB is

a two-sided PCB which allowed placement of components on both sides of the

PCB. Two sides allowed for components to be tightly coupled such as placing

capacitors near IC’s for high frequency decoupling. Care was taken to reduce

noise from coupling onto control signals which could be noise sensitive. Noise is

caused by high dv/dt and di/dt components and traces. It is coupled onto these

signals through parasitic paths created within the layout. In order to avoid coupling

noise onto these sensitive signals and components, parasitic paths were reduced

as much as possible. One method is to use an analog-signal ground to isolate

noisy signals from noise sensitive signals. This is a plane that sits directly

underneath the noise sensitive components. See Figure 4-8 for this portion of the

PCB layout. There is only one location where this ground is connected to the main

power return. This is to keep noisy return currents from passing through this portion

of the ground plane [10]. The complete layout is shown in Appendix D and E. The

PCB’s were purchased from the company Oshpark, (https://oshpark.com/).

24
Figure 4-8: Signal Ground and Power Ground

25
5. Hardware Test and Results

5.1. Initial Testing, Challenges, and Modification

When initially testing, during first power on, it is important to be on the

cautious side and assume that something could be incorrect. It could be something

as simple as an incorrect connection, configuration of instrumentation like power

supply set for abnormal source voltage, incorrect component, or many other

potential hazards when it comes to developing a new design. Prior to applying any

power, general PCB and component checkout should be performed. Continuity

checks should also be performed in order to verify proper connections are made

and there are no errors from the PCB manufacturer. If possible, portions of the

circuit should be powered on individually to verify functionality prior to operating

the entire system. If a portion of the circuit is not operating correctly, the hope

would be to identify the issue and resolve prior to power up the entire system.

When initially powering on, the setup should be double checked. Once this is done,

the input current should be set relatively low and the source voltage should be set

to zero volts. Slowly ramping up the source voltage allows you to monitor the input

current for abnormal current draw. Once this is done, it is best to turn on the source

voltage as it was intended to be done so that the circuit can operate in its intended

configuration.

During initial checkout and verification, it was noticed that the drive voltage

to the Power MOSFET’s was low, approximately 5.5V. For logic level MOSFET’s,

this would be sufficient. For the MOSFET’s used in the design, this drive voltage

was not sufficient to fully enhance the MOSFET. MOSFET’s that are not fully

26
enhanced have increased on state resistance and will increase losses in the

converter. Luckily, the design could be modified to drive the MOSFET’s at a higher

voltage. The configuration in Figure 5-1 (the original configuration) would need to

be connected as shown in Figure 5-2 [10]. As you can see, this included connecting

NDRV, INTVcc, and EXTVcc pins together as well as bringing in an external bias.

To achieve this, the EXTVcc pin needed to be disconnected from ground and the

NDRV MOSFET and Resistor shown in Figure 5-1 were removed.

Figure 5-1: Configuration for Generating Bias

Figure 5-2 Configuration for Using an External Bias

27
 Making the modifications outlined in the previous paragraph are straight

forward from an electrical standpoint, but from a physical standpoint, it was difficult.

Two traces needed to be cut (shown in Figure 5-3), components removed, and two

jumpers were needed. The biggest issue was with one of the cuts. It ran

underneath the LTC3814 controller which meant that the controller needed to be

de-soldered, the trace could then be cut, and the controller could afterward be re-

soldered.

Figure 5-3 Locations of Cut Traces for Modification

Removing the controller was only necessary for the first unit since it was

built and tested prior to any other converters being built. The following units had

the modifications done before soldering on the controller. The cut traces are shown

in Figure 5-4.

28
 Figure 5-4: Modifications to traces on PCB

In addition to the modifications of the circuit components and PCB

modification, there was also the need to apply an external bias to the chip. This

voltage was chosen so that the MOSFET’s could be fully enhanced and provide

the lowest on state resistance. Driving the gate to 10V was sufficient to fully

enhance the MOSFET according to the datasheet. In order to apply the voltage to

the circuit, wires were soldered to the appropriate nets. The modified circuit with

bias wires is shown in Figure 5-5. Once these modifications were made, the unit

could be tested and data could be collected. A completed converter with

modifications is shown in Figure 5-6.

29
Figure 5-5: Completed Modifications for Applying External Bias Voltage

Figure 5-6: Completed Converter with Modifications

30
 During first power on into regulation, it initially appeared as if the converter

was not boosting correctly. The waveform shown in Figure 5-7 was observed.

Figure 5-7: First Power on Observed Input (Ch. 1) and Output (Ch. 2)

Initially, the output voltage follows the input voltage by conducting through

the body diodes of the MOSFETS and through the Or-ing diode. Then, at

approximately 320ms, the input/output voltage reach the undervoltage lockout

threshold and the converter goes through its turn-on sequence. As you can see on

Channel 1 of Figure 5-7, the input voltage dips down. It dips down far enough to

shut down the converter by reaching the undervoltage lockout threshold. Although

there is hysteresis in that circuit to prevent such an event from occurring, the dip

was excessive and caused this “hiccup” to occur and prevent the converter from

properly regulating. The cause of this issue could be contributed to two reasons.

The first reason is the fact that the input lines, including source connection and

converter connections, created a significant impedance to the input to the

31
converter circuitry. This created a significant voltage drop when the converter

began boosting and drawing the increased current. A secondary cause to the drop

was the fact that the input source had the current limit set too low which forces the

source to reduce its output voltage to keep the current constant. The input current

is high during startup since the output is not in regulation and needs to boost the

output voltage to charge the bulk output capacitance. Once these two issues were

resolved, the output came into regulation as shown in Figure 5-8.

Figure 5-8: First Successful Turn on into Regulation.

32
 5.2. Results

5.2.1. Single Power Supply Performance

Initially, each power supply was tested individually to characterize the

performance as a standalone converter. It was also useful to perform this testing

on the first build in case changes were needed so they could be more easily

incorporated into subsequent builds. This occurred as in the case of making the

bias voltage modification as described in section 5.1.

When powering up the power supply with no load attached, the power

supply tended to regulate at a higher voltage than nominal. The output point of

regulation is prior to the Or-ing diode as described in section 4.1. When no load is

attached, the Or-ing diode is conducting a minimal amount of current, if any. This

means the voltage drop is minimal and the output capacitors are allowed to peak

charge with only parasitics left to discharge the output. The output voltage

regulation set-point must be set higher so that when the power supply is loaded to

full load the voltage drop across the Or-ing diode is not high enough to cause the

output to fall below the required regulation point. The power supply was

characterized and at approximately 50mA out, the output is within 0.1V of the

intended nominal output voltage. It is unlikely that there will be no load on the

output when there is a device connected to the output so the higher regulation

should not be an issue. Table 5-1 shows regulation characteristics of Power Supply

1 over the input line (22V-26V) and load range (0A-3.2A). The Power Supply is

within the requirements specified in Section 3.

33
 Table 5-1 Power Supply Regulation

Over Line and Load

(>50mA)

No
Max out Min Out
Load

Voltage (V) 49.286 48.0 47.9

Percent
2.7% 0.0% 0.2%
Regulation

When measuring output ripple, the oscilloscope should be bandwidth limited

to 20MHz to avoid erroneous measurements. It is also assumed that there should

be some sort of localized filtering at the point of measurement so that there is an

accurate reading of the output ripple. A small amount of capacitance will filter out

any external noise pickup and will provide a more accurate measurement for ripple.

Using a tight scope probe, the output ripple of the unit was observed during various

operating conditions. The worst-case ripple observed was during full load

operation and is shown in Figure 5-9. This ripple for a standalone converter meets

the requirement of Section 3, 286mV peak to peak measured compared to the

960mV peak to peak allowed.

34
 Figure 5-9: Power Supply 1 Ripple at Full Load

Data was also taken for the efficiency of the converter over the load and line

range. The data for Power Supply 1 is shown in Figure 5-10.

PS #1 Efficiency
98.0%
96.0%
94.0%
Efficiency

92.0% Nominal Input

90.0% (24Vin)
High Input
88.0% (26Vin)
86.0% Low Input
(22Vin)
84.0%
0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 180.0
Pout(W)

Figure 5-10: Power Supply 1 Efficiency

35
 Data was taken for each power supply and efficiency data for the remaining

power supplies are shown in Figure 5-11, Figure 5-12, and Figure 5-13.

PS #2 Efficiency
96.0%

94.0%

92.0%
Efficiency

90.0%
Nominal Input
88.0% (24Vin)
High Input
86.0% (26Vin)
Low Input (22Vin)
84.0%

82.0%
0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 180.0
Pout(W)

Figure 5-11: Power Supply 2 Efficiency

PS #3 Efficiency
96.0%

94.0%

92.0%
Efficiency

90.0% Nominal Input

(24Vin)
88.0% High Input
(26Vin)
Low Input (22Vin)
86.0%

84.0%
0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 180.0
Pout(W)

Figure 5-12: Power Supply 3 Efficiency

36
 PS #4 Efficiency
96.0%

94.0%

92.0%
Efficiency

90.0%
Nominal Input
88.0% (24Vin)
High Input
(26Vin)
86.0%
Low Input (22Vin)

84.0%

82.0%
0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 180.0
Pout(W)

Figure 5-13: Power Supply 4 Efficiency

From the data shown in Figure 5-10 through Figure 5-13, the curves

between each converter do not match up exactly. This is caused by various factors.

There could be differences in the component characteristics such as on state

resistance and forward voltage drop. The construction of the converter could have

caused increased conduction losses if components do not have adequate

soldering. Even the weather could play a role since component characteristics

could change over temperature. Since there were extended gaps between test

data being taken for each converter, the test setup and equipment also changed

since equipment was being purchased throughout the process. A list of the

equipment used is shown in Table 5-2. Although these various variables played a

role in the outcome of the data, the fundamental characteristics were similar.

37
 Table 5-2: List of Test Equipment Used

Part
Description Manufacturer Quantity
Number

Digital Multimeter Tekpower TP2844R 1

Digital Multimeter Tacklife DM05 3

Digital Multimeter Tacklife DM02A 4

Oscilloscope Rigol DS1054Z 1

Electronic Load Kunkin KL283 2

DC Power Supply
Eventek KPS305D 1
(5A)

DC Power Supply
Eventek KPS3010D 4
(10A)

5.2.2. Parallel Power Supply Testing

To reiterate what was previously stated, great care was taken in testing to

prevent unsafe and potentially damaging conditions. Before testing the converters

in parallel, voltage from a controlled source was back fed into the output of one of

the power supplies to ensure the Or-ing diode circuit was working properly and no

damage was observed on the unit under test. Once this was confirmed, parallel

testing could continue, and the converters were characterized.

Initially, only two power supplies were built and then those two were tested

in parallel to observe any interaction between the power supplies. This was done

prior to building up the remaining power supplies to more easily incorporate any

38
changes in the event of a change being needed to improve interoperability

between the power supplies. The test showed there was nothing that caused any

issues and the first two units operated successfully and the remaining power

supplies could be constructed.

The power supply load sharing data for the first and second power supply

is shown in Figure 5-14. Additional data was taken for the remaining power

supplies to ensure each power supply was built correctly and validate there were

no defects from the construction process. Each power supply was paralleled with

another power supply and similar data was taken as in Figure 5-14. The power

supplies tend to settle out and current share at approximately 7% load difference

for loads of 150W and above when two power supplies paralleled.

Percent load share difference (1 and 2)

16.0% PS 1 & 2 Load
Sharing
14.0%

12.0%
Percent Difference

10.0%

8.0%

6.0%

4.0%

2.0%

0.0%
0.0 50.0 100.0 150.0 200.0 250.0 300.0 350.0
Pout total
(W)

Figure 5-14: Power Supply 1 and 2 Load Share Difference

Each time power supplies were operated in parallel for the first time, the

regulation point had to be dialed in for those units in operation. That initial dialing
39
in of the regulation voltage is what provides the closest load sharing for that

configuration. Below is data for the various units operating in parallel. As can be

seen from Figure 5-15, Figure 5-16, and Figure 5-17, the load sharing between

two power supplies does not stay constant over the load range. The two paralleled

power supplies can alternate between providing more or less current. This is

caused by differences in component characteristics and differences in line

impedances.

Load Difference Vs Total Load PS 1 & 2

4
Individual PS Current (A)

3.5

2.5

2
PS 1
1.5 Load

1 PS 2
Load
0.5

0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0

Total Output Load Current (A)

Figure 5-15: Load Difference PS 1 & 2

40
 Load Difference Vs Total Load PS 2 & 3
3.5

Individual PS Current (A) 3

2.5

2
PS 2
1.5 Load

1 PS 3
Load
0.5

0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0

Total Output Load Current (A)

Figure 5-16: Load Difference PS 2 & 3

Load Difference Vs Total Load PS 2 & 4

4
Individual PS Current (A)

3.5

2.5

2
PS 2
1.5 Load

1 PS 4
Load
0.5

0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0

Total Output Load Current (A)

Figure 5-17: Load Difference PS 2 & 4

41
 As indicated in previous sections, incremental steps were taken prior to

operating the power supplies under the intended configuration to avoid mistakes

and potential damage to the power supplies. The next step in the testing process

was paralleling three power supplies and observing how they operate. Each

converters output current was documented as well as the shared output voltage.

The worst-case load sharing is shown in Figure 5-18. Initially the load sharing

difference is large, but as the load is increased the converters share load more

closely. Above approximately 270W out, the converters are sharing within 10% of

each other. The actual load curves while operating in parallel are shown in Figure

5-19.

Worst Case Percent Load Share Difference (2,3, & 4)

60.0%

50.0% PS 2, 3, & 4
Load Share
Percent Difference

40.0% Difference

30.0%

20.0%

10.0%

0.0%
0.0 50.0 100.0 150.0 200.0 250.0 300.0 350.0 400.0 450.0 500.0
Pout total
(W)

Figure 5-18: Worst Case Load Sharing for PS 2, 3, & 4

42
 Individual Load Vs Total Load PS 2, 3, & 4
4

Individual PS Current (A) 3.5

2.5

2
PS 2 Load
1.5

1 PS 3 Load

0.5 PS 4 Load
0
0.0 2.0 4.0 6.0 8.0 10.0 12.0

Total Output Load Current (A)

Figure 5-19: Load current of PS 2, 3, & 4 operating in Parallel

The final configuration tested was operating the four converters while

paralleling all outputs as shown in Figure 3-1. Testing parallel converters became

increasingly more difficult to test due to lack of equipment and available test

cabling. More Banana connectors were purchased to create custom wiring.

Individual cables were easily constructed using 14-gauge, stranded wire and

banana cables. To make paralleling outputs easier, a custom output wiring cable

was created. The beginning of the wires had banana plugs and would connect to

each of the outputs individually. Those four wires were stripped approximately 1.5ft

from the end and then that section was combined with all the wires. This section

provided a single point of connection between all the outputs. There was an

additional 2.5ft for the cable to reach the electronic load. A picture of the output

load cables is shown in Figure 5-20.

43
 Figure 5-20: Custom Load Cables

Four individual sources were used to power four individual converters, the

outputs of the converters were paralleled using a load cable. The load cable fed

into two electronic loads, which each had two adjustable loads. The output currents

of each of the converters were monitored using multimeters. The combined output

voltage, at the single point connection, was monitored using a multimeter. Some

of the input voltages and output voltages (at the converter) were also monitored

with leftover equipment. A block diagram of the setup is shown in Figure 5-21 and

the actual setup is shown in Figure 5-22.

44
Figure 5-21: Block Diagram of Four Paralleled Converters

Figure 5-22: Four Converters Paralleled Testing Setup

45
 Data was taken of each output converter current using the multimeter to

monitor the currents. The converters outputs were paralleled, and data was

collected for the individual currents as well as the combined output voltage. Figure

5-23 shows the individual load currents supplied by each of the converters over

the total load range. The data collected was also used to capture the worst-case

load difference over the entire power range and is shown in Figure 5-24. The output

current difference is less than 10% worst-case, above loads of approximately

320W out. In total, the four combined converters were able to supply a 48V load

with a measured 12.9A equating to approximately 620W which meets the

requirements outlined in Section 3. The efficiency of the combined converters was

not taken but efficiency of the converters should be approximately the same as the

individual converters. From Figure 5-10 through Figure 5-13, all converters have

efficiency above 92%. It is safe to assume that the combined efficiency of the 4

operating in parallel meets the requirement of 75% efficiency and will be an

average of the individual converter efficiencies.

46
 Individual Load Vs Total Load PS 1, 2, 3, & 4
4.00

Individual PS Current (A) 3.50

3.00

2.50

2.00
PS 1 Load
1.50
PS 2 Load
1.00
PS 3 Load
0.50
PS 4 Load
0.00
0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0

Total Output Load Current (A)

Figure 5-23: Four Paralleled Converters Shared Load Current

Worst Case Percent Load Share Difference (1,2,3, & 4)

50.0%
45.0%
Load Share
40.0%
Percent Difference

35.0%
30.0%
25.0%
20.0%
15.0%
10.0%
5.0%
0.0%
0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0
Pout total
(W)

Figure 5-24: Worst Case Load Difference of 4 Paralleled Converters.

47
 In addition to the current sharing data taken, ripple data was also taken.

Ripple data was taken in the paralleled configuration since the output ripple of each

of the converters could interact with each other. The ripple was taken at the

common point and a small amount of capacitance was added since the

measurement point had a large loop area susceptible to high frequency noise.

With 1µF of local capacitance, the worst-case ripple was measured to be 360mV

peak to peak and is shown in Figure 5-25. The ripple is larger than was seen on a

single converter since the individual converters are most likely operating at slightly

different frequencies and the ripple waveforms will create constructive

interference. So, Figure 5-25 demonstrates constructive interference when the

ripple is larger and destructive interference when the ripple is reduced. Even with

this interference, the ripple remained within the requirements of 2%, 960mVpp.

Figure 5-25 Ripple at 12.8A Total Load Current

48
6. Conclusion

Great care was taken in the development of this converter to make it a

feasible design. Overall this project was successful in developing and

demonstrating the possibility of using multiple input sources to provide a higher

power single output through the use of parallelable converters. The fundamental

requirements set out for this design were all met and included the use of 4

individual 24V nominal input sources, a 48V output providing 600W total output

power, greater than 75% efficiency, tight output regulation, and low output ripple.

The costs of the converters were kept relatively low by using readily available

commercial components. There were some issues seen during the development

which were rectified to ensure success of the design.

The issue which was seen regarding the low bias voltage in Section 5.1

should be incorporated into the design. This would imply incorporating the

modifications of the external bias configuration. In addition, it would be ideal to

have a local bias converter which is powered by the same input as the main

converter. Incorporating this bias would allow the converter to be a standalone

package. Since the efficiency of the converter is relatively high, it may be possible

to incorporate a simple linear regulator if the bias power is low and the additional

losses are tolerable. Additional thermal considerations would have to be made to

ensure proper cooling is provided for the linear regulator.

Although in practice this design works, there is room for improvement. Since

this project would be used in conditions where there would be extended use of the

product, it is important to make sure the design is robust. The improvements which

49
should be incorporated would be protection circuitry. Protection circuitry would be

needed to protect the converter from any abnormal conditions which could

potentially cause failures. A more robust design would reduce the likelihood of

failures, reduce overall costs, and increase the life of the product. In addition, there

are other improvements which could be made to allow the converter to be used for

additional scenarios. Here I would like to discuss some potential improvements to

the design.

The first improvement which should be made would be to have the ability to

operate the converter over a wider input voltage range. This could allow the use of

different input sources which operate over this larger range. When there are more

options to use, there is a larger group of individuals this product could serve.

Essentially it would increase the number of sources it would be compatible with

which would then expand the consumer base. This could be achieved by

increasing the current rating of the power components which would improve the

low voltage operational range. In addition, higher output power at nominal input

voltages could potentially be achieved with the increased current capabilities. In

order to operate at higher input voltages, components with higher voltage ratings

could be used. These improvements should only be incorporated if there is a need

and if the higher costs of better performing components is not significant.

In order to prevent the input source from damaging the converter when the

input limits are exceeded, there should be an over voltage protection circuit. The

circuit would shut down the converter if the input voltage is too high for the ratings

of the components. This could easily be incorporated by using the same circuit that

50
was used for the undervoltage protection circuit as discussed in Section 4.1. This

circuit consisted of a comparator circuit with hysteresis which toggled the Run/Soft

Start pin of the controller whenever the sensed threshold was exceeded.

There is also the potential for the output load to draw excess current which

could damage the converter and also the source. Although the controller chip had

the ability to incorporate overcurrent protection of the low side switch it was not

feasible to use in this design. The low side on-state resistance of the MOSFET

were kept low to reduce the power losses of the converter as much as possible.

This on-state resistance is also used by the controller chip to monitor the current

by sensing the voltage drop across the MOSFET. The overcurrent threshold is

inversely proportional to the on-state resistance so to use this functionality, another

MOSFET with higher on-state resistance would have to be used. This would have

an adverse effect to the power dissipation and efficiency so an external sense

resistor could be used to measure the output current of the converter. This sense

current could then provide feedback to shutdown the converter and restart after

some thermally insignificant rate in the event that the overcurrent condition is

cleared. It would also be nice to have some sort of indicator light which would notify

the user that an overcurrent condition was observed which could help in

troubleshooting the issue.

Overall this project met the intent of supporting the required performance

for the DC House project. The design shows the potential for further growth to

expand on the performance capabilities of the project. This includes the ability to

use more than four input sources to supply greater total output power capabilities.

51
BIBLIOGRAPHY

[1] D. A. Jones, "Electrical engineering: the backbone of society," in IEE

Proceedings A - Science, Measurement and Technology, vol. 138,
no. 1, pp. 1-10, Jan. 1991.
URL: http://ieeexplore.ieee.org.ezproxy.lib.calpoly.edu/stamp/stamp.jsp?tp
=&arnumber=61630&isnumber=2242

[2] Bose, Bimal K., “Power electronics” Engineering and Technology History
Wiki, University of Tennessee (Knoxville), 15 September 2014, [Online]
Available: https://ethw.org/Power_electronics#Power_Converters

[3] Sulzberger, Carl., “Milestones:Pearl Street Station, 1882”, Engineering

and Technology History Wiki, IEEE Region 1, 10 October 2011, [Online],
Available: https://ethw.org/Milestones:Pearl_Street_Station,_1882

[4] Rashid, Muhammad H., “Power Electronics Circuits, Devices, and

Applications Third Edition”, Pearson, 2003

[5] Dilip Ahuja and Marika Tatsutani, “Sustainable energy for developing
countries”, S.A.P.I.EN.S 2.1, 27 November 2009, [Online], URL :
http://journals.openedition.org/sapiens/823

[6] Taufik, “The DC House Project: An Alternate Solution for Rural

Electrification”, IEEE Global Humanitarian Technology Conference,
October 2014.

[7] Wong, Taffy Chi Kin. “Multiple Input Single Output DC-DC Converter for
the DC House Project”. October 2011.

[8] Jong, Owen. “Multiple Input Single Output (MISO) DC-DC Converter For
the DC House Project”. 2012

[9] B. T. Irving and M. M. Jovanovic, "Analysis, design, and performance

evaluation of droop current-sharing method," APEC 2000. Fifteenth
Annual IEEE Applied Power Electronics Conference and Exposition (Cat.
No.00CH37058), New Orleans, LA, USA, 2000, pp. 235-241 vol.1.doi:
10.1109/APEC.2000.826110
URL:
http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=826110&isnum
ber=17861

[10] Linear Technology, “60V Current Mode Synchronous Step-Up Controller,”

LTC3814-5 datasheet, Jan 2011

52
APPENDIX A

53
APPENDIX B

54
APPENDIX C

Qty Per Unit

Cost Per Board
Part Mfg PN Board Price
STAND1 8841
STAND2 8841 4 $ 0.27 $ 1.07
STAND3 8841
STAND4 8841

L1 74435582200 1 $ 7.83 $ 7.83

C3 100TXW220MEFC8X60
C9 100TXW220MEFC8X60 3 $ 1.51 $ 4.53
C14 100TXW220MEFC8X60
V_IN 150080SS75000 1 $ 0.29 $ 0.29
IN_GOOD 150080VS75000 1 $ 0.29 $ 0.29
R23 2-1614885-8 1 $ 0.18 $ 0.18

Q1 2N7002P,215 1 $ 0.13 $ 0.13

R2 4-2176093-9 1 $ 0.54 $ 0.54
PAD1 575-4
PAD2 575-4
4 $ 0.75 $ 2.98
PAD3 575-4
PAD4 575-4

D6 BAT46ZFILM
2 $ 0.28 $ 0.55
D7 BAT46ZFILM
C4 C0603C563J4RACTU 1 $ 0.31 $ 0.31

C11 EEE-FTH101XAP 1 $ 0.53 $ 0.53

C7 GCM21BR72A104KA37L
C1 GCM21BR72A104KA37L
4 $ 0.08 $ 0.30
C36 GCM21BR72A104KA37L
C37 GCM21BR72A104KA37L
C8 GRM1885C1H101JA01D 1 $ 0.03 $ 0.03
C5 GRM1885C1H151JA01D 1 $ 0.05 $ 0.05
C6 GRM1885C1H331JA01D 1 $ 0.05 $ 0.05
C19 GRM219R7YA105KA12D
C2 GRM219R7YA105KA12D 3 $ 0.12 $ 0.35
C33 GRM219R7YA105KA12D

55
 C15 GRM21BR72A103KA01L
C16 GRM21BR72A103KA01L
C17 GRM21BR72A103KA01L
C18 GRM21BR72A103KA01L
8 $ 0.04 $ 0.32
C20 GRM21BR72A103KA01L
C21 GRM21BR72A103KA01L
C34 GRM21BR72A103KA01L
C35 GRM21BR72A103KA01L
C10 GRM32ER72A105MA01L
C12 GRM32ER72A105MA01L
C22 GRM32ER72A105MA01L
C23 GRM32ER72A105MA01L 7 $ 0.38 $ 2.66
C24 GRM32ER72A105MA01L
C29 GRM32ER72A105MA01L
C30 GRM32ER72A105MA01L
PAD5 H9004-01
PAD6 H9004-01
PAD7 H9004-01 4 $ 0.33 $ 1.30

PAD8 H9004-01
Q2/Q3 IPB020NE7N3 G 2 $ 3.90 $ 7.80

1 $ 0.32 $ 0.32
D9 LM4040CYM3-2.5-TR
U2 LT1716CS5#TRMPBF
2 $ 2.50 $ 5.00
U4 LT1716CS5#TRMPBF

LTC3814IFE or 1 $ 7.18 $ 7.18

U1 LTC1814EFE
R7 MCR03ERTF3003 1 $ 0.02 $ 0.02
R16 MCR03ERTJ000
2 $ 0.01 $ 0.01
R22 MCR03ERTJ000
R15 MCR10ERTF1001 1 $ 0.02 $ 0.02
R18 MCR10ERTF1002
2 $ 0.02 $ 0.04
R21 MCR10ERTF1002
R8 MCR10ERTF1003
2 $ 0.02 $ 0.04
R19 MCR10ERTF1003
R3 MCR10ERTF4990 1 $ 0.02 $ 0.02
R9 MCR10ERTF4991
2 $ 0.02 $ 0.04
R14 MCR10ERTF4991
R25 MCR10ERTF49R9 1 $ 0.02 $ 0.02
R5 MCR10ERTF6491 1 $ 0.02 $ 0.02
R4 MCR10ERTF7872
R17 MCR10ERTF7872 3 $ 0.02 $ 0.06
R20 MCR10ERTF7872
56
 R1 MCR10ERTF8063 1 $ 0.02 $ 0.02
R24 MCR10EZPJ000 2 $ 0.02 $ 0.04
R6 MCR25JZHF1003 1 $ 0.16 $ 0.16
R13 MCR25JZHF4991 1 $ 0.16 $ 0.16
RPOT TC33X-2-101E 1 $ 0.27 $ 0.27
D1 V10P10-M3/86A
D2 V10P10-M3/86A
D3 V10P10-M3/86A 5 $ 0.73 $ 3.64
D4 V10P10-M3/86A
D5 V10P10-M3/86A
PWB 1 $ 12.40 $ 12.40

Total Converter
Cost
$ 61.57

57
APPENDIX D

58
APPENDIX E

59