DC-DC Converter for Electric Vehicle

By

Jason Zhou

Nicholas Mah

Senior Project

Project Advisor: Professor Taufik

ELECTRICAL ENGINEERING DEPARTMENT

California Polytechnic State University

San Luis Obispo

June 2019
Table of Contents

DC-DC Converter for Electric Vehicle ........................................................................... 1

List of Figures ..................................................................................................................... 3

List of Tables ..................................................................................................................... 4

Abstract .............................................................................................................................. 5

1. Introduction ............................................................................................................... 6

2. Background ............................................................................................................... 8

3. Design Requirements .............................................................................................. 11

4. Design and Simulation Results ............................................................................... 17

5. Hardware Test and Results ..................................................................................... 42

6. Conclusion .............................................................................................................. 54

7. References ............................................................................................................... 57

8. Appendix ................................................................................................................. 58

Code for Transformer Design ....................................................................................... 58

Timeline of Tasks and Milestones ................................................................................ 67

Bill of Materials ............................................................................................................ 68

Analysis of Senior Project Design ................................................................................ 76

2
List of Figures

Figure 3-1: Level 0 Block Diagram .................................................................................. 15

Figure 3-2: Level 1 Block Diagram .................................................................................. 15

Figure 4-1: Switch Configuration ..................................................................................... 26

Figure 4-2: Final Simulation Schematic ........................................................................... 30

Figure 4-3: Efficiency Plot of Optimal Configuration ...................................................... 30

Figure 4-4: DC-DC Top Schematic .................................................................................. 39

Figure 4-5: LT8311 Schematic ......................................................................................... 39

Figure 4-6: Forward Active Schematic ............................................................................. 39

Figure 4-7: Flyback Schematic ......................................................................................... 39

Figure 4-8: LT3752-1 Schematic ...................................................................................... 40

Figure 4-9: OptoIsolator Schematic .................................................................................. 40

Figure 4-10: Input Fusing and RVP Schematic ................................................................ 40

Figure 4-11: DC-DC Disable Schematic .......................................................................... 40

Figure 4-12: Sync Transformer Schematic ....................................................................... 41

Figure 4-13: Gate Driver Schematic ................................................................................. 41

Figure 5-1: Final Constructed Transformer ...................................................................... 42

Figure 5-2: Low Leakage Winding Depiction .................................................................. 43

Figure 5-3: Crossover Winding Depiction ........................................................................ 43

Figure 5-4: Transformer Measured Primary Side AC Resistance .................................... 45

Figure 5-5: Transformer Measured Secondary Side AC Resistance ................................ 45

Figure 5-6: Transformer Turns Ratio Validation .............................................................. 46

3
 Figure 5-7: DC-DC Main Board Layout ........................................................................... 47

Figure 5-8: DC-DC 3D Render ......................................................................................... 48

Figure 5-9: DC-DC Final Constructed Board ................................................................... 48

Figure 5-10: DC-DC Multimeter, Supply, and Load Block Diagram .............................. 49

Figure 5-11: DC-DC Oscilloscope Measurement Point Setup ......................................... 50

Figure 5-12: LT3752-1 and LT8311 Controller Signal Waveforms (Ignore Blue

Waveforms)................................................................................................................................... 51

Figure 5-13: DC1929A OUT and SWP ............................................................................ 52

Figure 5-14: DC1929A FG and FSW ............................................................................... 52

Figure 5-15 ........................................................................................................................ 52

Figure 5-16 ........................................................................................................................ 53

Figure 8-1 .......................................................................................................................... 68

Figure 8-2 .......................................................................................................................... 68

Figure 8-3 .......................................................................................................................... 79

Figure 8-4 .......................................................................................................................... 79

List of Tables

Table 3-1: Requirements List............................................................................................ 16

Table 5-1: Transformer Design Test Data ........................................................................ 44

Table 5-2: Efficiency Test Results.................................................................................... 51

Table 5-3: DC1929A Demo Board Waveforms ............................................................... 52

Table 8-1: Final Board Bill of Materials........................................................................... 69

Table 8-2: Original Cost Estimate .................................................................................... 78

4
Abstract

In this work, a DC-DC converter is designed for an electric vehicle. The DC-DC

converter is designed to provide 500W with a 200-400V input and a 12-15V adjustable output.

Electric vehicle sales are beginning to increase in popularity and the need for DC-DC converters

to siphon power from the tractive system is not yet fully satisfied, especially for single-seater

class vehicles. Additionally, improving performance in efficiency without sacrificing wide input

voltage range can benefit future DC-DC converter designs. In the end, a forward active clamp

DC-DC converter is designed and tested. Additionally, spreadsheet calculators, LTSpice

simulations, and Matlab scripts were made to assist in work for the DC-DC.

5
 1. Introduction

Electric Vehicle (EV) sales have grown dramatically since 2010 [1]. The first mass-

produced EV, the EV1, sported a range of 70 to 100 miles, while a base-model Tesla Model 3

has a range of 264 miles [2] [3]for similar costs. Improvements in total range and lifetime of

EV’s is mostly responsible for this massive increase in popularity. The two main ways of

improving EV range is through system efficiency and improved battery technology.

System efficiency improvements are mostly due to the use and improvement of power

electronics systems within the vehicle. Power electronics refers to the use of switching

electronics for power conversion and distribution. There are six main categories of power

converters: AC-DC converters (rectifiers), AC-AC converters, DC-AC converters, static

switches, and DC-DC converters. AC-DC, DC-DC, and DC-AC converters. These play

important roles for EV usage. AC-DC converters are often used for on-board charging systems,

which allows the EV to be charged from the grid. DC-AC converters are typically used to drive

the main motors of the vehicle, providing propulsion [4]. In most EV’s, there is a high voltage

system consisting of the vehicle’s battery pack and drive system, and a low voltage system,

containing the vehicle’s media entertainment, sensors and telemetry, control unit, etc. To reduce

weight and improve vehicle packaging, most EV’s use a DC-DC converter with power capability

on the order of kilowatts to efficiently step down the vehicle’s battery pack voltage, 200V-800V,

to the vehicle’s low voltage system operating voltage, usually 12 or 24V [5].

There are many different types of DC-DC converters, but they can be divided into two

main categories, isolated and non-isolated. Isolated designs incorporate a galvanic isolation

between the input and output while non-isolated systems do not. Isolated designs offer a high

6
degree of safety because failures on the input side are generally isolated from the output, but

non-isolated designs are typically smaller because they do not require a galvanic isolating device.

A standard DC-DC converter will contain a switching device, an inductor, and input and output

capacitors. In an isolated device, a transformer will typically be included as the method of

galvanic isolation.

In order to increase knowledge in experience in automotive development, the Society of

Automotive Engineers (SAE) at Cal Poly State University created the Formula SAE Student

Design Competition (FSAE). In FSAE, students design, build and compete a single-seater race-

car. Currently, there are both an Internal Combustion (IC) and EV competition. Internally, the

EV race-car contains many similar components to a commercial EV, including a HV (High

Voltage) to LV (Low Voltage) DC-DC converter. For race-car, development size and weight are

among the most important design metrics. Reducing weight and size allows the vehicle to turn

and accelerate much faster, improving the race-car’s chances of winning the race. By extension,

system efficiency is also very important because improving the range of a vehicle will allow

designers to reduce the total volume and mass of batteries required. This project will explore the

design of a DC-DC converter for an FSAE EV.

7
 2. Background

Cal Poly’s Formula SAE club, Cal Poly Racing, needs a smaller, higher voltage rated

DC-DC converter because their existing solution, the RSP-500-12 [6], is only rated at 370 VDC,

which a slight 0.4 V more than the expected maximum operating voltage of 369.6 VDC. The

purpose of the DC-DC converter is to use the “high-voltage” 316 V battery pack, which is used

to drive the electric motor, to power all of the low voltage electronics while the car operates.

During times when the high-voltage battery is disconnected, a “low-voltage” 14 V battery is used

to maintain low voltage electronics. Although a low voltage battery could be used instead of a

DC-DC converter, the size of such a battery is infeasible for the packaging constraints of a

Formula SAE Electric vehicle.

The Formula SAE Electric vehicle can draw 500W of power continuously on the low

voltage system and the high voltage battery pack voltage can vary from less than 246 V to 370

V. Additionally, the DC-DC converters must fit in a narrow enclosure next to the battery pack in

order to pass high voltage isolation rules. These requirements can be summarized as a DC-DC

converter that must be able to transform 200-400 VDC to 14 VDC at a power level of 500 W in

the smallest package possible.

Because climate change has become a greater concern in the 21st century, the automotive

industry is slowly moving towards hybrid and fully-electric vehicles. As a result, the use of DC-

DC converters in electric vehicles has been a topic of development and research in the

automotive industry. Most DC-DC converters used in hybrid or electric vehicles consist of both

unidirectional and bidirectional DC-DC converters. The unidirectional converters generally have
8
been used for vehicle sensors, controls, entertainment, utility, and safety purposes. Whereas the

bidirectional DC-DC converters have found their use in battery charging and regenerative

braking. Many modern day DC-DC converters use soft-switching technique to achieve higher

efficiency and are isolated to provide safety to the load. However, the use of transformers leads

to leakage inductance, which creates larger voltage stress on converter components due to

ringing and semiconductor output capacitance. The presence of leakage inductance has also led

to the use of passive snubbers, active-clamping, and soft-switching. Additionally, transformers

increase converter area, volume, weight, cost, and EMI. Although the development of

bidirectional full-bridge converters are in high demand, half-bridge topologies have seen

significant development, with their reduction in number of components. As a result, converters

such as full-bridge primary and half-bridge secondary with synchronous rectification using

coupled inductors have contributed a three times size reduction compared to its full-bridge

counterpart. Lastly, wide band gap semiconductors are slowly being incorporated into EV DC-

DC converters as they have better thermal conductivity, breakdown, and maximum operating

junction temperature than silicon. Thus, SiC (Silicon Carbide) and GaN (Gallium Nitride) based

semiconductors will reduce temperature issues, and have been used in hybrid vehicles like the

Toyota Prius II.

The University of Wisconsin-Madison’s Formula SAE team, Wisconsin Racing,

published their Electrical Safety Form, documenting their solution to the DC-DC stage design.

Wisconsin Racing utilized the Vicor Power DCM290P138T600A40 DC-DC module in

conjunction with the LTC4417 PowerPath Controller to supply power to the Low Voltage

system. With this DC-DC module, they were able to package a system with power flow control,

pre-charge and discharge circuitry, a shutdown circuit, and a battery management system, in an
9
enclosure approximately 3.6 by 5 in. enclosure [7]. Unfortunately, information from other

Formula SAE teams is not available. As such, most research is limited to general DC-DC

converter designs. However, the DCM290P138T600A40 is also used by Cal Poly Racing for this

year’s car design. The Vicor DC-DC module boasts over 90% efficiency at 500 W from 200-400

VDC input with internal transformer in a 1 by 2 in. enclosure. The Vicor DC-DC achieved this

with a 600 to 1000 kHz variable switching frequency, ZVS controller. However, the Vicor DC-

DC converter is not a fully standalone controller and it requires 1 mF of external output

capacitance and heat-sinks. Additionally, despite its high frequency operation, the Vicor DC-DC

has approximately 400mV p-p of ripple [8].

Another potential DC-DC module available is the Cosel SNDBS700B12. Unfortunately,

not much information can be gathered regarding the overall performance. The Cosel DC-DC is a

8.75 by 3.8 in. DC-DC converter with an efficiency of 90% from 200 to 400 V. An external heat-

sink is also required for the Cosel DC-DC. Although limited information is available for the

Cosel DC-DC, the given information implies that this solution is inferior to the Vicor DC-DC

[9].

This project is to design, build, and test of a 500W, 200-400 VDC in to 14 VDC out, high

efficiency, compact DC-DC converter. The Vicor DC-DC module is so much smaller than the

RSP-500-12 that Cal Poly Racing cannot take full advantage of the packaging. Our DC-DC

design will aim to produce a higher efficiency, lower output ripple DC-DC converter with a size

between the Vicor DC-DC and the RSP-500-12 that doesn’t require any external hardware.

10
 3. Design Requirements

This chapter explains technical design requirements for the project. It encompasses those

requirements related to electrical performance as well as physical dimension.

Technical Requirements

● Input Voltage Range

A 200-400 VDC input voltage range is a required design specification. The 88s8p LiIon

battery configuration means that the tractive system voltage will vary from 220V to 369V

during normal operation. Extending the range to 200-400V allows for a 20V safety

margin.

● Adjustable Output Voltage

A 12-15V adjustable output voltage is needed to integrate with the low voltage battery.

The low voltage battery used has its operating voltage vary from 11.2 to 14.4V. Cal Poly

Racing does not expect the LV battery to drop to the minimum 11.2V state of charge, so

a 12V minimum operating voltage is acceptable.

● Control and Sense Ports

DC-DC converter voltage and current data must be able to be constantly monitored to

ensure safe operation. Cal Poly Racing also expects adjustable output voltage with

external control.

● Continuous Output Power Rating

Cal Poly Racing expects to draw up to 400W of 12V power throughout the course of an

endurance. Although the power draw is not continuous throughout the entire race, the

11
 high power “bursts” can last for minutes, necessitating a continuous power rating of

500W.

● Input/Output Overcurrent Protection

Cal Poly Racing has a constant influx of new members. As such, assembly and operating

errors occur that project leads are not able to fully prevent. Because of this, output short

circuits are a possible event. In order to protect the DC-DC itself as well as the connected

harness, overcurrent protection must be included in the design.

● Input/Output Overvoltage Protection

It is possible for an overvoltage condition to exist during regenerative braking, or if an

incorrectly set voltage source is connected. The DC-DC must protect itself in order to

prevent the need for purchasing a replacement DC-DC.

● Input/Output Reverse Voltage Protection

As mentioned previously, assembly and operational mistakes are possible in the DC-

DC’s lifespan. Another common error that occurs is reverse voltage events. These events

are possible on the input and output of the device since energy sources in the form of

batteries are connected on both sides. Because of this, the device must withstand input

and output reverse voltage during a fault without permanent damage to ensure reliable

operation.

12
● Rated for Automotive Conditions

Device undergoes rigorous vibrations as it is operated within a moving vehicle. Thus,

most if not all components should be AEC-Q100 rated or better.

● Isolated Design

In accordance with 2019 Formula SAE Electric rules, EV.6.1.7, HV and LV on separate,

clearly defined areas of the PCB, HV and LV areas are clearly marked, spacing between

traces under conformal coating is more than 4mm. In accordance with EV.6.1.5, HV and

LV within an enclosure is separated by moisture resistant, UL rated insulating materials

rated for 150 degrees C or higher, and the spacing through air is greater than 30mm.

Additionally, at this power level, the design is isolated to prevent damage to low voltage

components during a fault.

● IP31 Rated

In accordance with 2019 Formula SAE Electric rules, EV.6.5.3, the enclosure and

connectors are protected from moisture with approximately an IP31 rating., meaning that

the device is resistant to dripping water and greater than 2.5mm probes Although the rain

test involves spraying water through a sprinkler directly into the car, The battery box

enclosure that will contain the DC-DC converter will provide spray protection.

● Small Volume

Because Cal Poly Racing requires smaller components to improve vehicle packaging, the

DC-DC converter must be roughly half the size of the currently used Meanwell RSP-500-

15 DC-DC converter, which implies physical size less than 115mm x 127mm x 40.5mm.

13
● Lightweight

In order to be a usable replacement for the current module RSP-500-12, the DC-DC

should be significantly lighter that the 2.866 lbs of the RSP-500-12. A target of less than

2 lbs total is a reasonable target.

● Efficiency

As a result of the previous requirements, converter’s overall efficiency of at least 90% is

required. The losses from the LV system is small compared to the total power

consumption.

● Line and Load regulations

Because of the large voltage and load range required of the DC-DC converter, a strict line

and load regulation is also required in order to keep the voltage from deviating too far

away from the nominal value. A maximum regulation of 5% was selected for both line

and load regulations.

14
 Figure 3-1: Level 0 Block Diagram

Figure 3-2: Level 1 Block Diagram

The Level 0 block diagram of the project shows that the proposed system consists of

three inputs and three outputs. The three inputs contain the “high voltage” input, DC-DC

converter enable control, and converter output voltage adjustment. The three outputs contain the

“low voltage” output, voltage and current sensing signals, and various temperature signals.

The Level 1 block diagram of the project, as illustrated in Figure 3-2, consists of the main

blocks on either side of the isolation. This includes the active clamp circuitry, various operating

15
protections, the two controller ICs, and various sensing blocks. Table 3-1 summarizes the design

requirements for the project.

Table 3-1: Requirements List

Requirement Value

Input voltage range 200-400V

Output Voltage Range 12-15V

Efficiency > 90%

Additional Features Control and Sense Ports

Power Rating 500W Continuous

Input/Output Protection Overvoltage, Overcurrent and Reverse

Voltage Protection

Ratings Isolated, IP31, and Vibration Resistant

Size < 115mm x 127mm x 40.5mm

Weight 2 lbs.

16
 4. Design and Simulation Results

The Linear Technology LT3752-1 Active Forward Clamp Controller met all of the design

requirements for the project. The IC is capable of ZVS and has a housekeeping controller. ZVS

allows for improved switching efficiency, and the housekeeping controller reduces the size of the

board by removing the need for an additional IC [6]. Additionally, the LT3752 has built-in

integration with the LT8311 Synchronous Rectifier Controller with Opto-Coupler Driver for

Forward Converters. This allows for synchronous rectification, which improves efficiency [7].

Excel calculators were made to determine component values.

Compared to other existing designs, our DC-DC converter uses a 300 kHz switching

frequency, compared to the LT3752 Demo Board’s 100 kHz switching frequency. This higher

switching frequency allows us to achieve a comparable size to the demo board while carrying 2.5

times the power [8].

17
Output Inductor

Equation 4-1: Output Inductor Value Based on Output Current Ripple

The output current ripple was originally designed to be the recommended 40% of IOUT(MAX), but

due to continuous and saturation current ratings of available inductors in our desired dimensions,

the current ripple was increased. The increased current ripple should be compensated by the

programmable slope compensation setting of the LT3752-1.

Primary MOSFET

Primary MOSFET selection depends on the maximum expected drain voltage and current.

Additionally, the power losses due to conduction, gate driver, and transition losses are used to

determine the best fit for the converter.

18
 Equation 4-2: Maximum expected steady state drain voltage

MOSFET BVDSS rating was at least 20% larger than the steady state VDS value to account for

tolerances in duty cycle, load transients, voltage ripple, and spikes caused by leakage inductance.

Equation 4-3: Total Primary MOSFET Losses at Maximum Output Current

Several MOSFETs with different RDS(ON) and QG values were considered for the primary

MOSFET, with the conduction loss being the primary loss. Additionally, the junction-to-case

and junction-to-ambient coefficients were considered when determining heatsinks and overall

package size. As a result, the LMG3410R050 Integrated GaN power state was chosen for its low

loss and low junction-to-ambient coefficient [9].

Equation 4-4: Primary MOSFET Conduction Losses

Equation 4-5: Primary MOSFET Gate Driver Losses

Gate driver losses were not able to be calculated for the chosen MOSFET, the LMG3410R050,

because of it’s fully integrated GaN design. Instead, the operating current value was used to

calculate the gate driver loss.

19
 Equation 4-6: Primary MOSFET Transition Losses

Transition losses only consist of turn-on losses, as it is assumed that ZVS is occurring on turn-off

and will prevent any turn-off losses. The ZVS on turn-off is guaranteed, but depending on clamp

switching timing, ZVS may be achieved upon turn-on, eliminating transition losses entirely.

Clamp MOSFET

Clamp MOSFET selection also depends on the maximum drain voltage and current. The main

optimization goal is also to reduce RDS(ON) and QG, while meeting peak drain and body diode

current ratings. Additionally, the maximum VDSS is the same value as the primary MOSFET.

Equation 4-7: Clamp MOSFET Maximum Drain Current

The maximum drain current of the clamp is significantly smaller than the primary MOSFET

drain current and is double for safety margin due to tolerances in magnetizing inductance and

transients.

20
 Equation 4-8: Clamp MOSFET Maximum Body Diode Current

The pulsed body diode current occurs due to the active clamp topology, where the energy in the

transformer’s leakage inductance is conducted through the clamp body diode after the primary

MOSFET turns off and before the clamp MOSFET is turned on.

Catch MOSFET

Catch MOSFET selection depends on the maximum drain-source voltage and current as well as

maximum gate-source voltage. The main optimization goal is also to reduce RDS(ON) and QG.

Equation 4-9: Catch MOSFET Maximum Drain-Source Voltage

The catch MOSFET margin was chosen to be 1.5, and is typically between 1.5 to 2, to allow for

a safety margin the VDS(MAX) equation. This accounts for voltage spikes from the leakage

inductance of the transformer’s secondary winding.

Equation 4-10: Catch MOSFET Maximum RMS Drain Current

21
 Equation 4-11: Secondary MOSFET Peak Current

The peak current, which is the peak inductor current, applies to both forward and catch

MOSFETs.

Forward MOSFET

Forward MOSFET selection, like the catch MOSFET, depends on the maximum drain-source

voltage and current as well as maximum gate-source voltage. The main optimization goal is also

to reduce RDS(ON) and QG.

Equation 4-12: Forward MOSFET Maximum Drain-Source Voltage

MOSFET VDSS rating was at least 20% larger than the calculated VDS(MAX) value due to energy

being transferred between the magnetizing inductance and the clamp capacitor causing voltage

“bowing”.

Equation 4-13: Forward MOSFET Maximum RMS Drain Current

22
 Equation 4-14: Total Gate Charge Power Loss for Both Catch and Forward MOSFETs

Active Clamp Capacitor

Active clamp capacitance selection depends on the voltage ripple allowed by the components

connected to SWP. Low clamp capacitance values will generate larger voltage ripple and thus,

larger drain voltage on the primary MOSFET, but will require less magnetizing current.

Equation 4-15: Clamp Capacitance Value

Equation 4-16: Clamp Snubbing Capacitance

Equation 4-17: Clamp Snubbing Resistance

A snubber circuit will limit the peak voltages at the primary MOSFET drain during transients.

23
Mosfet Heatsinking

The packages for the Clamp Mosfet and the Catch Mosfet were TO-220 and required heatsinks

to be chosen for them. The Primary Mosfet and the Forward Mosfets were surface mount and did

not require any additional heatsinking beyond the recommended layout.

Equation 4-18: Calculation of Ambient Temperature without Heatsinking

Equation 4-19: Calculation of Minimum Heatsink Performance

24
Active Clamp Switch Timing

In order to correctly perform ZVS for the primary MOSFET, AOUT, the clamp MOSFET gate

signal, to OUT, the primary MOSFET gate signal, (tAO) and OUT to AOUT (tOA) delays need to

be programmed. Both timings are generated from a single resistor RTAO, which programs the

clamp gate signal falling edge to the primary gate signal rising edge delay, or clamp MOSFET

turn-off to primary MOSFET turn-on.

Equation 4-20: Minimum Timing from AOUT to OUT (Clamp Off, Primary On)

The minimum time between the clamp gate signal to the primary gate signal is determined using

the chosen MOSFETs datasheet signal delay values. Having a larger tAO value allows the main

transformer’s magnetizing current to discharge the primary MOSFET’s drain voltage towards

VIN before turn-on.

Equation 4-21: Switching Timing AOUT to OUT Resistor

The tOA value is automatically generated and checked to make sure the clamp MOSFET does not

turn on until the primary and forward MOSFETs are turned off to prevent shoot-through.

25
 Figure 4-1: Switch Configuration
Source: Adapted from [6, p. 32].

M1 represents the primary MOSFET, M2 the clamp MOSFET, M3 the forward MOSFET, and

M4 the catch MOSFET.

26
Synchronous Rectifier Timing

To synchronize the primary switching to the secondary side synchronous rectification, SOUT,

the control signal to the secondary side, to OUT (tSO), and OUT to SOUT (tOS) need to be

programmed for highest efficiency. Both timings are generated from a two resistors RTAS, which

programs the catch gate signal falling edge to the primary gate signal rising edge delay, or catch

MOSFET turn-off to primary MOSFET turn-on, and RTOS, which programs the primary gate

signal falling edge to the catch gate rising edge delay, or primary MOSFET turn-off to catch

MOSFET turn-on.

Equation 4-22: Minimum Timing from SOUT to OUT (Catch Off, Primary On)

The minimum time between the catch gate signal to the primary gate signal is determined using

the chosen MOSFETs datasheet signal delay values. The catch MOSFET should always be

turned on when the primary MOSFET is turned off for continuous conduction. This results in

very short non-overlap times between the SOUT and OUT gate signals.

Equation 4-23: Minimum Timing from AOUT to OUT (Clamp Off, Primary On)

The tSO delay is defined by the AOUT to OUT and AOUT to SOUT individual timings, so once

the tSO value is found, the tAS value is determined to set the clamp off to primary on delay.

27
 Equation 4-24: Switching Timing AOUT to SOUT Resistor

Equation 4-25: Minimum Timing from OUT to SOUT (Primary Off, Catch On)

Lastly, the tOS is defined, which defines the primary MOSFET turn-off to catch MOSFET turn-

on delay. This timing is programmed through the RTOS resistor value.

Equation 4-26: Switching Timing OUT to SOUT Resistor

Simulation

Simulations were run many times over the course of the design in order to validate design

equations, determine optimal values for parameters not covered by the datasheet, and retrieve

data from hard to calculate expressions. The simulation depicted in Figure x is the final

simulated schematic we used. This schematic was the test bed for all previous simulations. The

schematic runs the converter through 5 load values, 0%, 25%, 50%, 75%, and 100% load, and

and with 3 different input voltage levels, 200V, 316V, and 400V. It should be noted that the

schematic is not representative of the final design due to simulation issues. The first difference is

28
that most of the OC and other non-critical set resistors were not used because simulation errors

sometimes resulted in a triggering of the various protection circuits that we do not expect to

happen in actual operation. The second difference is that many of the mosfets are not

representative of the actual mosfets we used. This is because most of the mosfets that we used

did not have LTSpice models. The most notable of these is the primary mosfet, which has an

auxiliary boost converter inside of it. As a result, the switch timing resistors were not

implemented because the timings would not be representative of the actual system. Third, the

input filter network is not included because our supply is an ideal voltage source in simulation

and thus, would only slow down the simulation time. Finally, the gate driver is not present in the

simulation. A gate driver LTSpice model was found, but the LTSpice solver was not able to

reach convergence despite many simulation setting changes.

From simulation, we were able to confirm that the clamp and snubber network were still

effective at 300 kHz, 500W. Additionally, we determined that a turns ratio of 4 and a primary

inductance of 200 uH was the optimal combination for the main transformer. Deviating from this

optimal point ended up in increased losses, due to the clamp network no longer performing

optimally. Additionally, we found that reducing the current sense resistor on the main mosfet

improved efficiency noticeably and improved operation at higher power levels. Finally, we were

able to determine the average winding currents, which was critical information for determining

the winding losses in the transformer. Components in the housekeeping supply, sync transformer

and optoisolator were found to not significantly affect performance. The peak efficiency of our

best simulation achieved 94.5% efficiency.

29
 Figure 4-2: Final Simulation Schematic
Source: Adapted from [10].

Figure 4-3: Efficiency Plot of Optimal Configuration

30
Transformer Design

The differing power and frequency between the demo board and our chosen design necessitated

the design of a transformer. The transformer core was chosen using the area product method and

the final design was finalized through a MATLAB script (see Appen. A).

The first step in designing a transformer is determining as many overall parameters as

possible. Through simulation, we determined that we would be using a 300 kHz active clamp

switching waveform, 400V max primary voltage, approximately 4:1 turns ratio, and 200 uH of

primary side inductance. From here, the next step is to choose a core. A common method of

selecting a core is through the area product method. The area product method looks at a core’s

cross-sectional core area and its cross-sectional winding area. This product can give a rough

estimate of the total power handling of the core.

Equation 4-27: AP Product Guidance Equation

Source: Adapted from [11].

From here, we chose a core topology and size. We ended up choosing a N87 core material, ETD

core (specifically, the B66358) from TDK. Choosing core topology will not be discussed further

in the report. However, the ETD core was chosen due to the ease of winding and increased

availability compared to PQ and other core types.

31
In the code, the first step is to calculate the skin depth. The skin effect is a phenomena in high

frequency wires, where the flowing current induces an eddy current within the wire, causing the

current density to increase around the edges of the wire, reducing the effective wire gauge. The

skin depth is the depth where 30% of the current is present within that distance from the surface

of the conductor.

Equation 4-28: Skin depth equation

For a copper wire at 300 kHz, the skin depth is approximately 119 micron.

Since we started with a maximum allowable duty cycle, we are able to correlate that duty cycle

to a maximum flux density, which was chosen to be 200mT. This flux density was chosen

because it is the highest flux density that is characterized in the core loss plots for our chosen

core.

32
 Equation 4-29: Calculation of required primary turns

The actual operating flux density is determined by multiplying the max flux density for

the chosen primary turns by the ratio of the operating duty cycle to the maximum duty cycle:

Equation 4-30: Calculation of Operating Flux Density

Now that we know the operating flux density, we can determine the core loss. TDK’s

website has core loss parameters for a given temperature and flux density, in units of mW/kg.

These plots were inserted into the MATLAB script, interpolated, and multiplied by the core mass

to determine loss. Lower flux densities will result in lower loss, which is why it is important to

make the distinction between the maximum allowed duty cycle and the actual operating duty

cycle.

Next, we must determine the core material and gapping required to achieve our desired

primary inductance. In our research, most transformer designs are made without accounting for

the transformer inductance, with the exception of designs for flyback supplies. However, our

simulation indicated that failure to achieve a 200uH primary inductance would lead to decreased

33
efficiency and possible failure to operate. By changing the core material and gapping the core,

we can reduce the effective permeability of the core enough to reduce the primary inductance to

an acceptable value. A figure of calculating the inductance of a winding is the Al value. The Al

value is in units of nH/N^2, where N is the number of turns. Our core datasheet gave us the Al

values for different cores and gaps. From here, we choose the gap that will give us an inductance

that is closest to 200 uH.

Next is to calculate winding losses. From simulation data, we determined that 3.865A

will flow through the primary windings on average during worst case, max power. Multiplying

that current by the turns ratio will give us the secondary winding current. The length of the

primary and secondary windings is determined by multiplying the number of turns by the mean

path length, a parameter given by the bobbin datasheet.

From here there two potential methods for designing the transformer windings. The first

method is to use single wire strands, the second is using Litz wire.

There are two different types of effects that increase resistance in transformer windings:

skin effect and proximity effect. The skin effect is the increase in resistance caused by eddy

currents causing the current to concentrate near the surface of the conductor. The proximity

effect increase resistance in parallel wires due to interaction in their magnetic fields, which

further crowds the current. The skin effect can be easily circumvented by using thinner gauge

wire. On the other hand, the proximity effect can be reduced by interleaving the primary

windings between secondary windings, since the secondary windings have currents that are out

of phase with the primary windings, or by not paralleling wires. Because these are competing

parameters and proximity effect losses are not easily calculable without EM simulation, the

safest option is to use a small gauge wire to reduce skin effect losses and to interleave primary

34
and secondary windings until the all of the winding area is consumed. However, this will lead to

a very difficult to manufacture transformer and the capacitance between the windings will be

greatly increased.

When designing this way, it is important to make sure that the J, current density, between

the primary and secondary windings are equal. This will ensure that the winding losses will be

equal, which is the lowest loss point.

Litz wire is a bundle of electrically insulated wires that are twisted in a special way so

every strand spends and equal amount of distance in every part of the cross-sectional area of the

bundle. This allows Litz wire to have a larger effective diameter wire that bypasses the skin

effect due to the thin wire and also fully use all of the wires in the strands because each wire is

exposed to the same proximity effect forces at some point in the wire. This causes the initial

resistance of the Litz wire to decrease as strand count increases. Because the proximity effect

will increase resistance, the effective resistance of the litz wire will eventually start to increase as

the number of parallel strands are added to the litz wire. Generally, the resistance of Litz wire is

represented by F, which is the scaling factor of Rac / Rdc. Taking the DC resistance of the Litz

wire and multiplying it by F will give you the actual resistance of the Litz wire.

Equation 4-31: Calculation of F for Litz Wire

35
 Equation 4-32: Calculation of Rac

Because Rac will only be minimal over a very specific bundle outer diameter for a given

strand diameter, and because only a limited combinations of Litz wire exist, it is not important to

enforce an equal current density. Instead, it is better to test every combination of litz wires and

interleaving, then picking the lowest loss combination. In our design, it was determined that no

interleaving was required to reach the optimal litz wire solution.

While choosing different litz wire options, it is important to ensure that the total occupied

area fits within the winding area. Winding are occupied is affected by the number of interleaving

layers, the cross sectional area of the wire, and the number of turns. The relationship between the

number of turns and the winding is why the area product equations exist. The larger the cross-

sectional area of the core, the less turns are required for any given flux density, which increases

the amount of winding area available for parallel wires. When using the litz wire, it is important

to recognize that the actual area of the litz wire is does not scale linearly linearly with the number

of strands. Additionally, we must take into account the wasted space between the round cavities

when laying wires together. This was accounted for by increasing the wire area by 25%.

36
 Equation 4-33: Empirical Formula for Calculating Litz Cross-Sectional Area

The MATLAB script takes all of these equations and uses them to find the lowest loss

solution. First, the script creates an array of different possible B. For each different B, we

calculate the core loss, then try every combination of litz wire and interleaving to find the best

winding combination that will fit within the winding area. With this information, we have the

best possible total loss in the transformer for a single B. As we iterate through the different B’s

in the array, we keep track of the losses. When the loop completes, the script selects the B and

winding combination with the lowest loss.

37
Final Design

Figures 4-x through 4-x represent the final design of the DC-DC converter. All component

values calculated through the previous simulation and design equations are used in the

schematic. Additionally, test points, reverse voltage and overcurrent protection, jumpers, and an

enable pin are included. When power is not supplied to the enable pin, it blocks the supply

voltage to the LT3752 undervoltage lockout pin in order to force the converter into shutdown

pin. The jumpers allow the user to force the DC-DC to be off through the same method as the

enable pin, or disable the overcurrent protection. During simulation overcurrent protection was

occasionally an issue, so a disable was added in the event that false overcurrents were occuring.

The reverse voltage protection and overcurrent protection are achieved through a fuse-diode

circuit. Under normal conditions, current flows through the fuse and the diode is reverse biased.

In the event of an overcurrent, the fuse blows, protecting the DC-DC. In the event of a reverse

voltage, the diode becomes forward bias and short circuits the supply through the fuse. This

causes the fuse to blow, disconnecting the circuit from ground.

38
 Figure 4-4: DC-DC Top Schematic Figure 4-5: LT8311 Schematic

Figure 4-6: Forward Active Schematic Figure 4-7: Flyback Schematic

39
 Figure 4-8: LT3752-1 Schematic Figure 4-9: OptoIsolator Schematic

Figure 4-10: Input Fusing and RVP Figure 4-11: DC-DC Disable
Schematic Schematic

40
Figure 4-12: Sync Transformer
Figure 4-13: Gate Driver Schematic
Schematic

41
 5. Hardware Test and Results

The construction of the DC-DC is separated into two parts: the construction of the

transformer, and the main board.

After purchasing the required transformer materials that were chosen in Chapter 4, we

began winding the transformer. The primary windings were the inner layer and the secondary on

the outer layer. As mentioned in Chapter 4, there was no need for interleaving. The inner layer

was used for the primary winding in order to reduce leakage inductance on the primary side.

Additionally, the gapping needed to achieve the correct leakage inductance will cause the inner

layer to experience higher proximity effect resistance. Therefore, it is better to have the higher

current, secondary winding further away from the center leg of the core since the secondary

winding experiences higher currents.

Figure 5-1: Final Constructed Transformer

While winding the transformer, we tried two separate methods for winding the

transformer, which we called low leakage, and crossover. The low leakage winding method

reduces the total leakage inductance while the crossover method lowers the maximum voltage

between adjacent windings. As seen in Table 5-1, there are no significant changes in Rac
42
between the two methods; therefore, we chose to follow the crossover winding method to reduce

the risk of arcing in the transformer.

The primary side inductance matched calculations very well. The chosen 0.2mm gap

gave us a primary side inductance of approximately 454uH, which is close to the predicted

400uH. Because the datasheet did not provide data on gaps formed by combinations of different

gaps, we tested all the different combinations of gaps and found that a 0.4mm total gap gave the

closest primary side inductance to 200uH. On the other hand, Rac measurements deviated

significantly from the calculated data. Most Rac measurements were found to be over 30 times

greater than expected.

Figure 5-2: Low Leakage Winding Depiction

Figure 5-3: Crossover Winding Depiction

43
 Table 5-1: Transformer Design Test Data
Primary L@300kHz
Winding Gap (uH) Rac@300kHz Rdc (mOhm) Winding

.4mm gap 268.8 2.2 Not Measured Lleak attempt

.3mm gap 334 2.436 Not Measured Lleak attempt

.2mm gap 454.6 3.097 Not Measured Lleak attempt

.1mm gap 721 5.3 Not Measured Lleak attempt
0 gap 2.5 123 Not Measured Lleak attempt
.4mm gap 269 2.53 50.93 Crossover
.3mm gap 336.3 2.77 50.93 Crossover
.2mm gap 454.2 3.6 50.93 Crossover
.1mm gap 709 6.8 50.93 Crossover
0 gap 2.49 124.4 50.93 Crossover

Secondary
Winding
.4mm gap 21.09 0.3439 4.065

Calc'd N1/N2 3.571394703

44
 Figure 5-4: Transformer Measured Primary Side AC Resistance

Figure 5-5: Transformer Measured Secondary Side AC Resistance

45
 Figure 5-6: Transformer Turns Ratio Validation

After AC resistance measurements were made, we validated the turns ratio by inputting a

sinusoid and measuring the output sinusoid amplitude as shown in Figure 5-6.

The DC-DC converter main board was laid out in Altium Designer, layout and board

shown in Figures 5-7, 5-8, and 5-9. Decoupling capacitors were used in close proximity to each

IC’s VIN and GND. The TAO, TAS, TOS, TBLNK, IVSEC, and RT resistors were placed close to

each pin of the LT37352-1. A single, closed off AGND plane was used to ground these resistors

to pin 18 of the LT3752-1. Kelvin connections were used to connect the primary current sense

resistor for the LT3752-1 to the ISENSEP and ISENSEN pins. The current sense resistor for the

housekeeping supply also had its ground connection close to PGND, pin 38. The gate driver

traces, HOUT, AOUT, SOUT, and OUT were made as short as possible, bringing the MOSFETs

closer to the LT3752-1. High current traces were expanded and connected on multiple layers

with vias to reduce current path resistance. Additionally, the low voltage and high voltage

portions of the PCB were separated by at least 4 mm to comply with Formula SAE rules [12].

The board was designed for a 1 oz, 4 layer PCB. Once the layout was finished, the PCBs were

46
etched by Bay Area Circuits. To solder the board, homemade stencils were created to decrease

manufacturing time. The base stencil material consisted of two layers of Kapton tape stuck on

either side of a piece of a paper. The Kapton-paper composite was then cut on a laser cutter using

the paste layer from the Altium board layout, forming a makeshift stencil. Solder paste was

applied and all non-through hole components on the top layer were placed and reflowed using a

microscope and toaster oven. The remaining through-hole and bottom layer components were

soldered by hand. For the through-hole MOSFETs, their heat sinks were bolted on using plastic

bolts and nuts with insulating thermal pads.

Figure 5-7: DC-DC Main Board Layout

47
 Figure 5-8: DC-DC 3D Render

Figure 5-9: DC-DC Final Constructed Board

To sufficiently monitor all relevant signals and rails of the converter, multiple

oscilloscopes and differential probes were used to measure and characterize the converter. The

basis of the test setup consisted of monitoring the input and output voltage and current. The final

test setup consisted of two 600V, 1.5A power supplies in parallel, sourcing the converter as

shown Figure 5-10. The HV current path was conducted through an ammeter, and eventually to
48
the DC-DC converter. Additionally, the input voltage of the DC-DC converter was monitored

using a multimeter after the input fuse of the converter to indicate whether or not the converter

had passed its overcurrent limit. The output voltage was monitored using another multimeter,

and the output current was measured using a BK Precision electronic load when the converter

was loaded.

Figure 5-10: DC-DC Multimeter, Supply, and Load Block Diagram

As for the signals and power rails of the converter, three different oscilloscopes were

used and synchronized to measure all the relevant signals. However, the ground prong of each

oscilloscope connected the chassis together and thus, the grounds of each oscilloscope probe

were connected together. Because of the nature of isolated power supplies, signals on the high

voltage and low voltage sides of the converter could not be normally measured by neither the

same oscilloscope, nor multiple oscilloscopes. To measure the signals on both sides of the

converter, the signals on one of the sides needed to be measured differentially, with probes from

all three oscilloscopes grounded to the same point. In this case, all probes from all three

oscilloscopes were grounded to the high voltage ground. This meant that high voltage signals

could be measured normally, while all low voltage signals needed to be measured differentially.
49
The high voltage signals were measured using Rigol 1, as shown in Figure 5-11. In order to use

probes efficiently, only one probes from the Rigol 2 and Keysight oscilloscopes were used to

measure the low voltage ground signal. A Matlab script was used to perform math on the

measured data to subtract the Rigol 2 and Keysight signals from their measured ground signal to

extract the signal waveforms after the fact. Additionally, this Matlab script synchronized the data

measured from each oscilloscope to the same time scale. To accomplish this, each oscilloscope

used the same time division scale and were externally triggered to each other. In total, 7 signals,

(9 including low voltage grounds), were measured, including OUT (primary MOSFET gate

signal), AOUT (clamp MOSFET gate signal), SWP ( primary MOSFET drain voltage and

bottom leg of primary side of the transformer), FSW (forward MOSFET drain voltage and

bottom leg of secondary side of the transformer), CSW (catch MOSFET drain voltage and top

leg of secondary side of the transformer), FG (forward MOSFET gate signal), CG (catch

MOSFET gate signal), and HV IN (high voltage input and top leg of the primary side of the

transformer).

Figure 5-11: DC-DC Oscilloscope Measurement Point Setup

50
 Due to multiple failures of the DC-DC converter, we were unable to properly characterize

the operation of the DC-DC. At the time of writing, the DC-DC converter suffered from excess

ringing that lead to the failure of the converter.

Table 5-2: Efficiency Test Results

Input Voltage Load Efficiency Output Voltage
200V 0% N/A 9V
200V 25% 93% 7V

Figure 5-12: LT3752-1 and LT8311 Controller Signal Waveforms (Ignore Blue
Waveforms)
Source: Adapted from [7, p. 14]

51
 According to Figure 5-12 from the datasheet, SWP and FSW mirror each other with FSW

being stepped down by the turns ratio. This is accomplished by transformer action on opposing

legs of the transformer. Additionally, CSW, the top leg of the secondary side, resembles a square

wave with ringing at the rising edge.

Table 5-3: DC1929A Demo Board Waveforms1

Figure 5-13: DC1929A OUT and SWP Figure 5-14: DC1929A FG and FSW

Figure 5-15: DC1929A CG and CSW

1
All signals tested using 316V input at 2A load.
52
 The DC1929A Demo Board was tested to serve as example control signals of how the

DC-DC converter should look. As shown in Figure 5-14, the FSW waveform closely resembles a

square wave with ringing on the rising and falling edge. The voltage steps shown in the datasheet

waveforms are small, and not clearly seen in the SWP and FSW waveforms in Figures 5-13 and

5-14. However, both waveforms roughly resemble a square wave.

Figure 5-16: 8A load test with ringing

Figure 5-16 indicates the major problem that occurred while testing the DC-DC

converter. As seen in the 3rd from top plot, the primary and secondary voltage of the transformer

do not match. This does not match expected transformer operation. Additionally, the CSW and

FSW waveforms have multiple voltage spikes, likely from ringing. This is the result of the

mismatched transformer operation, as the SWP signal is very close to a square wave while the

FSW signal is ringing.

53
 6. Conclusion

This project entails the design and construction of the HV (High Voltage) to LV (Low

Voltage) DC-DC converter for the Formula SAE Student Design Competition (FSAE) electric

vehicle car. The final DC-DC converter was able to demonstrate successfully a high frequency

power transformer design, which further provides a useful framework to assist in the speed of

development of future DC-DC projects. There were some issues encountered during the testing

of the converter, which include meeting the load specifications due to excess ringing on the

primary and secondary side leading to eventual failure of the power MOSFETs during light load

operation. Additional learning experience gained from the project is the knowledge needed for

proper heatsink selection and the balancing of switching and conduction losses to improve

efficiency. Finally, we also developed a measurement system to allow for the synchronization of

multiple oscilloscopes for differential probe measurements.

Future improvements falls into two categories: improvements that could help the DC-DC

meet specifications, and improvements that should be implemented after meeting specifications.

Improvements focused on helping the DC-DC meet specifications are centered around

fixing the ringing, as we believe it is the primary source of component failure and heating in the

DC-DC. The intensity of the ringing is potentially due to the measurement setup. Using high

voltage differential probes would significantly decrease the complexity of the measurement setup

by reducing the amount of probes required. Additionally, the parasitic capacitance of the

oscilloscope probes should also be reduced by using differential probes. Another source of

resonance is through the input capacitance of the secondary side MOSFETs.

54
 The frequency of the ringing also decreased after increasing the output inductor value, so

having a larger output inductor may help as well. In the Demo Board results, the voltage at the

top leg of the transformer resembled the top half of a sine wave with significant ringing. This

may indicate that a higher frequency resonance is superimposed on a lower frequency resonance,

with the higher frequency resonance being switching or leakage inductance related, and the

lower frequency resonance being related to the resonance between the output inductor and output

capacitance.

As another improvement, the transformer needs more testing and validation. Because we

lacked a proper test fixture, we were unable to test our transformer for leakage inductance, power

handling and load regulation capabilities. Transformer characterization is important because high

secondary side leakage inductance could be the reason why the ringing is only present on the

secondary side. Designing an H-bridge or fixed timing active clamp forward converter would

help characterizing the source of the ringing.

Improvements focused on improving a functional DC-DC center on performance

improvements and quality of life changes. One improvement includes reducing the overall size

of the converter in the width and length dimensions. These improvements will mainly decrease

the power path length, reducing resistance, and benefit packaging for future Formula SAE

Electric vehicles.

In terms of improving our ability to debug the converter, the clamp MOSFET should be

moved or become a surface mount package, which would assist the removal of the transformer

core during board reworks. Additionally, the clamp MOSFET heatsink was oversized and can be

reduced. Several nets of converter were missing that would benefit debugging should be added.

Some of these nets include the HV net before the input filter, the output voltage, and primary

55
MOSFET pins. The output connector system bolting wire lugs to M3 screws was tedious to

connect and remove and should be altered to improve removal.

In order to further reduce size, water cooling could be included to decrease the required

cooling effort. This would allow designers to use smaller switching IC's, which should improve

efficiency, as the switching energy will decrease. This would also decrease the total required

mass for heatsinks.

Another potential improvement is the addition of a complete microcontroller system. This

would significantly improve the tunability of the design as well as the amount of telemetry that

could be sent back. In this project, voltage and current sensing were not able to be easily

transmitted to an external processor. Additionally, the output voltage could only be adjusted

manually. Both of these operations could be accomplished using a microcontroller system.

56
 7. References

[1] L. McDonald, “US EV Sales Surpass 2% In 2018 — 9 EV Sales Charts,” CleanTechnica,

13-Jan-2019. [Online]. Available: https://cleantechnica.com/2019/01/12/us-ev-sales-surpass-
2-for-2018-8-more-sales-charts/. [Accessed: 01-Feb-2019].
[2] “EV-1 White Paper,” 11-Sep-2002. [Online]. Available:
https://web.archive.org/web/20090726034344/http://www.cleanup-gm.com/ev1.html.
[Accessed: 01-Feb-2019].
[3] “Model 3 | Tesla.” [Online]. Available: https://www.tesla.com/model3. [Accessed: 01-Feb-
2019].
[4] N. Sridhar, “Driving the Future of HEV/EV With High-Voltage Solutions.” [Online].
Available: http://www.ti.com/lit/wp/slyy052b/slyy052b.pdf. [Accessed: 18-Feb-2019].
[5] “DC-DC converter for Hybrid Electric Vehicle and EV,” 2013.
[6] “LT3752/LT3752-1 Active Clamp Synchronous Forward Controllers with Internal
Housekeeping Controller.” [Online]. Available: https://www.analog.com/media/en/technical-
documentation/data-sheets/3752fb.pdf. [Accessed: 10-Jun-2019].
[7] “LT8311 Synchronous Rectifier Controller with Opto-Coupler Driver for Forward
Converters.” [Online]. Available: https://www.analog.com/media/en/technical-
documentation/data-sheets/8311f.pdf. [Accessed: 10-Jun-2019].
[8] “DEMO MANUAL DC1929A LT3752-1/LT8311 200W Active Clamp Forward Converter
with Input Voltage 150V to 400V.” [Online]. Available:
https://www.analog.com/media/en/dsp-documentation/evaluation-kit-
manuals/dc1929afa.pdf. [Accessed: 10-Jun-2019].
[9] “LMG341xR050 600-V 50-mΩ Integrated GaN power stage with overcurrent protection.”
[Online]. Available: http://www.ti.com/lit/ds/symlink/lmg3410r050.pdf. [Accessed: 10-Jun-
2019].
[10] “LT3752 LT3752-1 Datasheet and Product Info | Analog Devices.” [Online]. Available:
https://www.analog.com/en/products/lt3752.html#product-tools. [Accessed: 10-Jun-2019].
[11] “Power Transformer Design.” [Online]. Available:
https://www.ti.com/lit/ml/slup126/slup126.pdf. [Accessed: 10-Jun-2019].
[12] “FSAE Rules.” [Online]. Available:
https://www.fsaeonline.com/page.aspx?pageid=e179e647-cb8c-4ab0-860c-ec69aae080a3.
[Accessed: 10-Jun-2019].

57
 8. Appendix

Code for Transformer Design

Transformer.m

%% Initializations
clearvars
load('wireSizes.mat'); % Loading wire gauge tables
load('litzOptions.mat'); % Loadings Litz wires of form (Nstrand,
Gauge)
% Litz wires available from
https://www.mikeselectronicparts.com/litz-wire/
Bdes = (0.005:0.001:.2)'; % Desired Flux Swing (Arbitrary Starting
Value)
K = 0.014; % Compensation Factor

%% Constants
rho_cu = 1.68e-8; % Copper Resistivity (ohm m)
%% Desired Values
Lpri_des = 200e-6; % Desired Primary Inductance
TR = 4; % Desired Turns Ratio
Jdes = 2.2; % Desired Current Density (A/mm^2)
eta = 0.3; % Expected Efficiency
DutyLim = 0.75; % Max Duty Cycle
DutyOp = 0.3; % Operating Duty Cycle
Ninter = 4; % Max interleaving layers
T = TransformerModel2(500, 300e3, DutyLim, DutyOp, 400,200, 0.75);
AP = (T.Po./(K.*Bdes.*T.ft)).^(3/4); % Estimated Area Product
Required (cm^4)
%% Core Properties
% B66358
T.Core = Core(0.93, 70.4, 76, 5350, 97, 19.4, 52.8, 28,...
[1470 1610 1670 457 281 148 91], 0.3, ...
[2000 2200 2250 621 383 201 124],...
[0.025, 0.050, 0.100, 0.200;
0.00333, 0.014433 0.084536 0.500732],...
[0 0 0 .1 .2 .5 1]);

58
T.wPri = LitzWire();
T.wSec = LitzWire();
%% General Design Calculations
%T = V*s/m^2
Np1 = ceil(T.VinMax .* T.Dlim.* 1e4
./(2.*Bdes.*T.ft.*T.Core.Ae*0.01));
Bmact = T.VinMax .* T.Dlim.* 1e4 ./(2 .* Np1 .* T.ft .*
T.Core.Ae*0.01); % Actual Max Allowed B
Bact = Bmact.*T.Ddes./T.Dlim; %Getting Actual used B
Np2 = ceil(Np1/TR);
TR_act = Np1./Np2;
Al_req = Lpri_des ./ Np1.^2 *10^9;
% Determining the correct Al for the job
gChosen = zeros(1,length(Al_req));
AlChosen = zeros(1,length(Al_req));
LChosen = zeros(1,length(Al_req));
for i=1:length(Bdes)
Altemp = T.Core.Al(T.Core.Al > 1.2.*Al_req(i));
muiChosen = T.Core.mui(length(Altemp));
gChosen(i) = T.Core.g(length(Altemp));
AlChosen(i)= Altemp(end);
LChosen(i) = AlChosen(i) * Np1(i)^2*1e-9;
end
clear Altemp
T.Ipri = 3.865; % Measured Values
Isec = T.Ipri*TR_act;
LenPri = Np1 * T.Core.LN * 0.001; % Primary Wire
Length in m
LenSec = Np2 * T.Core.LN * 0.001;
Pcore = T.Core.getCL(Bact);
A = T;
A = repmat(A,1,length(Bdes));
for Bindex = 1:length(Bdes)
T.Np1 = Np1(Bindex);
T.Np2 = Np2(Bindex);
T.Isec = Isec(Bindex);
T.Bop = Bact(i);
T.Lpri = LChosen(i);
T.Core.gOpt = gChosen(i);
59
 T.Core.AlOpt = AlChosen(i);
T.TR = TR_act(i);
T.Pcore = Pcore(i);
T.Bmax = Bmact(i);
A(Bindex) = T.findLitz(LitzOptions, LenPri(Bindex),
LenSec(Bindex), ...
Ninter);
end
% Finding Optimal Operating Point
[val, i] = max([A.eta]);
T = A(i);
%% Plots
figure(1)
plot(Bdes,[A.eta])
xlabel('Bmax')
ylabel('Efficiency(%)')
title('Efficiency vs. Max B')
TransformerModel.m

classdef TransformerModel2
%TRANSFORMER Holds all of the relevant information to the final
%transformer design
% Detailed explanation goes here

properties
%% Objects
Core % Model of Transformer Core
wPri % Primary Wire Model
wSec % Secondary Wire Model
%% Vars
AwPri % Area used by primary windings
AwSec % Area used by secondary windings
Bmax % Max B used
Bop % Nominal B used
Dlim % Max Possible Duty Cycle
Ddes % Desired Duty Cycle
eta % Efficiency
ff % Winding Fill Factor

60
 ft % Transformer Operating Frequency
Ipri % Primary side peak current
Isec % Secondary side peak current
JPri % Current density of the Primary
JSec % Current Density of the Secondary
Lpri % Primary side inductance
Ninterleave % Number of interleaving section PER PHASE
Np1 % Primary Number of Turns
Np2 % Secondary Number of Turns
Po % Output Power in Watts
Pcu % Power Dissipated in windings
Pcore % Power Dissipated in the core
PcuPri % Power Dissipated in the primary windings
PcuSec % Power Dissipated in the secondary windings
Ptot % Total Dissipated Power
sd % Skin Depth
TR % Turns ratio
VinMax % Max Primary Voltage
VinMin % Minimum Primary Voltage
end

methods
function obj = TransformerModel2(Po, ft, Dlim, Ddes, VinMax,
VinMin,...
ff)
%UNTITLED3 Construct an instance of this class
% Detailed explanation goes here
obj.Po = Po;
obj.ft = ft;
obj.sd = getSkinDepth(obj.ft);
obj.Dlim = Dlim;
obj.Ddes = Ddes;
obj.VinMax = VinMax;
obj.VinMin = VinMin;
obj.ff = ff;
end
function obj = findLitz(obj,LitzOptions,lenPri,lenSec, N)
litzWires = LitzWire(LitzOptions, [obj.Np1, obj.Np2],
obj.Core.Ww, obj.sd, [lenPri,lenSec])';
61
 Ninterleaves = (1:1:N); %Number of interleaved blocks
per side (4 primaries)
AwPriM =
repmat(reshape(.01:.01:(obj.Core.Aw*0.8*obj.ff/obj.Np1),1,1,[])...
,1,length(Ninterleaves),1)./Ninterleaves; %List of
optional Aw's
AwSecM = (obj.Core.Aw*0.8*obj.ff -
AwPriM.*Ninterleaves.*obj.Np1)/obj.Np2./Ninterleaves;
RPriAC = repmat(litzWires.getRac(1)'./Ninterleaves, 1, 1,
length(AwPriM));
RsecAC = repmat(litzWires.getRac(2)'./Ninterleaves, 1, 1,
length(AwSecM));
RPriAC(AwPriM - ...
repmat(Ninterleaves.*[litzWires.Ao]', 1, 1,
length(AwPriM)) < 0) = NaN;
RsecAC(AwSecM - ...
repmat(Ninterleaves.*[litzWires.Ao]', 1, 1,
length(AwSecM)) < 0) = NaN;
PpriAC = obj.Ipri^2.*RPriAC;
PsecAC = obj.Isec^2.*RsecAC;
[obj.Pcu, indexAw] = min(min(min(PsecAC)+min(PpriAC)));
[~, indexInter] = min(min(PsecAC)+min(PpriAC)); %
Determining the optimal configuration
[obj.PcuSec, indexSec] =
min(PsecAC(:,indexInter(indexAw),indexAw));
[obj.PcuPri, indexPri] =
min(PpriAC(:,indexInter(indexAw),indexAw));
obj.Ninterleave = indexInter(indexAw);
obj.AwPri = AwPriM(indexAw);
obj.AwSec = AwSecM(indexAw);
obj.wPri = litzWires(indexPri);
obj.wSec = litzWires(indexSec);
obj.JPri = obj.Ipri/obj.AwPri;
obj.JSec = obj.Isec/obj.AwSec;
obj.Ptot = obj.Pcore + obj.Pcu;
obj.eta = (obj.Po-obj.Ptot) / obj.Po*100;
end
end
end
62
LitzWire.m

classdef LitzWire < Wire

%LITZWIRE Class that holds information of the Litz Wire to be
used
% Detailed explanation goes here

properties
Ao % Total Area (not copper area)
Aw % Total wire Area
Fr % Rac/Rdc (will have two values(one for primary, one
for secondary)
Nstrand % Number of strands in the Litz Wire
OD % Outer Diameter of wire + 10% safety factor
Rac % AC Resistance
Rdc % DC Resistance of bundle
end

methods
function obj = LitzWire(objArray, Ns,b, sd, len)
%LITZWIRE Construct an instance of this class
% Detailed explanation goes here
if nargin ~= 0
m = size(objArray,1);
obj(m) = obj; % Create an array of objects
for i=1:m
obj(i).Nstrand = objArray(i,1);
obj(i).awg = objArray(i,2);
obj(i).d = objArray(i,3);
obj(i).a = objArray(i,4);
obj(i).len = len;
obj(i).R = obj(i).calcR();
obj(i).Fr = obj(i).calcFr(Ns, b, sd);
obj(i).Rdc = obj(i).calcRdc();
obj(i).Rac = obj(i).calcRac();
obj(i).OD = 1.1*(1.28 *
sqrt(obj(i).Nstrand)*obj(i).d);
obj(i).Ao = pi*(obj(i).OD/2)^2;
obj(i).Aw = obj(i).Nstrand * obj(i).a;

63
 end
end
end
function Fr = calcFr(obj,Ns,b, sd)
Fr = 1+((pi.*obj.Nstrand.*Ns).^2.*(obj.d).^6)...
./(192*(sd.*1000).^4.*b^2);
end
function Rdc = calcRdc(obj)
Rdc = obj.R./obj.Nstrand;
end
function Rac = calcRac(obj)
Rac = obj.Rdc .* obj.Fr;
end
function Rdc = getRdc(obj, index)
Rdc = zeros(1,length(obj));
for i = 1:length(obj)
Rdc(i) = obj(i).Rdc(index);
end
end
function Rac = getRac(obj,index)
Rac = zeros(1,length(obj));
for i = 1:length(obj)
Rac(i) = obj(i).Rac(index);
end
end

end
end

Wire.m

classdef Wire
%WIRE Summarizes property of Wire
% Detailed explanation goes here

properties
a % Area of wire (mm^2)
awg % Wire Gauge number
64
 d % Diameter of wire (mm)
len % Required Length of wire for T
R % Single Wire Resistance
rho_cu = 1.68e-8; % Copper Resistivity (ohm m)

end

methods
function obj = Wire(awg, d, a)
%WIRE Construct an instance of this class
% Detailed explanation goes here
if nargin ~= 0
obj.awg = awg;
obj.d = d;
obj.a = a;
end

end
function R = calcR(obj)
R = (obj.rho_cu .* obj.len ./ (obj.a*0.001^2));
end

end
end

getSkinDepth.m

function [skinDepth] = getSkinDepth(f)

%GETSKINDEPTH Returns the skip depth, assuming the material is made
out of
%copper
% Detailed explanation goes here
rho = 1.678e-8; % resistivity of copper (ohm *m)
mur = 0.999991; % relative permeability
mu0 = 4*pi*1e-7; % free space permeability
skinDepth = sqrt(rho / (pi*f*mur*mu0));
end

65
Core.m

classdef Core
%CORE Summary of this class goes here
% Detailed explanation goes here

properties
CF % Core Factor (mm^-1)
Le % Effective magnetic path length (mm)
Ae % Effective Cross-Sectional Area (mm^2)
Ve % Effective Core Volume (mm^3)
Aw % Winding Area
Ww % Minimum winding width
LN % Average Turn Length (mm)
m % Mass (g/set)
AP % Area Product
mui % Initial Permeability
Bsat % Saturation Magnetic Flux Density (Estimate)
Al % AL-Value @ 1kHz, 0.5mA(nH/N^2)
AlOpt % Optimal Al
CL % Core Loss @ 80C, 300kHz, W/g
g % Air gap in
gOpt % Optimal Air Gap
end

methods
function obj = Core(CF, Le, Ae, Ve, Aw, Ww, LN, m, mui, ...
Bsat, Al, CL, g)
%CORE Construct an instance of this class
% Detailed explanation goes here
obj.CF = CF;
obj.Le = Le;
obj.Ae = Ae;
obj.Ve = Ve;
obj.Aw = Aw;
obj.Ww = Ww;
obj.LN = LN;

66
 obj.m = m;
obj.AP = getAP(Ae,Aw);
obj.mui= mui;
obj.Bsat = Bsat;
obj.Al = Al;
obj.CL = CL;
obj.g = g;

end

function Pcore = getCL(obj,Bact)

%METHOD1 Summary of this method goes here
% Detailed explanation goes here
Pcore = obj.m * interp1(obj.CL(1,:)',obj.CL(2,:)', Bact);
end
end
end

getAP.m

function [AP] = getAP(Ae,Aw)

%getAP Returns Area Product in cm^4 from mm^2 inputs
% Detailed explanation goes here
AP = (Ae*.01*Aw*.01);
end

getCL.m

function CL = getCL(CL_table,B)
%GETCL Returns the core loss for a given B using interpolation
% Detailed explanation goes here
CL = interp1(CL_table(1,:)',CL_table(2,:)', B);
end

Timeline of Tasks and Milestones

67
 Figure 8-1: Winter Quarter Timeline

Figure 8-2: Spring Quarter Timeline

Bill of Materials

The final price of a single board (purchased in bulk quantity of 5 boards) is

$66.79

68
 Table 8-1: Final Board Bill of Materials

Manufacturer Part Reference

Manufacturer Customer Reference Quantity Unit Price Extended Price Total Price
Number Designator

FDP083N15A-F102 ON Semiconductor Catch Mosfet x2 M3 1 4.32 $4.32 $66.79

2.5A Gate Driver Optical Coupling

ACPL-W346-500E Broadcom Limited U2 1 5.05 $5.05
5000Vrms 1 Cha

C4532X7T2J304K2 Capacitor MLCC 0.3uf 630V CAP

TDK Corporation C27, C28, C29, C30 4 1.252 $5.01
50KA 1812/4532

IHLP6767GZER1R5 1.5µH Shielded Molded Inductor 42A

Vishay Dale L2 1 5.87 $5.87
M01 1.88mOhm Max

Panasonic
Capacitor Aluminum Hybrid Radial 330uf
EEH-ZA1E221P Electronic C5 1 2.7 $2.70
25V CAP E
Components

Wurth Electronics
7.44E+08 FIXED IND 4.7UH 6A 19.5 MOHM SMD L1 1 3.6 $3.60
Inc.

IPP60R190P6XKSA Infineon
Clamp Transistor M5 1 3.23 $3.23
1 Technologies

IXYS Integrated Solid State Relay SPST-NO (1 Form A) 4-

CPC1393GRTR U6 1 2.63 $2.63
Circuits Division SMD (0.30

CL10B474KA8NFN Samsung Electro- Capacitor MLCC 0.47uf 25V CAP

C18 1 0.073 $0.07
C Mechanics 0603/1608

Renesas
PS2802-1-A PC817B U5 1 1.52 $1.52
Electronics America

C3216X7R1E106K1 Capacitor MLCC 10uf 25V CAP

TDK Corporation C6, C7, C8, C9 4 0.566 $2.26
60AB 1206/3216

B66359B1013T001 EPCOS (TDK) BOBBIN COIL FORMER ETD 29X16X10 T2 1 1.84 $1.84

69
 Toshiba
TPHR6503PL,L1Q Semiconductor and Forward Mosfet M2 1 1.82 $1.82
Storage

IPN80R750P7ATM Infineon
AUX Transistor M1 1 1.64 $1.64
A1 Technologies

EKXJ501ELL100MJ 10µF 500V Aluminum Electrolytic

United Chemi-Con C26 1 1.03 $1.03
30S Capacitors Radia

HRG3216P-5600-D-
Susumu RES 560 OHM 1% 1W 1206 R2 1 0.97 $0.97
T1

TLR2H15DR001FT TE Connectivity
RES 0.001 OHM 1% 1.5W 2010 R3 1 0.97 $0.97
DG Passive Product

Supplier Device Package ETD 29 x 16 x

B66359S2000X000 EPCOS (TDK) T2 1 0.59 $0.59
10

Amphenol Piher
PT6WV-471A2020 TRIMMER 470 OHM LINEAR R36 1 0.81 $0.81
Sensing Systems

Johanson Capacitor MLCC 11pf 250V CAP

251R15S110GV4E C3 1 0.73 $0.73
Technology Inc. 0805/2012

Rohm
PMR18EZPFU8L00 RES 0.008 OHM 1% 1W 1206 R5 1 0.73 $0.73
Semiconductor

SMC Diode
SD560BTR DIODE GEN PURP 600V 5A SMB D7 1 0.51 $0.51
Solutions

C1210C473KBRAC Capacitor MLCC 50.047uf 630V CAP

KEMET C14 1 0.49 $0.49
7800 1210/3225

BAT46WJ,115 Nexperia USA Inc. BAT46WJ D1, D2 2 0.39 $0.78

TMK316BBJ226ML- Capacitor MLCC 22uf 25V CAP

Taiyo Yuden C2 1 0.58 $0.58
T 1206/3216

Uncoated N87 Ferrite Core ETD Type

B66358G0200X187 EPCOS (TDK) T2 2 0.99 $1.98
30.60mm Lengt

70
GRM21BR61E226 Murata Electronics Capacitor MLCC 22uf 25V CAP
C16 1 0.49 $0.49
ME44L North America 0805/2012

CL10B224KB8VPN Samsung Electro- Capacitor MLCC 0.22uf 50V CAP

C11, C15 2 0.21 $0.42
C Mechanics 0603/1608

OSRAM Opto
LH R974-LP-1 LED RED DIFFUSED 0805 SMD D5 1 0.33 $0.33
Semiconductors Inc.

CL21A475KAQNNN Samsung Electro- Capacitor MLCC 4.7uf 25V CAP

C1, C17, C21 3 0.129 $0.39
E Mechanics 0805/2012

CL10X106MO8NRN Samsung Electro- Capacitor MLCC 10uf 16V CAP

C10 1 0.29 $0.29
C Mechanics 0603/1608

SR0805JR-7W10RL Yageo Resistor 10 0805 1/4W 10% R16 1 0.27 $0.27

DIODE GEN PURP 100V 300MA

1N4148W-7-F Diodes Incorporated D4, D6 2 0.16 $0.32
SOD123

SMC Diode
RS1MTR DIODE GEN PURP 1KV 1A SMA D3 1 0.17 $0.17
Solutions

RNCP0805FTD1K0 Stackpole
Resistor 1.0k 0805 1/4W 5% R25 1 0.072 $0.07
0 Electronics Inc

CC0805JRNPOAB Capacitor MLCC 68pf 200V CAP

Yageo C4 1 0.18 $0.18
N680 0805/2012

CC0603JRNPO8BN Capacitor MLCC 220pf 25V CAP

Yageo C25 1 0.069 $0.07
221 0603/1608

Capacitor MLCC 0.022uf 16V CAP

TMK107B7223KA-T Taiyo Yuden C22 1 0.068 $0.07
0603/1608

Panasonic
ERJ-3EKF4990V Electronic Resistor 499 0603 1/10W 1% R14 1 0.066 $0.07
Components

CL10B105KA8NNN Samsung Electro- Capacitor MLCC 1.0uf 25V CAP

C24 1 0.06 $0.06
C Mechanics 0603/1608

71
 Rohm
MCR18ERTFL1R30 RES 1.3 OHM 1% 1/4W 1206 R6 1 0.1 $0.10
Semiconductor

RNCP0805FTD4R7 Stackpole
Resistor 4.7 0805 1/4W 1% R7 1 0.1 $0.10
0 Electronics Inc

CRM0805-FX-
Bourns Inc. Resistor 51 0805 1/4W 1% R41 1 0.1 $0.10
51R0ELF

Stackpole
RMCF1206FT715K Resistor 715k 1206 1/4W 1% R15 1 0.036 $0.04
Electronics Inc

TE Connectivity
CRGCQ0603F150R Resistor 150 0603 1/10W 5% R10 1 0.031 $0.03
Passive Product

TE Connectivity
CRGCQ0603F560R Resistor 560 0603 1/10W 1% R40, R43 2 0.031 $0.06
Passive Product

CL10B104KA8NNN Samsung Electro- Capacitor MLCC 0.10uf 25V CAP

C19 1 0.029 $0.03
C Mechanics 0603/1608

Wurth Electronics Capacitor MLCC 68pf 25V CAP

8.85E+11 C20 1 0.027 $0.03
Inc. 0603/1608

Wurth Electronics Capacitor MLCC 0.033uf 16V CAP

8.85E+11 C23 1 0.026 $0.03
Inc. 0603/1608

Wurth Electronics Capacitor MLCC 470pf 50V CAP

8.85E+11 C13 1 0.023 $0.02
Inc. 0603/1608

TE Connectivity
CRGCQ0603J220R Resistor 220 0603 1/10W 5% R9 1 0.022 $0.02
Passive Product

Stackpole
RMCF0603FT360R Resistor 360 0603 1/10W 1% R32, R33, R37 3 0.017 $0.05
Electronics Inc

Stackpole
RMCF0603FT806R Resistor 806 0603 1/10W 1% R12 1 0.017 $0.02
Electronics Inc

Stackpole
RMCF0603FT820R Resistor 820 0603 1/10W 1% R35 1 0.017 $0.02
Electronics Inc

72
 Stackpole
RMCF0603FG1M00 Resistor 1.0M 0603 1/10W 1% R17 1 0.017 $0.02
Electronics Inc

Stackpole
RMCF0603FT1K50 Resistor 1.5k 0603 1/10W 1% R20, R42 2 0.017 $0.03
Electronics Inc

Stackpole
RMCF0603FT100K Resistor 100k 0603 1/10W 1% R22, R23 2 0.017 $0.03
Electronics Inc

Stackpole
RMCF0603FT10K0 Resistor 10k 0603 1/10W 1% R11, R34, R38, R8 4 0.017 $0.07
Electronics Inc

Stackpole
RMCF0603FT15K0 Resistor 15k 0603 1/10W 1% R30 1 0.017 $0.02
Electronics Inc

Stackpole
RMCF0603FT2K80 Resistor 2.8k 0603 1/10W 1% R13 1 0.017 $0.02
Electronics Inc

Stackpole
RMCF0603FT20K0 Resistor 20k 0603 1/10W 1% R18, R19 2 0.017 $0.03
Electronics Inc

Stackpole
RMCF0603FT24K9 Resistor 24.9k 0603 1/10W 1% R26 1 0.017 $0.02
Electronics Inc

Stackpole
RMCF0603FT3K00 Resistor 3.0k 0603 1/10W 1% R24 1 0.017 $0.02
Electronics Inc

Stackpole
RMCF0603FT36K5 Resistor 36.5k 0603 1/10W 1% R21 1 0.017 $0.02
Electronics Inc

Stackpole
RMCF0603FT39K0 Resistor 39k 0603 1/10W 1% R4 1 0.017 $0.02
Electronics Inc

Stackpole
RMCF0603FT47K0 Resistor 47k 0603 1/10W 1% R39 1 0.017 $0.02
Electronics Inc

Stackpole
RMCF0603FT51K0 Resistor 51k 0603 1/10W 1% R28 1 0.017 $0.02
Electronics Inc

Stackpole
RMCF0603FT68K0 Resistor 68k 0603 1/10W 1% R27 1 0.017 $0.02
Electronics Inc

73
 Stackpole
RMCF0603FT7K50 Resistor 7.5k 0603 1/10W 1% R31 1 0.017 $0.02
Electronics Inc

Stackpole
RMCF0603FT76K8 Resistor 76.8k 0603 1/10 1% R29 1 0.017 $0.02
Electronics Inc

GRM188R61E225K Murata Electronics Capacitor MLCC 2.2uf 25V CAP

C12 1 0.17 $0.17
A12D North America 0603/1608

Assmann WSW
V6560X 1 na 1 0.67 $0.67
Components

Aavid, Thermal
578622B03200G Division of Boyd 1 na 1 1.4 $1.40
Corporation

Essentra
NMS-306 1 na 3 0.171 $0.51
Components

Essentra
HN-M3-79 1 na 3 0.167 $0.50
Components

Stackpole 2.5 mOhms ±1% 0.25W, 1/4W Chip

CSRF0603FT2L50 R1 1 0.63 $0.63
Electronics Inc Resistor 0603 (16

Omron Electronics
XG8T-0231 W1,W2 2 0.19 $0.38
Inc-EMC Div

Sullins Connector
QPC02SXGN-RC W1,W2 2 0.1 $0.20
Solutions

Zener Diode 14V 500mW ±5% Surface

DDZ9701-7 Diodes Incorporated D8 1 Sample $0.00
Mount SOD-123

FUSE BOARD MNT 3.15A 600VDC

04853.15DR LittelFuse F1 1 Sample $0.00
2SMD

Wurth Electronics 10µH Shielded Wirewound Inductor

74404020100 L3 1 Sample $0.00
Inc. 500mA 850mOhm Nonstandard

Wurth Electronics Auxilliary Transformer for DC-DC Senior

750817020 T1 1 Sample $0.00
Inc. Project

74
 Pulse Electronics
PE-68386NLT PULSE XFMR 1:1 785UH T3 1 Sample $0.00
Power

Keystone
5029 Test point TP1 Thru TP35 32 Sample $0.00
Electronics

XLMG3410R050R N-Channel 600-V 50mOhm GaN with

Texas Instruments U1 1 Sample $0.00
WHT integrated driver and protection

Forward Converter Regulator Positive,

Isolation Capable Output Step-Up/Step-
LT3752FE-1 Linear Technology U3 1 Sample $0.00
Down DC-DC Controller IC 6.5-100V
Input

Power Supply Controller Secondary-Side

LT8311FE Linear Technology U4 1 Sample $0.00
Controller

43650-0402 Molex X1 1 Sample $0.00

Wurth Electronics
7466113 1 Pin Shank Terminal M3 Surface Mount X2, X3 2 Sample $0.00
Inc.

Mike's Electronics
N/A 6604/44 Litz Wire T2 1.558656 1.294 $2.02
Parts

Mike's Electronics
420/46 Litz Wire T2 5.715072 0.764 $4.37
Parts

75
Analysis of Senior Project Design

a) Summary of Functional Requirements

This project will accept a 200-420VDC input from the Cal Poly Racing, Formula

SAE Electric Vehicle's battery pack and output 12-15VDC at loads up to 500W.

The output voltage will be programmable and telemetry data reporting the status

of the DC-DC's performance will be outputted. Many shutdown conditions are

possible for the DC-DC. Overcurrent, overvoltage, and reverse voltage protection

circuits will be included on the input and output in order to shut the device down.

Additionally, a discrete shutdown/enable input will be included in order to disable

the device when necessary.

b) Primary Constraints

The primary constraint for the design of the DC-DC is the Formula SAE

Rulebook. Nearly all safety requirements for the DC-DC are derived the Formula

SAE Rulebook. Failure to abide by these constraints will result in a failure to be

useful for our largest market. Furthermore, failure to follow these rules will also

significantly degrade the general safety of the product. Additionally, the system

must be able to operate under harsh environmental conditions, such as rain and

high temperature. Finally, the DC-DC must be small and light in order to be

competitive in the automotive industry.

76
c) Economic

What economic impacts result?

i) Human Capital

This device will create new jobs in engineering, manufacturing, sales and

technical support.

ii) Financial Capital

This product will not create a lot of business for suppliers because small-

series EV manufacturers and Formula SAE teams represent a very small

portion of the total EV market. However, this product will assist the

customer due to its $40-$200 price improvement.

iii) Natural Capital

Aluminum sheet or plastic will be required to create the enclosure of the

DC-DC. FR-4 material and a board fab house will be needed to create the

main board. Finally, the stock from various IC and passive companies are

needed to populate the board. Significant amounts of water is used in to

create ICs.

When and where do costs and benefits accrue throughout the project lifecycle?

Costs to the senior project team includes, design costs, manufacturing

costs, testing costs, and material costs. Costs to the customer include the upfront

cost, repackaging costs, harness integration cost, and the code costs. These will all

occur at the initial implementation of the product. After successful integration, the

customer will benefit from improved efficiency, reduced weight, and increased

packaging headroom throughout the lifetime of the product.

77
 What inputs does the experiment require? How much does the project cost? Who

Pays?

Table 8-2: Original Cost Estimate

Item Quantity Cost

DAW System/Switching 1 $15

Controllers

Switching Device 8 $5

Gate Drivers 1 $50

Magnetics 1 $50

Other ICs $60

Printed Circuit Board 1 $150

Connectors, Enclosure, $60

and Hardware

With a very aggressive budgeting, we believe that a total cost of

approximately $400 will be required to develop this project. These prices are

within the $200 per person grant given to EE senior projects. Moreover, we

expect to be able to nullify a large portion of the costs through existing materials

in the Cal Poly Racing inventory, and partnerships with Linear Technology and

Bay Area Circuits.

How much does the product earn? Who profits?

The predicted cost of manufacturing is approximately $95. To obtain a

margin of 40%, we will be selling our device at $130. Therefore, each device will

result in $35 of profit. The profit from selling of these devices will be split

between the two designers of the project: Nick Mah and Jason Zhou. We expect

78
to sell 20-30 units to other Formula SAE teams, leading to a potential profit of

$800-$1200.

Timing

The DC-DC will be released after the senior project expo. Afterwards, we

will begin marketing to Formula SAE teams, while probing the commercial

market for interest in our device. If there is interest, we will begin further testing

to certify our device for commercial use. After completion of testing, the device

will immediately be released to Cal Poly Racing, as they are the primary

contractor.

Figure 8-3: Winter Quarter Timeline

Figure 8-4: Spring Quarter Timeline

79
d) If Manufactured on a Commercial Basis:

Estimated number of devices sold per year: 500 (40-50 from Formula SAE and

Formula Student, the rest from small EV manufacturers)

Estimated Manufacturing Cost for Each Device: $95

Estimated Purchase Price for Each Device: $130

Estimated Profit Per Year: $17500

𝐶𝑜𝑠𝑡 $0.12
Estimated Cost for User to Operate Device: 𝐻𝑜𝑢𝑟 = 𝐶𝑒𝑙𝑒𝑐 ∗ 𝑃𝑚𝑎𝑥 ∗ 𝜂 = ∗
𝑘𝑊ℎ

$0.54
0.5𝑘𝑊 ∗ 0.9 = ℎ𝑜𝑢𝑟

e) Environmental

This product will impact the environment at all points of its use. The

creation of the DC-DC will involve the manufacturing of electrical components,

PCBs and metal enclosures. All of the above stress the environment with water

usage, waste chemicals and materials, emissions and energy usage. During usage,

the efficiency of the device will cause some wasted electricity. Since this is an

electric vehicle-oriented device, the environmental impacts of electric vehicles

should also be considered. The large battery packs of electric vehicles can lead to

very large electrical waste deposits. On the other hand, the high efficiency of

electric vehicles could offset the pollution costs associated with ICE vehicles.

Finally, at end of life, the device will need to be recycled. Because the device is

used in student-built vehicles. There is a high chance of proper electronics

recycling occurring. Additionally, because of the big ewaste hazard electric

80
 vehicles pose, there is a strong likelihood that the DC-DC will be properly

recycled along with the rest of the EV.

f) Manufacturability

A lot of effort will need to be put into the construction of the device.

Although vibration and temperature environments are not as extreme as in an ICE

vehicle, a high amount of care must be put into the reliability of the device.

Therefore, our manufacturing process will require rigorous vibration, thermal,

endurance, and isolation testing throughout the manufacturing process to ensure

customer safety.

g) Sustainability

Describe any issues or challenges associated with maintaining the completed

device or system

The primary challenge with maintaining the device is environmental

conditions. Automobiles undergo very harsh environmental and startup

conditions. Our DC-DC must be able to withstand high vibration and ambient

temperatures as well as constant power cycling. Additionally, a balance between

an environmentally proof enclosure and effective thermal dissipation will need to

be found.

Describe how the project impacts the sustainable use of resources

This product stresses the natural resources during production and end of

life. The primary stresses are water usage, ewaste, pollution/chemical emissions,

and energy usage. However, successful implementation of the DC-DC could

81
result in increased EV usage, which would help reduce the total pollution impact

of personal transportation.

Describe any upgrades that would improve the design of the project

In order to further reduce size, water cooling could be included to decrease

the required cooling effort. This would allow designers to use smaller switching

IC's, which should improve efficiency, as the switching energy will decrease. This

would also decrease the total required mass for heatsinks.

Another potential improvement is the addition of a complete

microcontroller system. This would significantly improve the tunability of the

design as well as the amount of telemetry that could be sent back.

A final potential improvement is the splitting the DC-DC into two,

separate, lower power outputs. By having two separate outputs, more flexibility of

design is afforded. The output has two separate output voltages, reducing the

amount of voltage regulators required on other boards in the car.

Describe any issues or challenges associated with upgrading the design

Water-cooling will increase the difficulty of manufacturing significantly.

Another issue with water-cooling is that we lose flexibility from our value

proposition.

The addition of a microcontroller will reduce the ease of integration for

our device. Additionally, our personal costs of development time and

manufacturing costs will go up now that we will need to support a more

complicated firmware package.

82
 The addition of a microcontroller will reduce the ease of integration for

our device. Additionally, our personal costs of development time and

manufacturing costs will go up now that we will need to support a more

complicated firmware package.

Splitting of the outputs will improve the flexibility of design. However, it

will increase the size and weight of the DC-DC, which is a primary component of

our value proposition.

h) Ethical

The DC-DC has multiple ethical concerns. The DC-DC must be made

such that it is compliant with the Formula SAE Rulebook. Workarounds that

"technically" pass the rules are also unethical as they betray the intent of the rule.

Additionally, the DC-DC must be extremely reliable. Any product going into

automotive use is purchased in faith that the device will work reliably. Any failure

modes possible with the DC-DC must be documented for the customer. Any

failure modes that could result in potentially dangerous situations must be

accounted for and prevented before release to the customer. An example of a

dangerous failure mode is a loss of isolation, which could result in the death of the

driver. Unfortunately, the designers of the product cannot control how the DC-DC

is operated. Modifications, such as reflashing the controller, could result in a

harmful system. Finally, usage of the device must be safe. High voltages are

present at the input of the device. The designers must be cognizant of this danger

and make installation safe and well documented.

83
i) Health and Safety

The primary safety concern for the DC-DC are failure modes and

interaction with the DC-DC. Loss of isolation of the DC-DC could result in

permanently damaging electric shock. Additionally, if the failure of a component

is not well contained by the enclosure, it could result in the destruction of nearby

electronics. Finally, connecting the input leads to the DC-DC could result in

electric shock as well if proper installation procedures are not followed.

j) Social and Political

Because this device is designed to be completely Formula SAE rules

compliant while maintaining a competitive price compared to existing solutions,

the competition may see a decrease in revenue. This can affect the livelihood of

employees. Because integration of the device is very simple, we expect our DC-

DC to help equalize the disparity between well-funded and underfunded race

teams.

k) Development

This project will require development in knowledge of thermal dissipation

in electronics, isolated power supply design, power electronics, environmental-

rated components, automated testing, and magnetic design.

84