EE152 Green Electronics - Stanford...

66
EE152 Green Electronics Embedded Software 9/28/16 Prof. William Dally Computer Systems Laboratory Stanford University

Transcript of EE152 Green Electronics - Stanford...

Page 1: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

EE152 Green Electronics

Embedded Software9/28/16

Prof. William DallyComputer Systems Laboratory

Stanford University

Page 2: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Logistics• Class is NOT full. See me for “permission numbers”• Lab group assignments

– Sign up. Tentative assignments will be announced next week.• Sign up for Lab signoff time

– Lab 1 must be signed off next week - individually• Homework 1 due next Monday 10/3• Sign up for Piazza

– piazza.com/stanford/fall2016/ee155• Discussion sections

– Fridays 4:30-5:50PM– Herrin T185

2EE155/255 F15 L2

Page 3: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Course to Date• We need sustainable energy systems• At the core they are voltage converters• Periodic steady-state analysis, buck and boost• Intelligent control + power path

EE155/255 Lecture 3 - Power Devices

Page 4: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Today• Logicstics• Brief Review• Embedded software• Power Devices

EE155/255 F15 L2 4

Page 5: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

EE155/255 F15 L2 5

Page 6: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Review of Monday– PSSA, Buck, Boost• Green energy systems are made from voltage converters

– PV systems, electric/hybrid cars, wind power, etc…– Power path + intelligent control

• Separate behavior into fast and slow– Fast – within switching cycle– Slow – f << fcy

• Within cycle solve for periodic steady state– State variables the same at start and end of cycle.

• For slow transient response– Write difference equations

• Buck/Boost Topologies

L

V2

+-

iLV1

+-

a

b

L

V2+-

V1+-

a

b

iL

V1 =V2Da

=V2

1−Db( )V2 =V1Da

0 5 10 15 20 25 30 35 400

0.5

1

1.5

2

2.5

3

3.5

time (µs)

V 2 (V)

ω =1e+06ζ =0.2

f = 1.59e+05Q = 2.5

0 5 10 15 20 25 30 35 40

−0.1

−0.05

0

0.05

0.1

0.15

I L (V)

ip

i0

t0 tb tcy

iL

b/a

tb ta

6EE155/255 F15 L2

Page 7: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

How Does This Work?

EE155/255 F15 L2 7

VS

+

- vC Load

+

C

a b

iLoadLiL

Page 8: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Another View

EE155/255 F15 L2 8

VS

+

-

vC Load

+

C

a b

iLoad

LiL

Page 9: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Embedded Software

9EE155/255 F15 L2

Page 10: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Embedded Software• Most software is embedded

– 50 microprocessors in the average car– Several in many appliances– At the heart of most green electronic systems

• Embedded software forms the “slow” but “complex” part of most control loops– The fast, simple parts are in hardware (analog or digital)

10EE155/255 F15 L2

Page 11: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Embedded System

Microcontroller

Fast Hardware

Sensors Actuators

11EE155/255 F15 L2

Page 12: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Example: PV Maximum Power Point Tracker

MicrocontrollerPanel Voltage

PWM to Boost ConverterPanel Current

12EE155/255 F15 L2

Page 13: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Sometimes Embedded Systems Interact with Humans

Microcontroller

Fast Hardware

Sensors Actuators

User Interface Debug Log

13EE155/255 F15 L2

Page 14: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

UI Design• It should be obvious how it works• Minimum number of “modes” (1 is best)• Display clearly indicates

– Current mode– Relevant data– Progress indicator if “waiting”– “Button” labels

• If the user needs to read the manual it’s a bad UI• Refresh display every 50ms• Check buttons and update “state machine” every 50ms

EE155/255 F15 L2 14

Page 15: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Event Driven• Embedded systems execute code in response to events

– A real-time clock interrupt (e.g., every 1ms)– An I/O device completes an operation (e.g., A/D converter)– An input transitions (logic low-high, or analog threshold)

• Code is of the form– When trigger, do action– When 1ms clock tick, refresh display– When A/D completes, process data & start next conversion– When current over limit, shut down PWM

15EE155/255 F15 L2

Page 16: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Embedded vs Conventional Software

Embedded• Event driven

– Most code runs in response to interrupts

• Real-world I/O– A/D in, D/A and PWM out

• Real-time– Update PWM within one cycle

Conventional• Run code in program order

– In response to user input (e.g. mouse click)

• User Interface I/O• No hard time constraints

EE155/255 F15 L2 16

Page 17: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Real-Time Clock• Microcontroller has a counter-timer that can be programmed to interrupt

the processor periodically (e.g., every 1ms).

17EE155/255 F15 L2

Page 18: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

STM32 Timers

EE155/255 F15 L2 18

Advanced-control timers (TIM1) RM0364

381/1121 DocID025177 Rev 1

Figure 106. Advanced-control timer block diagram

1. The internal break event source can be:- A clock failure event generated by CSS. For further information on the CSS, refer to Section 7.2.7: Clock security system (CSS)- A PVD output- SRAM parity error signal- Cortex-M4 LOCKUP (Hardfault) output.- COMPx output, x = 1,2,3,5 and 6.

Page 19: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Library Code

void my_callback1();timer_init();timer_id_t my_timer1 = timer_register(1, &my_callback1, GE_PERIODIC);timer_start(my_timer1);

19EE155/255 F15 L2

Make sure to check my_timer1 for error code

Page 20: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

/** * @brief Registers a callback function to an internal Timer* @details Sets up a timer based on the required period. The* function will then assign the callback function to the* appropriate timer interrupt. The function uses TIM3 to generate* the interrupts.* * @param ms Period in milliseconds (int)* @param function Callback function* @param type SINGLE_SHOT or PERIODIC* @return Timer ID of associate timer or error code.*/

timer_id_t timer_register(uint32_t period, void (*function)(void), uint8_t type) {

EE155/255 F15 L2 20

Page 21: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Timer Interrupt

rtc

RTC interrupt

Every 1ms

Every 3ms

Every 50ms

rtc rtc rtc

rd rd

tick50()

rtc

21EE155/255 F15 L2

Page 22: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

I/O Completion• I/O devices interrupt when they complete an operation• Handler should

– Process input data (if an input device)– Start next operation

22EE155/255 F15 L2

Page 23: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

EE155/255 F15 L2 23

Analog-to-digital converters (ADC) RM0316

220/1037 DocID022558 Rev 3

13.4 ADC functional description

13.4.1 Block diagram of a single ADCFigure 28 shows the ADC block diagram and Table 36 gives the ADC pin description.

Figure 28. ADC block diagram

Page 24: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

A/Dvoid my_adc_callback(uint16_t *buf) ;

adc_set_fs(10000); // 10kHz sample frequencyadc_callback(&my_adc_callback); // register callbackadc_enable_channels(chan_to_conv, NUM_ADC); // specify channels adc_initialize_channels(); adc_start();

24EE155/255 F15 L2

Page 25: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

A/D Timing

25EE155/255 F15 L2

ADC Start Clk

ADC1ADC2

Convert VConvert I

compute_power()

20us100us

Page 26: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

• What happens if timer interrupt handler is still running when ADC interrupt occurs?

EE155/255 F15 L2 26

Page 27: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

• Make sure your interrupt driven code runs quickly!!!

• You can toggle a GPIO pin at entry and exit of your code to see how long it takes on a scope.

EE155/255 F15 L2 27

Page 28: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Volatile Variables• If you run code with “interrupts enabled” other code may run “at the same

time”.– And long running code (more than 200us) should not disable interrupts– Enable interrupts after “critical section”.

• That “other code” may change variables• Don’t expect variables to be stable while “background” code runs!

28EE155/255 F15 L2

Page 29: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Integer Arithmetic• Much embedded code uses fixed-point arithmetic. • Floating-point takes too long (on an integer uC)

– Some modern uCs (e.g., STM32F3) have fast FP• Fixed point has plenty of precision for typical control values

– int (16 bits) [-32,768, 32,767]– long (32 bits) [-2,147,483,648, 2,147,483,647]

• But you need to scale your variables and check for overflow– Example – 16b represents voltage in mV, current in mA– Multiply mV and mA to get 32b power in uW– Divide by 1,000 and truncate to get 16b power in mW

• Shift and multiply – don’t really divide – its expensive• Multiply by 1.024 and shift right 10

EE155/255 F15 L2 29

Page 30: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Fixed-Point Arithmetic• Scale numbers to be in [-1,1)• Binary point is just to the right of the sign bit

– s.bbbbbbb…• On multiply retain high half of result• General fixed-point representation

– saaaaaaaaa.bbbbbb…– Represents a value to 1/64– Need to shift after a multiply to put binary point in the correct place

• Need to think about the range and precision of your variables

EE155/255 F15 L2 30

Page 31: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Floating Point• ST32F3 has fast 32-bit floating point• Max value of ~ +/-3 x 1038

• Min non-zero value of ~ 1.2 x 10-38

• Represents numbers to better than six sigfigs

• Scale inputs and outputsfloat_volts = (float) int_volts * VOLTS_PER_DIV ;

– 12b ADC with 3V range – 3V/4096 = 0.000732 volts/div

• Always check for range, never equality

EE155/255 F15 L2 31

Page 32: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Organization of Code• Control• User Input• Display• Monitoring• Housekeeping

32EE155/255 F15 L2

Page 33: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Layered Control: Example Battery Charger

State Selection

Target Selection

State

Current 3A

PWM Control PWM to Buck

Battery V, IInput V

Battery V, I

Driver I

User Interface

Mode Status

33EE155/255 F15 L2

Page 34: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Layered Control: Example Battery Charger

State Selection50ms

Target Selection1ms

State

Current 3A

PWM Control10us FPGA

100ns - ComparatorPWM to Buck

Battery V, IInput V

Battery V, I

Driver I

User Interface50ms (or foreground)

Mode Status

34EE155/255 F15 L2

Page 35: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Example Photovoltaic Controller• Mode selection

– Line test, Panel test, MPPT, etc…• Monitors

– Line (current, voltage, frequency, …) (anti-islanding)– Internal operation (voltages, currents, temps) – Panels (current, voltage)

• Maximum Power-Point Tracking– Search algorithm for peak power point

• Inverter control– Convert 400V DC to 240V AC – in sync with line

• User interface– Display status, serial output, network connections, …

35EE155/255 F15 L2

Page 36: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Summary – Embedded Software• Embedded software consists of event handlers

– When trigger do action• Timer is used to perform periodic tasks

– Schedule A/D converter every 20us– Report status on serial port every 10ms– Update mode every 50ms– Etc…

• I/O devices trigger handler on completion– Process input data– Start next I/O operation

• Layered Control– High-level control loop sets parameters of lower level control loops– Fastest loops may be in hardware

36EE155/255 F15 L2

Page 37: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Power ElectronicsPart 1 – Power Devices

EE155/255 F15 L2 37

Page 38: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Real Switches

EE155/255 F15 L2 38

L

V2

iLV1

+-

a

b

R

C

+

_

M1

M2

V1

High-SideGate Drive

Low-SideGate Drive

VGL

VGH

inH

inL

GND

X

G1

G2

Page 39: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Real Switches• Finite switching time• Finite voltage blocking• Non-zero on resistance• Parasitic L and C

EE155/255 F15 L2 39

L

V2

iLV1

+-

a

b

R

C

+

_

M1

M2

V1

High-SideGate Drive

Low-SideGate Drive

VGL

VGH

inH

inL

GND

X

G1

G2

Page 40: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Quick Summary in a Few Pictures

Page 41: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

DC I-V Characteristics of On Switches

V(d)0.0V 0.1V 0.2V 0.3V 0.4V 0.5V 0.6V 0.7V 0.8V 0.9V 1.0V 1.1V 1.2V 1.3V 1.4V 1.5V 1.6V 1.7V 1.8V 1.9V 2.0V

0A

5A

10A

15A

20A

25A

30A

35A

40A

45A

50AIx(m:1) Ix(h:2) Ix(i:C) I(Dd)

600V FETFCB36N60

600V IGBTFGH40N60

400V DiodeSTTH20R04

60V FETIRLB3036

FETs CharacterizedBy RON

IGBT Like a DiodeLittle Current Until ~0.7V

HV FET has high RONR ~ kV2

Diode and IGBTResistive at high Current

Page 42: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Transient Response of FET and IGBT

0ns 40ns 80ns 120ns 160ns 200ns 240ns 280ns 320ns0V

50V

100V

150V

200V

250V

300V

350V

400V

450V

500V

550V

0A

5A

10A

15A

20A

0V

50V

100V

150V

200V

250V

300V

350V

400V

450V

500V

550V

0A

5A

10A

15A

20A

0KW

1KW

2KW

3KW

4KW

5KW

6KW

7KW

8KW

9KW

10KW

V(di) Ix(i:C)

V(dm) Ix(h:2)

ix(i:c)*v(di) ix(h:2)*V(dm)

600V FETFCB36N60

InstantaneousPower

Turn On Turn Off

600V IGBTFGH40N60

36µJ 36µJ92µJ 428µJ

Page 43: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Boost Configuration for Transient Test

Page 44: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Diode Reverse Recovery

EE155/255 F15 L2 44

I1

IRM

ta tb

trr

Qrr

iD

(Amps)

t(sec)t1

t2

t3

Page 45: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Diode Reverse Recovery

0ns 3ns 6ns 9ns 12ns 15ns 18ns 21ns 24ns 27ns 30ns-220A

-200A

-180A

-160A

-140A

-120A

-100A

-80A

-60A

-40A

-20A

0A

20A

40AI(D1) I(D2) I(D3) I(D4)

0ns 10ns 20ns 30ns 40ns 50ns 60ns 70ns 80ns 90ns 100ns 110ns 120ns-810A

-720A

-630A

-540A

-450A

-360A

-270A

-180A

-90A

0A

90A

180AI(D1) I(D2) I(D3) I(D4) I(D5) I(D6) I(D7)

400V DiodeSTTH20R04

Medium Speed Diode

Area under curve (Charge)is approximately constant QRR

Page 46: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Power MOSFETs

Page 47: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Power MOSFETs are your friends

Page 48: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

MOSFET Properties• Fast switching time 10-50ns

– Low switching losses• Low conduction losses at low voltages

– V2/R of 1-2MW • e.g., at 20V R = 0.4mOhm (4mV drop at 10A)

– Typically better than IGBT up to ~400V• Easily paralleled for lower Ron, More I, or easier cooling• Integral body diode in DMOS FET• Avalanche breakdown can be used to “snub” overshoot

Page 49: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

MOSFETs are Switches

Gate

Drain

Source

Drain

Source

GateRon

Dbody

(a) (b)

Page 50: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

(a)

gaten n

p

VDS

(b)

gaten n

p

VDS

- - - - -

IDS

g=0 g=1

IDSg

s d

g

s dIDS=0

+

+

How do they Work

VGS

Page 51: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Power MOSFET (DMOS) Structure

n+

p+n+

n-

p+n+

gatesource source

drain

channel

Page 52: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Gate Charge vs VGS

EE155/255 F15 L2 52

0 10 20 30 40 50 60 70 80 900

1

2

3

4

5

6

7

8

9

10

QG (nC)

V GS (V

)

Page 53: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

CDG

LS

LGRG

LD

CDS

CGS

source

gate

drain

Parasitics

EE155/255 F15 L2 53

130 Green Electronics

CDG

LS

LGRG

LD

CDG

CGS

source

gate

drain

Figure 17.4: MOSFET parasitic circuit elements. At high operating frequen-cies these parasitic circuit elements have a large e↵ect on the performance andswitching losses of a MOSFET.

Symbol Value Units DescriptionCDS 200 pF Drain-source capacitanceCDG 70 pF Drain-gate (Miller) capacitanceCGS 3600 pF Gate-source capacitanceQG 86 nC Total gate turn-on chargeQGD 35 nC Gate-drain turn-on chargeLS 7 nH Source inductanceLD 3 nH Drain inductanceLG 7 nH Gate inductanceRG 1.5 ⌦ Gate resistance

Table 17.1: Values of parasitic circuit elements for a typical 600V 100m⌦ powerMOSFET in a TO-220 package.

600V 0.1W FET TO220 Package

Page 54: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

CDG

LS

LGRG

LD

CDS

CGS

source

gate

drain

Parasitics

EE155/255 F15 L2 54

CDS – CV2 energyLD, LS – I2L energy

slows switching, overshoot,corrupts gate drive

CDG – slows turn onRG, LG – slow device turn-on

Page 55: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

EE155/255 F15 L2 55

Page 56: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Typical MOSFETs

EE155/255 F15 L2 56

Copyright

(c)

2011-15

by

W.J

Dally,allrights

reserved

133

Device 20V IRLB3036 IRFB4227 FCB36N60N EPC2010 C2M0025120 Units DescriptionVDSmax

20 60 200 600 200 1200 V Maximum VDS

Ron

1.9 20 81 18 25 m⌦ On resistanceQG 91 70 86 5 161 nC Gate chargeCoss 1020 460 80 270 220 pF Output capacitance CSD + CGD

IDmax

195 65 36 12 90 A Maximum continuous drain currentIDM 1100 260 108 60 250 A Maximum pulsed drain currentEAS 290 140 1800 mJ Single-event avalanche energyP

max

380 330 312 463 W Maximum power dissipationV 2/R 1.9 2.0 4.4 2.2 58 MW Figure of meritV 2/RQG 21 29 52 440 360 mV/s Figure of merit

Table 17.2: Key parameters of six field-e↵ect transistors.

GaN SiC

Page 57: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Power MOSFETs should be ON or OFFThey are not happy in between

• IRLB3036 • Can handle 60V (when its off)• Can handle 195A (when its on – if you can cool it)

– I2R = (195)2(0.002) = 76W• But it can’t handle 60V and 195A at the same time

– P = VI = (60)(195) = 11.7kW– At least not for very long

• Turn them on and off quickly• Best circuits are “soft switching”

– Zero-current switching (ZCS) or zero-voltage switching( ZVS) or both.

Page 58: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Power Diodes

Page 59: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Diode Properties• Self-controlling switch

– Allows current in one direction– Turns off when current reaches zero (in theory)

• Relatively fixed voltage drop independent of current– 0.5 to 2.0V– High losses at low voltages

• Care required to operate in parallel– Current hogging

• Turn-off delay– Must clear space charge out of junction

• Turn-on delay– Negligible for most fast diodes, but some are problematic

Page 60: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Key Parameters• Reverse breakdown voltage

• Maximum current

• Reverse recovery time

• Junction capacitance

Page 61: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Diode Reverse Recovery

EE155/255 F15 L2 61

I1

IRM

ta tb

trr

Qrr

iD

(Amps)

t(sec)t1

t2

t3

Page 62: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Diode Forward Recovery

EE155/255 F15 L2 62

VFP

VF

tfr

vD

(Volts)

t(sec)

1.1VF

Page 63: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Diode Forward Recovery

Good Diode Bad Diode

Page 64: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

Two 2A Diodes < One 4A Diode

EE155/255 F15 L2 64

0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 1.05 1.10

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

25°C

75°C

125°C

VD (Volts)

i D (A

mps

)

Page 65: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

CDG

LS

LGRG

LD

CDS

CGS

source

gate

drain

Summary – Power Devices• Finite, non-zero

– Switching time– Blocking voltage– On-voltage (resistance)

• Parasitics – L and C• MOSFETs – switches

– Turn on/off as fast as gate can be charged

– R = kV2

• Diodes– Self-controlled switches– Reverse recovery loss QRR

EE155/255 F15 L2 65

Gate

Drain

Source

Drain

Source

Gate RonDbody

(a) (b)

I1

IRM

ta tb

trr

Qrr

iD

(Amps)

t(sec)t1

t2

t3

Page 66: EE152 Green Electronics - Stanford Universityweb.stanford.edu/class/ee152/lecture_slides/Embedded_092816.pdf · Embedded Software • Most software is embedded – 50 microprocessors

No Date Topic HWout HWin Labout Labck Lab HW1 9/26/16Intro(basicconverters) 1 1 IntrotoST32F3 PeriodicSteadyState2 9/28/16EmbeddedProg/PowerElect.3 10/3/16PowerElectronics- 1(switches) 2 1 2 1 ACEnergyMeter PowerDevices4 10/5/16PowerElectronics- 2(circuits)5 10/10/16Photovoltaics 3 2 3 2 PVMPPT PVSPICE6 10/12/16FeedbackControl7 10/17/16ElectricMotors 4 3 4 3 MotorcontrolMatlab Feedback8 10/19/16IsolatedConverters9 10/24/16SolarDay 5/PP 4 5 4 Motorcontrol- Lab/ IsolatedConverters

10 10/26/16Magnetics11 10/31/16SoftSwitching 6 5/PP 6 5 PS MagneticsandInverters12 11/2/16ProjectDiscussions13 11/7/16Inverters,Grid,PF,andBatteries 6 P 6 Project14 11/9/16Thermal&EMI15 11/14/16QuizReview C116 11/16/16Grounding,andDebuggingQ 11/16/16Quiz- intheevening C2

11/21/16ThanksgivingBreak11/23/16ThanksgivingBreak

17 11/28/1618 11/30/16 C319 12/5/1620 12/7/16Wrapup

TBD Projectpresentations PTBD Projectwebpagedue EE155/255 F15 L1 66

In Upcoming Lectures