Project Power Supply

1
Project Power Supply

• J.T. Hwang
• Paul Dykstra

2
Presentation Overview

• Discuss Specifications
• Discuss Circuit design
• Discuss Individual Parts of circuit and some
  design issues
• Discuss Problems experienced
• Discuss Results
• Conclusions

3
Project Specifications

• Input range of 100-240Vrms
• Outputs of 5,-5,12,-12,24 volts
• Total power of 100W (20W per output)
• Ripple /-1
• Try and implement a feedback control to keep
  outputs constant

4
Design considerations

• Want to use power electronics
• Want to isolate input from output
• Want to isolate outputs from each other
• Want to keep ripple low

5
Circuit designed to use
6
Breakdown of ckt. operation

• 60Hz transformer with turns ration of 12025.2
  Vrms
• Full bridge rectifier
• Flyback converter
• Gate drive circuit
• Control circuit

7
Rectifier

• 60Hz transformer steps voltage down from 120 to
  25.2 Vrms (17035.2V peak)
• Full bridge fully rectifies signal
• Capacitor makes output look like ideal voltage
  source

8
Description of flyback

• When FET is on, current goes through primary
• When FET is off, current flows through
  secondarys, giving desired ouptputs.
• Current can only flow one way on each side due to
  switch design

9
Load Resistor Values

• PV2/R
• For 5V output, Rload25/201.25ohms
• For 12V output, Rload144/207.2
• For 24V output, Rload 576/2028.8 ohms

10
Capacitor Values

• CIdt/dv
• We wanted ripple below 1
• Switching at 200kHz (5us), with duty ratio of ½,
  and dtDT
• IP/V
• Minimum C values are found at 1000uF for 5V,
  173uF for 12V, and 4.3uF for 24V

11
Avoiding Discontinuous Mode

• When L is to small, current can swing from Ipeak
  to 0, which is discontinuous mode and is shown
  below

If L is big enough, current will look like below,
this is what we want
12
Finding Lcrit

• LcritVdt/di
• di is maximum swing of current on primary side,
  which will be largest for 240Vrms (inductor sees
  70.7Vdc) input at 1.4A,
• dt is DT5us.333 for 240V input
• Lcrit gt/83uH
• We designed to have much less swing, and used
  inductor value of about 1.8mH to get rid of a lot
  of ripple

13
Gate drive of FET

• Variable duty and frequency allow for different
  switching capabilities.
• Snubber gives smooth transitions in the switching
  action
• 35V comes from rectified signal
• 12V zener used for voltage regulator

14
Duty Ratios

• For a flyback converter, VoutVin(D1/1-D1) where
  D1 is the duty of the FET
• For 100V input D1.6
• For 120V input D1.5
• For 240V input D11/3

15
Feedback control circuit

• Voltages are compared to see if duty need to be
  adjusted
• Output10(VrefVout
• The output gets fed into the duty of the PWM
• This circuit was built, and tested, but we
  couldnt get it to function correctly in our
  converter

16
Parts Used

• Bridge Rectifier used 4 MUR 820s (200V, 8A)
• Output Diodes were MBR1045s 45V, 10A)
• FET used was IRF740 (400V, 27A)

17
Problems with cores

• Originally tried toroid core.
• This core was designed for lower frequencies than
  what we were switching at, and had too many
  losses because of low permeability.
• Switched to pot core.
• Higher permeability of pot cores solved our
  problems and gave better outputs.

18
Wire information

• Current density used was 500A/cm2
• Magnetic wire of 3 different gauges was used
• 5V and primary used 18 gauge (rated for 4.2A)
• 12V outputs used 20 gauge (rated for 2.06A)
• 24V outputs used 22 gauge (rated for 1.6A)

19
Pot Core specs

• Inductance factor8.1uH
• 15 turns (primary) gives 1.82mH
• 5V 3 turns
• 12V 7 turns gives 15.6V
• 24V 12 turns gives 26.8V
• Minimum of 2V to use voltage regulators

20
Other Problems Experienced

• Our circuit was drawing excessive amounts of
  current, causing strange behavior and FET getting
  extremely hot. Why?
• Insulation on Magnetic wires was wearing off,
  causing some leakage flux and current, which
  accounted for this

21
Ripple considerations

• Needed to minimize ESR of capacitors to lower
  ripple
• Added L-C filters to further minimize ripple
• Below shows this adjusted circuit at output

22
Ripple results
23
Efficiency Results
24
Conclusions

• We were able to use a flyback to design this
  project power supply
• We experienced trouble with certain types of
  inductor cores
• Overall, we were reasonably successful at getting
  good efficiency and low output ripple