Component damage on PSFB

Thread Starter

harinikrishna

Joined Apr 3, 2020
33
HI,
Hope you are well.
I am working on a PSFB converter for a 70V, 700 W design.
When making a no-load test, The components on the primary side are continuously failing. It is either the MOSFETs or the diodes on the primary side.

1617453963184.png

I measured the gate signals on the fet A at the following input

at input AC voltage 70 V( rectified DC 100V)

We were driving the gate with 13V

( we measured FET A gate voltage)

70Vac.jpeg-320x240.jpg

At 80 V AC ( 113V DC) observed the following Waveform at gate A and after some time the diodes went bad ( D1 and D2)

80Vac.jpeg-320x240.jpg


I am confused about what the diodes or the fets are constantly going bad while running the no-load test.
I am giving a rectified input to the converter.
The fet I am using is IPA60R280P7SXKSA1
http://www.farnell.com/datasheets/2718753.pdf

And the diode D1 and D2 are BYV29B-500,118
http://www.farnell.com/datasheets/2893371.pdf

(Moderator's note: You schematic has been rotated so people won't have to roll their monitors over on the side to read it.)
 
Last edited by a moderator:

Thread Starter

harinikrishna

Joined Apr 3, 2020
33
I want to make sure there is no voltage spike that is causing the fets or the diodes to go bad.
Can I directly probe the transformer terminals to measure the voltage at the input of the primary? ( the voltages are above 200V DC at the primary side)
 

Irving

Joined Jan 30, 2016
2,105
How/from what are you driving the gates? Can you show 1 cycle of all 4 gate drives on one screen - and make the images much bigger, say at least 1024 x 768 or so else they are hard to read.

The Siglent 1104x-e (great 'scope, just got one this week) is rated for 400v peak input with x10 probes. To monitor the primary voltages you could use 2 probes (CH1 + CH2) differentially and display CH1 - CH2, with the grounds connected to your primary input 0v (assuming that is isolated from mains ground). If you think the primary side voltages are likely to be >400v pk (or if not sure) use a HV divider made from 10 x 100k 1% 0.25W resistors in series permanently wired from each end of transformer primary to ground. You can then safely monitor x100 (1000V 9:1), x50 (500V 8:2) and x20 (200v 5:5) using a x10 probe on the 10v/division range.

I suggest you also monitor drain/source currents in each MOSFET by including an 0.5ohm 10W resistor in each source connection and using a differential 2-probe connection to monitor the voltage across the resistor (again via a HV divider for the upper side MOSFETs).

Post pictures of gate drive and matching current and voltage traces....
 

Irving

Joined Jan 30, 2016
2,105
Here is what I'm suggesting...

1617473266029.png

How did you measure the gate voltages, were they all relative to 0v (- rail)?

Can you show the gate drive circuits...
 

Irving

Joined Jan 30, 2016
2,105
Seems to me that anytime A turns on, D2 has full power supply current, same with B and D1.
D1 and D2 are reverse biassed and blocking when the MOSFETs are on and serve to provide a return path for the back emf of the collapsing magnetic field in the transformer (though the MOSFETs have built-in diodes so technically may not be needed).
 
Top