500W offline Full Bridge SMPS with general purpose diodes

[cupoftea]

Hi,
Has anyone ever done a 500W Full Bridge SMPS (275Vin, 28Vout, 250kHz) with general purpose diodes placed across the Full Bridge Primary FETs?
This schem has CGRKM4005-HF diodes in those 4 places.

CGRKM4005-HF

https://www.comchiptech.com/admin/files/product/CGRKM4001-HF%20Thru.%20CGRKM4007-HF%20RevG.pdf

If you have done this, may i ask what happened?

 

[Easy peasy]

assuming the fets are fets - do they really need anti-parallel diodes ?

certainly the ones in the link are likely to go poof at 500kHz, if they ever carry any fwd current and are then forcibly commutated off ...

 

[cupoftea]

Thanks, the FETs are IRF740L

https://www.vishay.com/docs/91053/91053.pdf

These FETs have internal diodes which have trr = 570ns.
...even these look bad in that respect.

JTLYK , This Full Bridge has a output inductor ripple frequency of 250kHz. As such, i call this a 250kHz Full Bridge...i appreciate that this is not standard terminology

 

[Easy peasy]

If the output inductor has 250kHz ripple, then the fets are switching at 125kHz

The slow mosfet diodes are fine as long as standard PWM is used

however the turn on losses in the IRF740 may overwhelm the fets ....

--- Updated Nov 14, 2022 ---

p.s. turning these fets on faster - will up the dv/dt in the ckt and likely interfere with the control chip and signal sensing - esp if thru a single layer of insulation to an earthed heatsink - this almost always catches out designers who try to go to HF with std pwm converters ....

there is also the effect on the output diodes if the mosfets are turned on too quickly ....

--- Updated Nov 14, 2022 ---

Calc losses for these fets @125kHz, 275VDC bus, 2A load, just about 5W each @ 80 deg C junc, 50nS turn on time ( current rise time - not Vds fall time ).

Turn off losses are assumed to be < 0.5W ( fast turn off, < 30nS gate fall time )

--- Updated Nov 14, 2022 ---

Will be extra for any snubbers needed for fets and o/p diodes ....

 

[FvM]

Any diodes other than schottky won't carry relevant share of total current when paralleled with Si MOSFET and are more or less useless.

 

[Easy peasy]

We can see here for the mosfet - top, and the diode, bottom, that in the range of interest - say 0 - 2A the mosfet will indeed carry the lions share of current ...
This says something about the experience of the designer who placed the diodes in the 1st place