Soft Switching

[shauns87]

Is there a way to implement soft switching (ZVS/ZCS) on the secondary side of phase shifted full bridge converter?

 

[cupoftea]

yes just wait till the diode conducts, then turn on sync fet in pllel with it.

 

[shauns87]

yes just wait till the diode conducts, then turn on sync fet in pllel with it.

Thanks, but won't that be just synchronous rectification? The MOSFET will still switch hard?

 

[cupoftea]

no , if the diode across it is on...the fet will turn on with v low voltage across it...not counted as hard switching....its just like in an LLC...so its effectively resonant switching.

 

[shauns87]

no , if the diode across it is on...the fet will turn on with v low voltage across it...not counted as hard switching....its just like in an LLC...so its effectively resonant switching.

Got it! Thanks. Is there any off the shelf controller available to implements this control strategy? Alternatively, in case there is no controller available then is there any reference papers or application note available that you're aware of? It would be quite helpful.

 

[cupoftea]

I think ti.com do a PSFB controller that i think includes sync rects....but you can just do your own, and use some hardware with delay blocks, to drive the sync rects ultimately from the primary fet gates...just have a little delay so that the diode gets chance to conduct b4 u short it out....also, have a load current sense, and disable the sync rects so that the current doesnt go backward in the sync rects....remember your synch rect fets must have a high ratio of CISS/CGD, otheriwse they may turn on spuriously.....

I am sure you know how to do a pulse delay in hardware...just like a diode shorted resistor into a capacitor, and a buffer either side.....so you get an rc filter one way, and just a diode and cap the other way.

 

[shauns87]

Yes. I am using LT3722 from AD which has SR controller built in but I wasn't sure if it is able to allow delay for the body diodes to turn on and achieve soft switching. However, I will test and check if it does that. Thanks again for helping me understand. Cheers!

 

[cupoftea]

LT3722 i think you can use the free LTspice sim to check it?

 

[shauns87]

Yes. I can!

 

[cupoftea]

Great, remember, you need to put a switching load in to really check out those synch rects....make it go suddenly from full load to no load, and see what happens...check that you dont get current reversing in the sync rects.
Plus make sure that the diode really does conduct b4 the fet gets switched on.
PSFB has a high leakage inductance, so make sure the ringing doesnt overvoltage your sec side sync rects.

 

[Calm down]

You can refer to an article on the "PE17" circuit on this website.

 

[shauns87]

Hi. I tried to search PE17 on edaboard but couldn't find anything.

 

[cupoftea]

search "new type of resonant converter", i think that was the name of the post

 

[Calm down]

I have a reservation that I hope can help you with

 

[dick_freebird]

ZCS (nZCS), ZVS work by delaying the low side catch-FET turnon until
the inductor current has done all the work. That determination wants
sensing and possibly more complex feedback (like, for ZCS you will see
a servo loop positioning low side switch turnon to obtain a small negative
voltage pedestal on the SW node - desire being both to minimize FET
channel conduction losses on the way down, and be smart about when
to catch the current so that it spends as little time as possible "upside
down" on the body or backstop diode.

So, yes, you -could- superimpose a ZCS or ZVS control on the bridge
by intercepting and delaying the high side, low side controls (if separate).
A more integrated bridge driver that controls both HSS and LSS from one
digital input and some break-before-make logic, might defeat your efforts.