Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Reverse recovery losses in synchronous rectifier

Status
Not open for further replies.

smazzer

Newbie
Joined
Nov 4, 2019
Messages
3
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
40
Hi everyone. I'm struggling with a tedious question about recovery losses associated withe body diode of the freewheling switch.

Let's assume positive current entering the choke of a buck converter.

During the dead time, body diode of freewheeling switch is ignited. What happen when the freewheeling MOSFET (low side device) is then gated? Current is transferred from body diode to MOSFET, but what about reverse recovery? I think this contribution will be dissipated along the channel. It is true? This thing cannot be directly verified with any SPICE simulators, and is not mentioned explicitly in any scientific article that I've consulted.

What's your opinion?

Thanks in advance

SM
 

Other people may give you the semiconductor physics but I'll point out that the 'ON' isn't where reverse recovery comes into play. There are no losses when you turn on the fet and add a channel parallel to your conducting diode.

The losses come when you turn off the sync rectifier fet, now all the current must flow through the diode again. Now when the opposite switch turns on its forced to supply the reverse recovery charge and it takes up the losses associated with the RR of its counterpart.

I stared a thread on this subject a while back:
https://www.edaboard.com/showthread.php?378926-Practical-Understanding-of-Reverse-Recovery&p=1624787
 
I've carefulley read your previous post and the ST application note. I've concluded that the body diode of synchronous fet is not experiencing reverse recovery when the fet itself is triggered, since current will be diverted from diode to channel, but operation in the 3rd quadrant never puts the diode under reverse bias, hence no reverse recovery.

Now the question gets more interesting: since the diode was carrying a certain amount of current (that must substained by a certain amount of minority carriers stored in the intrinsic region of the PiN body diode), and then it will be suddendly reduced since it is diverted to the fet channel, where is the abundant charge going? I mean...this is the same mechanism for which reverse recovery exist, but in this case the diode is not reverse biased, simply it experience a fast current transient....
 

Well, yes, as you must have a finite dead time - the losses in the bottom fet are related to the dead time - too long a dead time and the current will commute to the body diode and you will get some rev rec losses if you turn the top fet on really fast - empirically; the narrower the dead time the less of this effect you see ( due to Cds ) - but very narrow dead times are hard to maintain over temperature ... this is why you sometimes see a schottky in parallel with the lower fet ...
 

Well, yes, as you must have a finite dead time - the losses in the bottom fet are related to the dead time - too long a dead time and the current will commute to the body diode and you will get some rev rec losses if you turn the top fet on really fast - empirically; the narrower the dead time the less of this effect you see ( due to Cds ) - but very narrow dead times are hard to maintain over temperature ... this is why you sometimes see a schottky in parallel with the lower fet ...

This is true at all but actually does not answer my question.
 

This also won't answer your question: What I think is that it doesn't matter too much. The future for a mostfet that's on is that some day it's going to be turned off. At that point any positive current will forward bias the diode again so it's sort of irrelevant what the 'recovery state' was before that time.

It's slightly more interesting if there is negative current, then there is also a small reverse bias on the diode. Though also if there is a negative current then recovery losses will be less important as it will soft switch.
 

Dear smazzer, you asked several questions, which in particular do you consider was not answered ?

- - - Updated - - -

surely this was the main question :
a tedious question about recovery losses associated withe body diode of the freewheeling switch
recovery losses in this switch can only occur if there is diode forward current and you turn the top switch on fast ...
 

smazzer your particular concern, as described by others above, is not relevant in the case of the sync buck...but yes, your point is very relevant in cases such as the phase-shift-full-bridge, where your concerns can result in failure of the fets....because often there is not enough reverse bias across the (internal and/or external) diode when the fet is on...then the top fet turns on and their is severe reverse current in the intrinsic (or external) diode...sometimes resulting in failure......................................your point is also a concern in LLC converters when they operate at lower f(sw) than the lower resonant peak.

But yes, you are well cognisant of the fact that in order for a fast diode to turn off properly, it must undergo reverse recovery.....those charges must be swept out, as you imply.

Of course, with a low voltage sync buck, many would say just use an external schottky, and its problem solved because they dont have reverse recovery
 
Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top