DCM flyback and switching loss at turn-off

[treez]

Hello,

The switching loss of a DCM flyback is virtually zero at turn on (because theres no current flowing)...........however, it has high switching loss at turn off.......supposing I add an RCD snubber to the fet...............can I confirm that this doesn't make for overall efficiency increase, but does make dissipation in the fet less?

 

[ahsan_i_h]

Your question is not very clear to me so,
Lets discuss how a RCD across mosfet keeps the mosfet cool..

Whithout RCD across mosfet :
When the mosfet is turning off, at that time high rate of voltage rise at drain causes a capacitive current to flow towards gate through drain-gate capacitance. This current dont let the mosfet to get turned off quickly. As a result mosfet get heated.

With RCD across mosfet :
When the mosfet is turning off, at that time RCD causes relatively lower rate of voltage rise at drain. As a result current through drain-gate capacitance is low. So the mosfet can easily get turned off within zero (possible minimum) time. So mosfet stays cool.
But as there is a finite voltage rise rate at drain so there is a switching loss. This energy loss actually get stored in the C of RCD and later get dissipated as heat by the R of RCD.

To make confirm that RCD is working, you check the voltage rise rate at drain without RCD then with RCD. You will get a difference.

 

[mtwieg]

A slew-limiting RCD snubber can actually improve overall efficiency, in some cases. If you use a non-dissipating snubber then significant improvements can be seen, but at the cost of more components.

Also you will, in fact, dissipate power at turn-on due to the capacitance at the drain node, though usually this is small at lower switching frequencies.

 

[treez]

So the mosfet can easily get turned off within zero (possible minimum) time. So mosfet stays cool

..thanks, I have been thinking more about this..i am not sure that the RCD snubber makes the fet turn off any quicker, but rather, I believe that the RCD snubber simply diverts much of the current away from the fet and into the snubber capacitor (instead of it going through the fet).....so the switching loss in the fet is less since there's less current through the fet at switch off.
As you know , the following is to what I refer...


The RCD snubber....R2 D2 C2 in following..
https://www.daycounter.com/Calculators/Snubbers/Snubber-Design-Calculator.phtml


.......as Ahsan_i_h and Mtwieg kindly point out, there is dissipation on the turn-on due to the snubber capacitor discharging through the snubber resistor....but the point is that this discharge is a "low-rms" current discharge, as there is a filtering effect of the RC....so therefore, we have a lower rms current and therefore a slight overall increase in efficiency?

So anyway, the RCD snubber is simply a devce which "diverts" current out of the fet at switch off.....and thus decreases fet dissipation , but does not really improve efficiency much?

 

[ahsan_i_h]

..thanks, I have been thinking more about this..i am not sure that the RCD snubber makes the fet turn off any quicker, but rather, I believe that the RCD snubber simply diverts much of the current away from the fet and into the snubber capacitor (instead of it going through the fet).....so the switching loss in the fet is less since there's less current through the fet at switch off.

I also agree and unfortunately overlooked this in my previous post. As voltage rate snubber let the drain voltage to rise slowly so switching will happen at lower voltage and hence lower loss but additionally the mosfet turns off quickly also (what i said before) results a minimum loss.

Besides I understand that, diversion of current away from fet is not correct. Average switching current through fet during turn off is (current before switchoff + current after switchoff)/2 which is same for without RCD or with RCD. So average switching current is not reduced when we put RCD but switching voltage is reduced due to slow rise rate of voltage across switch.

 

[treez]

Thanks Ashan_i_h....but I really think we are making the same point, just from a different perspective.

at switch off of fet, with rcd, the fet gate drops through its miller plateau and linear region with the drain at a lower voltage than without the rcd.....the current diversion argument is valid.........at switch off, the current that would usually flow through the fet is now flowing into the snubber cap (say its boost converter)......the current is the peak current level of the inductor current......until the boost diode conducts, that inductive current has to flow somewhere..it either flows through the fet or into the snubber capacitor..so what I am saying is, that its the diversion of current out of the fetduring switch off whichresults in less fet dissipation...but as you rightly imply, the fet still went thru' its miller region and its linear region, but just that the drain voltage was less when It did this.

IMHO, I believe that the "current diversion" perspective is more intuitive.