2 Transistor forward with double coil gate drive

[cupoftea]

Hi,
Why do they use pulse transformer output for top and bottom?

UCC28511: Two transistor forward converter - Power management forum - Power management - TI E2E support forums

Part Number: UCC28511 Dear all, I am designing PFC and two transistor forward converter with UCC28511 controller. PFC works well, but I have problem with forward

e2e.ti.com

It just means more components.
You could pllel those two output coils and get lower leakage to drive just the top FET with.

 

[D.A.(Tony)Stewart]

It looks like the FAN IC is pushing 12 A during 20 ns to drive the coils which amplify even more. The controlled delay time with coils and FAN must give them the low dead-time needed at high Amp-MHz rates in low RdsOn FETs with high CIss.

FAN312x drivers incorporate the MillerDrive™
architecture for the final output stage. This bipolar /
MOSFET combination provides the highest peak current
during the Miller plateau stage of the MOSFET turn-on /
turn-off proces

 

[Easy peasy]

The Rt for that chip sets the osc freq to 100kHz or so

the Tx for the GD has two outputs to make the delays exactly the same - this gives better turn on and turn off performance ( balanced ) on the fets in the 2 sw forward.

also - if something blows up - it would usually not take out the controller - V useful

 

[cupoftea]

the Tx for the GD has two outputs to make the delays exactly the same - this gives better turn on and turn off performance ( balanced ) on the fets in the 2 sw forward.

Thanks, i guess by "balanced"you mean the switching losses are more evenly shared between top and bottom fets?

But ayk, if you have unequal delay, then you just get the top fet taking the turn on switching loss, and the bottom fet taking the turn off switching loss, which is pretty well "balanced" of itself?

Take your point about the controller protection though...having said that....if the PSU gets a problem like that, its finished anyhow.....kaput. (i know i shouldnt say that and should be into repairing and rejuvinating them, the environment, etc etc)

There are more cynical views where its said the more wound coils the better...the more hard manual work it involves, and the more put-off-doing-it-themselves that western people will be....so the more they will order from China etc....not that UK industry is pretty well completely finished off anyway due to the leaching of the importing western middle men over the years...who have diminished it. (no criticism)

 

[Easy peasy]

if the bottom fet is driven directly, it will likely turn off first, and turn on first

hence it carries nearly all the sw losses (?)

 

[cupoftea]

Thanks yes you are right, i was stupidly looking at my other circuit with an extra bit, where i have top on first, bottom off first. (its got a delay block added to the bottom fet, with RC forward and diode,C back).

Sorry, Not had the first coffee yet.....if the bottom fet turns on first, then the top fet takes all the ON switching losses
If the bottom fet turns off first, then the bottom fet takes all the OFF switching losses....in other words, direct drive for lower fet shares the switching losses well....
...and so back to my initial question.....why do people have pulse transformer drive for top AND bottom fets.?

i also wondered, since 2 tran forward is for low powers like up to around 300W, then do we even care if the bottom fet did take all the switching losses?
I mean, ayk, its still better switching losses than we'd get with a flyback.

 

[cupoftea]

So........We have the Two Transistor forward with upper FET driven by Gate Drive Transformer. (as attached LTspice and jpeg)
The lower FET, which isnt pulse transformer driven, will ovbiously turn off first...so there is no need to turn off the upper fet quickly.
However, we do need it to turn off definetely, so we use a PNP turn off anyway.
Given this, would you agree that MMBT2907 is better than FMMT591A for the PNP turn off transistor?

MMBT2907

https://www.onsemi.com/pdf/datasheet/pzt2907a-d.pdf

FMMT591A

https://www.diodes.com/assets/Datasheets/FMMT591A.pdf

The FMMT591A has much higher hfe at High Ic than MMBT2907.
As such, it will be slower.

There is much leakage inductance based ringing at the gate drive transformer secondary, and so the PNP turn off
actually suffers from the effect of slightly and unwontedly discharging the FET gate just after turn ON.
This bad effect, will be less pronounced with the MMBT2907, and so therefore it is preferred over the
FMMT591A...would you agree?

 

[Easy peasy]

You are nearly correct,

If the bottom fet turns off first, it sees the full bus volts at turn off, and interrupts the current - so more losses

when the bottom fet turns on 1st, it has to collapse all the bus volts due to total Cds of both devices ( instead of half the volts for balanced), which then shortly puts all the volts across the other device which has increased losses at turn on.

Under balanced gate drive ( balanced in time ) each fet sees half the bus volts and the same current at turn on and turn off, if you turn off really fast you can negate most of the turn off losses.

 

[cupoftea]

Thanks, and do you believe its worth having an extra set of gate drive circuitry for that benefit?
Maybe a 1W improvement in eg a 125W 2TF converter......
A single output gate drive transformer is surely more likely to have less leakage inductance.
(ie less leakage for cheaper)
Also, with double drive, arent the leakages likely to be different anyway...so delays not equal?
Also, the CDS is very low at high voltage......eg cds of STF18N60M is very low....even at low voltage.

https://www.st.com/resource/en/datasheet/stf18n60m2.pdf

We can also do stuff like "Joint footprinting" where we can use either a dual or single output pulse transformer, (eg pllel the dual outputs to give a single output) and therefore have more second sources.

 

[cupoftea]

Thanks, i confess that "if" there are advantages to using a pulse transformer gate drive for top and bottom fets of the 2 tran forward, then those "advantages" cannot be depicted in LTspice simulation, having tried to do so.
It appears that lowest losses of all, comes from doing direct drive for the bottom fet, and turning both top and bottom fets off as quickly as possible, and having a RCD turn off snubber on both fets.
Realistically, the leakage inductance of the gate drive(s) when using top and bottom pulse transformer drive, is going to be greater then if you just do pulse tran drive for top fet only........unless of course you go expensive on a "special" dual coil pulse transformer. -Questionable whether this is worth it though.

--- Updated Feb 3, 2024 ---

The situation of turn-on, with direct driven lower fet discharging both Cds voltages doesnt happen on the sim, -cant be made to happen.
It collapses its own "Half" of the bus volts, and the top fet only has to collapse half of the bus volts at turn on, even though it turned on later.

--- Updated Feb 3, 2024 ---

Also, at turn off, they both interupt the full bus volts at turn off, even though the top one turns off later