Driver IC for two transistor flyback?

[grizedale]

Hello,

I wish to do an offline, 230VAC input, two transistor flyback converter.

Why are there no specific high and low side gate drivers that cater for this?

-I mean the type that have a capacitive charge pump high-side driver on them.


There are planty of high and low side driver ICs for half bridges, but not two transistor flybacks.

 

[FvM]

Can you please help me to imagine what a two transistor flyback converter topology looks like? I never came across it.

 

[ea.arun]

never heard of such a configuration.....
do you mean push-pull or something like?

 

[grizedale]

2 tran flyback is just like the primary side of a 2 tran forward converter......but then with the flyback secondary

 

[FvM]

I don't know what a "flyback secondary" is, but driving the primary in a push-pull way makes the converter act as forward converter.

 

[grizedale]

2 switch flyback:

**broken link removed**

-But i can't see a chip like this for running off the PFC bus (400v)

 

[Orson Cart]

Two transistor flybacks are very common for high input volts/ wide range inputs (similarly for 2 xtor forward), usually a gate drive transformer is used to turn the xtors on and off simultaneously, Regards, Orson Cart.

 

[grizedale]

Thanks,

--the problem is that gate drive transformers at SMPS frequencies are very expensive.

-they have to be custom wound to have low leakage inductance.

-its a shame that there are no controllers with charge pump high side drives, meant for driving high-side FETs hanging off 400V PFC bus's

 

[Orson Cart]

If you get a tri-filar wound toroid in volume from Taiwan etc the volume price is pretty good - Regards, Orson Cart.

 

[FvM]

Thank you for pointing to the LM5015 example. I admit, that I wasn't aware of the two-transistor series circuit used for flyback converters, although some advantages are obvious. Wide range operation seems to be limited compared to the standard single switch flyback topology, however.

For the driver problem, I think that a combination of a standard current mode controller with a bootstrap high-side gate driver should work.

 

[Orson Cart]

Unfortunately when you look at the detail of the 2 sw f/back operation - std bootstrap is not possible for the high side device as when the bottom device is on - so is the high side - when both devices turn off there are substantial volts across the high side device (& the low side) either way there is no gnd reference to allow a charge of a bootstrap cap.
2 sw f/back is often used with 600V or 800V mosfets for aux supplies off a rectified 3 phase line (565 - 750VDC) where a single switch 1kV mosfet would be a bit short of volt headroom (also the lower volt devices are cheaper) - Regards, Orson Cart.

 

[FvM]

At least in operation, the bootstrap capacitor should be charged during diode conduction time. But I agree, that an auxilary circuit to precharge the bootstrap capacitor can be required. I assume, that activating the low side switch without the HS would be sufficient in most cases.

 

[ea.arun]

i was not aware of such a configuration....

this is really advantageous considering the mosfet selection and savings from automatic clamping of the leakage...

 

[Orson Cart]

At least in operation, the bootstrap capacitor should be charged during diode conduction time

As with a normal flyback - the intrinsic diodes in the mosfets of a 2 sw f/back do not conduct during operation.

 

[FvM]

As with a normal flyback - the intrinsic diodes in the mosfets of a 2 sw f/back do not conduct during operation.

I didn't talk about intrinsic diodes rather than the primary clamping diodes. They can be expected to conduct due to leakage inductance, even if the transformed output flyback voltage is considerably lower than the bus voltage.

A more general approach to force recharging of the bootstrap capacitor would be to introduce a lead time in activating the low side switch.

 

[Orson Cart]

If you have clamp diodes from the primary Tx ends to the rails they need to be properly ON for a certain time (~1uS or possibly more) each cycle to keep the bootstrap cap sufficiently charged - at low output power or low output volts (e.g. under current limit) you would need a fair bit of leakage in the Tx to ensure this (requiring big snubbers on the o/p diodes at full power), also such clamp diodes can only be used where you have fixed maximum o/p voltage and a fairly narrow range of input volts - otherwise the primary clamp voltage (the input rail) will limit the max o/p volts available - and considerable power will cycle between the Tx and the input bus.
Usually a 2 switch is used where it is desirable to have voltage sharing across the devices - which means it may not be possible to turn one on first - Regards, Orson Cart.