Single MOSFET Driver IC for 2SW FWD OFFLINE CONVERTER

[tkutluz]

Hello All,

I am working on a 2 switch forward offline SMPS design with 5 V 20 Amps output and 190-230VAC input.
I tried a gate drive transformer to drive both high and low side n channel mosfets but it is very tricky and i decided to use integrated circuits designed for this purpose.
However i could not find a single IC that will drive both high and low side mosfets in this topology.There are many for driving them in bridge topologies but not for 2 SW forward.
Am i missing something or are ther really not any ICs there on the market for such a purpose.

Thanks in advance for comments
Kutlu

 

[MicrosiM]

Hello All,

I am working on a 2 switch forward offline SMPS design with 5 V 20 Amps output and 190-230VAC input.
I tried a gate drive transformer to drive both high and low side n channel mosfets but it is very tricky and i decided to use integrated circuits designed for this purpose.
However i could not find a single IC that will drive both high and low side mosfets in this topology.There are many for driving them in bridge topologies but not for 2 SW forward.
Am i missing something or are ther really not any ICs there on the market for such a purpose.

Thanks in advance for comments
Kutlu

Showing a schematic or at least a block diagram, will help a lot to understand what are you doing.

 

[km-r]

ive opened up a Cougar ATX power supply uses a UC3843 on two-switch forward design. uses a small GDT+transistor for the high side switch. if you want to avoid transformer winding maybe use a bootstrapped buffer for the high side switch?

 

[tkutluz]

ive opened up a Cougar ATX power supply uses a UC3843 on two-switch forward design. uses a small GDT+transistor for the high side switch. if you want to avoid transformer winding maybe use a bootstrapped buffer for the high side switch?

My concern is that bootstraped circuit will not work with the two switch forward smps as the bs capacitor will not charge since both switches are either open or closed.
If this is not a concern than IC s including bs capacitor may be used.But has anyone tried this before.

Regards,
Kutlu

 

[blasphemy000]

This is the basic design of what he is building here. I don't think there are driver ICs made that can drive the FETs in this kind of topology, I could be wrong though. Both FETs are switched on/off together so I don't think that the bootstrap capacitor with an IC like IR2110 would charge. I've designed a few supplies using this topology with IGBTs and driven them with a gate drive transformer with a wave-shaping circuit between the GDT and the IGBTs that allow the gates to be driven just slightly negative for faster turn off. One way you could drive these with an IC would be to use two opto-isolated low-side driver ICs. Use one on the low-side switch as normal. For the high-side switch though, you would need a separate, floating, power supply to drive it. This would add more complexity to your project, but small DC-DC converters can be made or bought, that include transformer isolation, for very cheap and a small footprint. GDT overall is the most simple design for this topology. I'm heading to work right now but if you're more interested in the isolated supply method of driving the high-side switch, let me know, and I can include a basic drawing of how to set it up. The FOD3180 is a fairly good opto-isolated driver IC. Tr/Tf for it are 75ns/55ns respectively and it is capable of delivering 2A of drive current to the FETs. The propagation delay of this IC is between 100-200ns, but that isn't that important in this topology since both switches change states together.

 

[tkutluz]

Yes i am interested in an isolated design method and will appreciate if you could share your experience.
Thanks in advance,

 

[KX36]

Just about any high voltage high+low side gate driver IC with separate inputs and outputs rather than a single input. Just tie the input logic pins together.
Off the top of my head, IR2110/IR2113 should work. They're quite ubiquitous. Gate driver ICs tend to be relatively expensive and often only come in surface mount packages but they certainly have a place.

I don't think there should be a problem with the bootstrap circuit on a 2 switch forward converter at least in a steady state. All you need for it to work is the source of the high side FET to come sufficiently below the low voltage power rail for some significant time every cycle. This happens when the transformer resets and the upper FET's source swings down to ground, clamped by the reset diode. The reset time should be the same as the FET's on time, so as long as the duty cycle isn't too low it should be fine.

Not entirely sure how the thing would behave at startup though.

 

[KX36]

Looking into this, it looks like it probably would work even if it has to shudder into life over the first few cycles when the bootstrap cap isn't charged and the upper FET doesn't turn completely on.
ICs designed for this often include a low side FET in parallel with the reset diode which is turned on whenever the main FET is turned off so that at low duty cycle, even when the reset diode stops conducting the bootstrap capacitor can still charge sufficiently. (e.g. LTC3705 is this topology at a lower input voltage, see also SiLabs AN622)

It probably depends on the lowest expected duty cycle. I'd hope that with your relatively narrow 190-230Vac input voltage range, it wouldn't be an issue as the duty cycle shouldn't have to go too low.

Gate drive transformers have issues with duty cycle too as it changes the max and min voltages seen by the FET gate, so either way it gets slightly messy. Again, this hopefully wouldn't be an issue with your input voltage range.

Eagle-eyed readers may note that the reset diode appears to be the wrong polarity for current to conduct through the bootstrap capacitor, but as long as it's forward biased by a bigger transformer reset current it will still be able to conduct the bootstrap current.

Here's a short post of mine about my GDT driven 2sw forward.
http://www.diysmps.com/forums/showt...ansformer-Issues&p=11985&viewfull=1#post11985

 

[tkutluz]

http://www.irf.com/technical-info/appnotes/an-978.pdf

I ve seen above app.note from IRF and seems it is possible to drive the transformer in 2 switch forward configuration with additional 2 diodes and 2 mosfets on page 24 under figure 26 by an IR 2110.

To overcome charging issues of the cap they used 2 mosfet. Capacitor is grounded via Q1.This is the most explanatory document i ve found so far for such a configuration.
I think i will give a try.

Thanks so far for the comments.

 

[KX36]

So Q2 and R form a single transistor inverter/NOT gate so that whenever the main FETs are off, Q1 is on and vice versa. It seems reasonable, the only potential problem is the risk of shoot-through if Q1 doesn't turn off fast enough or turns on too early with respect to the upper switching FET. I don't think that'd be very likely to be a problem as Q2/R can sink Q1's gate charge quite well but can't source much current through R so it should mean a slow turn on, fast turn off, but I also wouldn't like to say how well it would work without prototyping it.

D1 would be there anyway in the bootstrap circuit, so the addition is actually 2 small FETs a diode D2 and a resistor. Q1 Vds would have to be rated for the high voltage but relatively low current (transformer magnetising current, I think), Q2 would only have to be a low voltage low current signal transistor. D2 probably isn't necessary either, I could speculate that it's there to discourage Q1 body diode conduction during reset and allow the faster reset diode to do that at least until Q1 turns on. I'd trust the app note schematic from IRF as much as any app note schematic (I've seen plenty that work in theory but not practice) so I would include it.