Gate Driver using TC4420

[mundachakka]

Hi. I am trying out an PFC converter which uses a back-to-back connected MOSFETs in common source configuration. The switching frequency has been selected as 50kHz. I can chose FCH041N60F or IRFP460 or 20N60 (available in lab). I was wondering if there is any chance for me to drive these using TC4420. Out of these 3, which would be best suited for TC4420.
Also, could i drive both the FETs using an single driver?
Thank you.

[fourtytwo42]

Power dissipation will be a function of static & dynamic losses. You need to compare the mosfets rdson (static) and gate capacitance (dynamic) to get a handle on these. Once you have the gate capacitance you then choose the driver to be able to turn on/off the device in the desired time to limit dynamic losses. Of course  TC4420 can drive multiple mosfets in theory but whether the consequent dynamic losses are practically acceptable is for you to decide.

[DavidAlfa]

These drivers are pretty capable, add several decoupling caps very close to the V+ pins, they can switch several amps, so i.e. 10uf+100nF+1nF.
Poor decoupling will make switching slower, causing a lot of heat.
I've run these close to 1MHz just for testing.
It ran pretty hot, completely normal as it was wasting a lot of  energy in the rising/falling edges.
At 100KHz it did fine, barely warm to touch.

Check the gate capacitace, the lower the better, faster switching, lower losses, however normally the capacitance increases as you pick higher VDS parts.

I would test them it in a small protoboard, with 2.2-10ohm gate resistor to prevent ringing.
Ideally, a gate transformer will do a lot better, as it turns off the gate with negative voltage, making it less sensitive to ringing.

[analityk]

What does it mean it this mosfet are connected back to back? Drain to drain or source to source? In the second case you can drive it with tc4420 but you have to to use gate driver transformer. It can be cheap and easy to build. I make some and achieve 14 ns rise/fall time on unconnected secondary side, driving by tc4420 through reasonable big capacitor in series with primary.
IRFP460 is a bit old transistor. I bet you can find some with smaller capacitance and drain turn on resistance.

[mundachakka]

These drivers are pretty capable, add several decoupling caps very close to the V+ pins, they can switch several amps, so i.e. 10uf+100nF+1nF.
Poor decoupling will make switching slower, causing a lot of heat.
I've run these close to 1MHz just for testing.
It ran pretty hot, completely normal as it was wasting a lot of  energy in the rising/falling edges.
At 100KHz it did fine, barely warm to touch.

Check the gate capacitace, the lower the better, faster switching, lower losses, however normally the capacitance increases as you pick higher VDS parts.

I would test them it in a small protoboard, with 2.2-10ohm gate resistor to prevent ringing.
Ideally, a gate transformer will do a lot better, as it turns off the gate with negative voltage, making it less sensitive to ringing.

I have decided to use a TC4422 driver and will test it based on the above info. Thank you.

What does it mean it this mosfet are connected back to back? Drain to drain or source to source? In the second case you can drive it with tc4420 but you have to to use gate driver transformer. It can be cheap and easy to build. I make some and achieve 14 ns rise/fall time on unconnected secondary side, driving by tc4420 through reasonable big capacitor in series with primary.
IRFP460 is a bit old transistor. I bet you can find some with smaller capacitance and drain turn on resistance.

Yes, they are connected source to source. Why is there a need to use gate transformer in such case (do you have some reference design i can look into)? Thank you.
Also, i need a wide range of duty ratio. So i don't think gate transformer would be suitable here.

[analityk]

I think it is floating potential on source of this mosfet. You can take this potential like your earth for the driver or if this is impossible (ground path can connect these sources with another potential) you should use isolated gate driver. It is a bit more complicated but isolated gate driver is more elastic for overall design and you can easily delivery signals in respect to common ground potential for all mosfet drivers. This means you can use only one isolated power supply.
You can also have to remember that mosfet including diode between drain and source so if this unipolar two transistors switch work under AC current there are many possibilities to find weird potential on sources of fets.

[mundachakka]

I think it is floating potential on source of this mosfet. You can take this potential like your earth for the driver or if this is impossible (ground path can connect these sources with another potential) you should use isolated gate driver. It is a bit more complicated but isolated gate driver is more elastic for overall design and you can easily delivery signals in respect to common ground potential for all mosfet drivers. This means you can use only one isolated power supply.
You can also have to remember that mosfet including diode between drain and source so if this unipolar two transistors switch work under AC current there are many possibilities to find weird potential on sources of fets.

Tried it out with an TC4420 driver with FCH041N60F FET. Got the converter running. Thank you all.