High voltage half Bridge mosfet problem.

Hello ,

Are there anybody experienced at high voltage half bridge ? please check my circuit, the mosfets( IRF830) are very hot even the half bridge not connect to any load. what maybe the reason? please give tips or suggestions, thanks a lot.

Volt: V400 =400v
U2: I R 2 3 04
switch frequency: 30khz - 100khz.

 

Thanks Bertus.

driver chip U2: I R 2 3 04
PB0, PB1 PWM pulse with dead time 1 us.

 

Datasheet for IR2304: http://www.irf.com/product-info/datasheets/data/ir2304.pdf
Max sink current is 130mA.
Total gate charge for the IRF830 is 38nC.
Turn-off delay time is 42nS.

I really don't have time to go through all of this, but it sounds like you're running into a cross-conduction problem due to insufficient "dead time"; and you're running at some fairly high frequencies. You may need a more capable gate driver IC.

Have a look at this application note for high-speed MOSFET drivers:
http://focus.ti.com/lit/ml/slup169/slup169.pdf
It will be well worth your time to read it through.

 

Maybe you can try to turn 1 mosfet gate off first. remove gate voltage. Then see if the other still gets hot. If it does then it is not due to cross conduction.. If it does then cross conduction is highly likely. Try also the other mosfet. So you can troubleshoot the circuit somehow?

 

Thanks a lot , could you please recommend some proper drivers or mosfets for this occassion? thanks again.

 

Sorry, I'm about ready to leave for a couple of days. It's a long drive home.

 

Thank you.
It seems not the cross conduction problem, 1 us deadtime is enough. I ve tried to connect a 1K ohm 1/4w resistor between V400 and "D" of Q2(H mosfet), I test the resistor not too hot but mos still very hot.

 

Have a good trip, friend.

 

hello all, the mosfet still hot when Vds=400v, F=100khz, I ve changed driver to FAN7382 and mosfet to FCP4N60, and I ve changed circuit to following schematic: Rgate=22 ohm, Don= SR240 , Qoff = 8550(1.5A)

would somebody give me some suggestions to cool the mosfet? thanks?

 

In the first circuit, remove CV7, DV7, CV4 and DV4, place DV2 where D8 is and then try.

 

Thanks Tahmid, I ve bandoned CV7, DV7, CV4 and DV4( for the purpose of produce negative voltage to ensure the mosfet shut) and take the new circuit, please check schematic of my last post , the mosfets seems cooler but still hot.

 

If you are using that for low side drive it seems ok and DOn and QOff don't seem necessary as the driver itself has internal totem-pole transistors for driving the MOSFET.
You can also try with IR2110/IR2113 from IR or L6385/L6386/L6387 from ST. I've used these drivers myself for high/low side MOSFET driving.
Hope I could help.

 

Hello Tahmid, Thanks for your message.
Don and Qoff make the mosfet to shut more quickly.
Have you tried L6385 at 100khz and under 400v voltage? what is the temperature of mosfet?
Also I wonder if L6385 support input logic of 3.3v?

Regards,

 

Hi,
When I used any of the drivers mentioned above (I used upto 50khz, didn't try any higher 'coz i didn't need to), I could use upto 10-15W without any heatsink or the heating of the MOSFETs. It would get pretty hot at 100W and above but not THAT hot.
Could you tell me how much power you are using as your load? Maybe you could try using a MOSFET with a lower RDs(on).
3.3v? The datasheet states a minimum of 3.6v.

Tahmid.

 

I had not connect circuit to any load( no power consume) the mosfet already hot. that is what I worry about.

 

That is the problem. This is a half bridge converter. So if there is no load the bootstrap capacitor for the high side MOSFET does not get charged (as the cap charges through the load to gnd) so the MOSFETs heat up. Connect a load and then check. It will probably be fine then.
Good luck.

 

bootstrap capacitor charged by diode byv26C

 

any new ideas?

 

Hi iamhere,
In the diagram attached, I've shown two configurations for halfbridge circuit. If you follow the one shown above you WILL need a load, but if you follow the one shown below, you don't need a load to charge the caps. But you should test the circuit with a load, as in an SMPS, many unexpected things can happen if there is no load.
With thanks.
Tahmid.