overheating transistors

[MsStateDawg]

I have a full-bridge inverter circuit that will output a 340V peak-peak 60Hz sine wave. The problem is that the MOSFET transistors (IRFP460's) are getting extremely hot with a 20k-ohm load resistor. I was thinking that the dead time between switching legs was not enough. After taking a closer look, the measured dead time is about 260ns. The Trise and Tfall times for the transistors are 59ns and 58ns, respectively. That lead me to believe that the dead times were sufficient. The two FETs of of each leg of the bridge are being driven by IR2181 high/low side gate drivers. I chose these to provide isloation for the high side transistors. Could the problem be in sizing the bootstrap capacitor on the IR2181 gate driver?

 

[Styx]

mmm,

not likely to be the bootstrap caps. All they will do is act as a charge tank. Too small and it might slow down the turn-on of the FET's since caps have been discharged too much, too big and you might have problems of inrush at turn-on if your PSU to the electronics is low-current limit.

A transistor will only get hot if it has to dissipate power. Either conduction losses or switching losses. Conduction losses will be low if the device is driven hard into saturation (does yr driver provide enough gate voltes to drive the FET into saturation). Switching losses will be high if you are switching the FET too fast thus the rms switching power due to high volts and high amps at turn-on and turn-off will be high.

However, you say you have a 20k load???? that will be next to no drain current in the FETS.

Instructing to turn both devices of a leg on is not the only way to get a shoot-through. parasitic capacitance can turn a FET on as well.
If you are not holding off (ie applying negative gate voltage) your bottom FET, after the dead-time and the top FET switches on the millar capacitance of the bottom FET will conduct that dv/dt and try to turn your bottom FET on.

do you provide negative gate volts to yr FETS? I looked a the FET datasheet but couldn't find any millar cap values.

 

[Optikon]

I have a full-bridge inverter circuit that will output a 340V peak-peak 60Hz sine wave. The problem is that the MOSFET transistors (IRFP460's) are getting extremely hot with a 20k-ohm load resistor. I was thinking that the dead time between switching legs was not enough. After taking a closer look, the measured dead time is about 260ns. The Trise and Tfall times for the transistors are 59ns and 58ns, respectively. That lead me to believe that the dead times were sufficient. The two FETs of of each leg of the bridge are being driven by IR2181 high/low side gate drivers. I chose these to provide isloation for the high side transistors. Could the problem be in sizing the bootstrap capacitor on the IR2181 gate driver?

Before suspecting the more complicated parts of your circuit (gate drives etc..) See if the power dissipation makes sense or not. Sure, they are hot but are they supposed to be? Find the power dissipation and then use the thermal resistances to calculate the temperature rise.

The 20kOhm load means that you expect the currents to be somewhat low but if the voltage across the fet at that time is very high, you still have alot of power dissipation. Check basic dissipation first to see if it is correct or not. It will save you time hunting for a problem that might not exist.

 

[Styx]

yep that is one thing forgot to post. what you class as "hot"