MOSFET Overheated?

Thread Starter

BlackMelon

Joined Mar 19, 2015
168
Hello,

I was driving MOSFETs in the H-bridge configuration without a load. The MOSFETs were kind of hot to my hand. So, what I did was just to connect only half of the bridge and used the same gate driver as in the H-bridge. Please refer to the attached file. The circuit to the right is the turn-on delay block, having a level-shift zener diode and 1N4148 to speed up discharging the gate pin. I wanted to know if there was supply shoot-through or not and if the heat I got previously was acceptable. I ran the circuit. The fuse did not break. No explosion, No smoke. Right after I shutdown the circuit, I touched the MOSFETs and felt the same heat (80 degree celcius or 176 fahrenheit I guess). The MOSFET package is TO-247.

So I would like to know if this is normal or not?

Thank You
BlackMelon
 

Attachments

sagor

Joined Mar 10, 2019
903
First of all, those MOSFETs (2SK3827?) are only rated 100V maximum, so even 2 in series with 220V, you are exceeding their ratings.
Where is the load? All we see is a short circuit from 220VDC to ground (assuming both are conducting at a point in time). However, you show inverted coils in your diagram, so that may mean there is never a full path to ground is both are not conducting.
A +24V to -24V swing thru a 16:8 will produce a +12V to -12V swing, which should be ok for the maximum gate voltages. Have you checked with a scope what the gate voltages actually are?
 

Thread Starter

BlackMelon

Joined Mar 19, 2015
168
Where is the load? All we see is a short circuit from 220VDC to ground (assuming both are conducting at a point in time). However, you show inverted coils in your diagram, so that may mean there is never a full path to ground is both are not conducting.
First of all, thank you for warning me about the rating. About the load, yes the load is missing, and the reason is I just wanted to know if my gate driver malfunctioned and caused shoot-through (from 220Vdc to ground) or not. Actually, these two MOSFETs belonged to an H-bridge. I ran the H-bridge without a load and noticed the same heat. So first, I inspects just only this half.


Have you checked with a scope what the gate voltages actually are?
I checked the Vgs directly. It swings from -17V to +7V. I think it makes sense, since we got a zener diode to add DC offsets of the square wave.


You need 300V transistors.
You must turn off before you turn on!
That is how the turn-on delay circuit works. There are 1N4148 diodes shorting across 10 Ohm resistor to accelerate the discharge rate of the gate pin when we turn that mosfet off. For the one we turn on, 1N4148 will not conduct, so the gate needs to charge through the 10 Ohm resistor. So in a time instant, the one getting turned on will be slower than the one getting turned off. I expected not shoot-through from supply to ground and no excessive heat at the MOSFETs. But yet, there is some heat.
 
Top