Hey everyone!
I'd appreciate the help with a MOSFET that I use as a discharge circuit that is burned every time.
i use the fet that on the attached file in linear mode to discharge batttery 1-150W the circuit works fine at 2A 0-60V. The fet is used in linear mode as a resistor to discharge the battery. When I set 3A 50v, the fet got damaged very quickly, and i dont understand why. I thought it might be a Spirito effect, so i slowed down the gate, and yet it got burned again. It happens at 50VDS 3A. What could be the issue? Based on the DS, everything should be fine and well beyond the limit
*i veryifed the circuit loop for a valid current loop. The current rise steadily with no osilation 0-5A the rise is 68ms
*during all testing i had a prob on the drain there was no osilations.

Thank you in advance for the help.

 

A circuit diagram might help. What are you using for a heat sink? 150 Watts is a lot of heat to dissipate!

 

A circuit diagram might help. What are you using for a heat sink? 150 Watts is a lot of heat to dissipate!

JQ1 is the connection through a connector to the fet, DAC_CC is just for the current target i increase R78 to 10k and C73 to 1uF for a slower ramp, also I added pull up of 10Mohs on R84 + input to force the initial state to be 0v at the gate
. I use a copper plate-based heatsink with thermal paste. The MOSFET burns quickly under 1s, so the heatsink has no time to react.

 

found the issue, the mosfet i used was not fit to use in linear mode at 50V for SOA graph at 50v max power was 80W

 

found the issue, the mosfet i used was not fit to use in linear mode at 50V for SOA graph at 50v max power was 80W

I wouldn't expect it to be a problem for that MOSFET. My guess is that there was an oveshoot as you have no overall current limit.
While U30 is OFF, the output of U29 will go high, and so when U30 switches on the MOSFET will be switched on at full power, and the integrator on your error amplifier will take a while to reduce it.
I would suggest using the switch to pull the input of U29 to zero, so that its output is at zero at the point the discharge is switched on.
I would also suggest using a MUCH larger resistor for R88, so that it can act as a current limit in case of MOSFET failure and not need amplifying, which could lead to instability.
Resistors are meant to get hot, MOSFETs aren't. Dissipate power in a resistor for reliability.

 

I wouldn't expect it to be a problem for that MOSFET. My guess is that there was an oveshoot as you have no overall current limit.
While U30 is OFF, the output of U29 will go high, and so when U30 switches on the MOSFET will be switched on at full power, and the integrator on your error amplifier will take a while to reduce it.
I would suggest using the switch to pull the input of U29 to zero, so that its output is at zero at the point the discharge is switched on.
I would also suggest using a MUCH larger resistor for R88, so that it can act as a current limit in case of MOSFET failure and not need amplifying, which could lead to instability.
Resistors are meant to get hot, MOSFETs aren't. Dissipate power in a resistor for reliability.

I changed the gate resistor to 1k, and the pull-up to 10Mohm forced the initial state output to zero; it still damaged the fet.
I calculated and verified what I claim, and calculatedthat the fet is not fit to work in DC conditions. The absolute max calculated power in linear mode for VDS of 50 was approximately 100W, the effect that causes it call Spirito effect