Why does this MOSFET keep burning out?

Thread Starter

Involute

Joined Mar 23, 2008
106
I'm trying to drive a solenoid with one of these MOSFETs. The circuit is below. I've blown three of my four MOSFETs (sometimes immediately, sometimes after a handful of cycles) as I become increasingly careful with my test setup. Before I try the last one I thought I'd check here to see what I'm doing wrong. Maybe it's not my testing but rather an inappropriate MOSFET for the application? Wrong circuit? Tips appreciated.

Solenoid circuit.jpg
 

crutschow

Joined Mar 14, 2008
34,281
What voltage source is driving the gate?
It needs to have a low impedance to rapidly turn the transistor on and avoid high dissipation during that time.
 

ArakelTheDragon

Joined Nov 18, 2016
1,362
The power dissipation is "0.76 to 1.26W". The transistor has "6Ohms" resistance at "3.3VDC" at the gate. Can you measure the current (by measure I mean simulate it) do not risk your life with a real measurement.
 

Thread Starter

Involute

Joined Mar 23, 2008
106
Likely due to excess Power Dissipation. Are you sure that the MOSFET is driven to saturation?
I’m driving the gate with a 5V pulse straight from the signal generator. If I’m reading the data sheet correctly, that should be more than enough to drive the thing to saturation.
 

Thread Starter

Involute

Joined Mar 23, 2008
106
What voltage source is driving the gate?
It needs to have a low impedance to rapidly turn the transistor on and avoid high dissipation during that time.
A single 400 msec. pulse straight from the signal generator. It takes it about 2 msec. to ramp from 0 to 5V, and another 2 msec. to ramp back down.
 

Thread Starter

Involute

Joined Mar 23, 2008
106
The power dissipation is "0.76 to 1.26W". The transistor has "6Ohms" resistance at "3.3VDC" at the gate. Can you measure the current (by measure I mean simulate it) do not risk your life with a real measurement.
I’m afraid I don’t know how to use a simulator like LTSpice, but I’m driving the gate with 5V, so the resistance should be 3.5 ohms. At 162 mA, there should be .57V across the MOSFET, for a power dissipation of only .092W.
 

bertus

Joined Apr 5, 2008
22,270
Hello,
A single 400 msec. pulse straight from the signal generator. It takes it about 2 msec. to ramp from 0 to 5V, and another 2 msec. to ramp back down.
The rampup and rampdown time may be to slow to avoid heating of the mosfet.
You could try to make a circuit using a NE555 as pulse generator.

Bertus
 

Thread Starter

Involute

Joined Mar 23, 2008
106
So I rigged up a 555 as a one-shot to generate a 400 msec. pulse. The rise and fall time are about 2 usec. I tested the MOSFET with 12V through a 10K resistor and it worked fine. At 180V and the solenoid, poof. I just ordered some more, and some different ones with similar specs. It's infuriating because these things look perfect on paper.
 

ebp

Joined Feb 8, 2018
2,332
That FET has no avalanche rating which is rather unusual these days. It may be just because it is a tiny die and they'd be embarrassed by the correspondingly tiny avalanche rating. This is a rating for the amount of energy that it can handle when the body diode is made to go into reverse breakdown and conduct like a zener diode. It may be that in spite of the diode across the coil you are still getting a voltage spike high enough to kill the FET.

The 1N400x series diodes are slow. Normally what is considered is "reverse recovery time" which is unlikely an issue in this circuit (though it could allow a narrow very high current spike at FET turn-on), but slow diodes usually also are fairly slow to begin to conduct (so-called "forward recovery"). A UF4004 ultrafast diode might be better (don't go with the higher voltage parts - they are slower).

Another more certain option, though a bit more expensive, is to use a zener or transient suppressor diode (a zener designed for high transient energy) right across the FET. I'd use something rated at about 200 V to give some margin from the 180 V supply but still keep the peak clamping voltage below the FET rating. This might be worth trying experimentally if for nothing other than gaining an additional clue. If the FET still fails with such a clamp in place, something else is killing it. Keep the path between the FET terminals and the TVS or zener as short as possible. I'd still keep the diode across the solenoid to handle the bulk of the energy so the TVS can be "small."

If the solenoid is located at a distance from the FET, be sure the connecting wires are twisted or otherwise arranged as a pair. The same goes for the wires from the source of the FET and the positive to the solenoid from near the FET to the power supply. You can sometimes get some nasty high voltage ringing that can cause misoperation and even damage. What you want to do is kept "out" and "return" paths for current close together so their opposing magnetic fields will partially cancel. This reduces the inductance and hence the stored energy that can cause problems.

Your original slow rise and fall would have gone a very long way to eliminating some of the issues mentioned above. The FET might, however, have been being killed by excessive transient power dissipation.
 
Last edited:

crutschow

Joined Mar 14, 2008
34,281
How long is the wire between the solenoid and the MOSFET?

Can you look at the drain with an oscilloscope to see if there are any voltage spikes during turnoff?
 

Thread Starter

Involute

Joined Mar 23, 2008
106
How long is the wire between the solenoid and the MOSFET?

Can you look at the drain with an oscilloscope to see if there are any voltage spikes during turnoff?
The wire’s about 12” long (as long as it will be in the application). I can scope the drain when I get the new parts in. In fact, I was scoping it when it blew up the last time but I didn’t realize it had blown until I reset the scope for another test, so I didn’t save the screen. I’ll be more careful next time.
 

ArakelTheDragon

Joined Nov 18, 2016
1,362
I’m afraid I don’t know how to use a simulator like LTSpice, but I’m driving the gate with 5V, so the resistance should be 3.5 ohms. At 162 mA, there should be .57V across the MOSFET, for a power dissipation of only .092W.
Now that you have the calculation, like suggested before, its good to take the real measurements of the voltage and current and see if it meats your expectations.
 

PhilTilson

Joined Nov 29, 2009
131
It may sound a silly question, but are you QUITE SURE that the freewheeling diode is connected the right way round?!

If it isn't, that would be a very good way to fry multiple MOSFETs!
 

rphare

Joined Nov 20, 2015
11
And another silly question: Are you sure about the solenoid resistance? Can you measure it? Where did the suspicious value of "1111" come from?
 
Top