MOSFET multi-meter test accuracy

[mikec]

Hi,

There are a couple of standard tests for MOSTFETs that use a multi-meter. For example:

Using the multi-meter in DIODE test, turn on the MOSFET and measure the drain-source resistance. Turn of the MOSFET by shorting gate-source and measure drain-source resistance again.

Other tests are generally checks for shorts/opens check body-diode etc, the first test above however seems to cover the actual functionality of the MOSFET.

My question is using the test above if I verify that the MOSFET actually switches ON and OFF (and other basic tests like body-diode ok), can I conclude that the MOSFET in any switching application is OK and should be discounted as the problem in a faulty circuit/device?


I currently repairing a circuit with 8 K2837 power MOSFETs (Transformer coupled Gate Drive circuit for an Inverter welder), 4 MOSFETs are split into 2 pair (paralled) for the HIGH-SIDE and the remaining 4 are split into 2 pair for the LOW-SIDE switching. At the moment I suspected a short-circuit is occurring at some point in the switching cycle (i.e. related to the MOSFETs), the control circuit seems to shut down and restart. Instead of probing the MOSFETs, I went with using my FLIR camera to watch the heating of the MOSFETs (didn't feel like probing the high-voltage circuits  ) and this did seem to point to unexpected behavior. 1 pair of the HIGH-SIDE got warm and the other pair didn't warm up, on the LOW-SIDE only 1 MOSFET of each pair warmed-up.

I may have convinced myself from the thermal image that the MOSFETs must be damaged in someway and the basic MOSFET test above can't detect the issue. I am wondering are there some important measurements of the MOSFET for this particular application that are missed by the basic MOSFET test?

[Manul]

If all mosfets are measured to switch correctly with a test circuit and multimeter, I would say it is unlikely, that they are the problem. Mainly because in high energy circuits any obscure failure would very likely escalate to catastrophic. You can not conclude this, but you can guess with high probability.

I would suspect driver problem, possible shoot-through or some other problems down the line.

[mikec]

Yes, I had kinda figured it was unlikely a failure that leads to something catastrophic, its just enough to pull the power down on the control and drive circuits.

The shoot-through would be where a high-side FET has not turned off fully before a low-side FET has turned on, resulting in an almost DC short circuit?

I guess in the end the MOSFET tests are pretty decent test, I'll assume for now MOSFETs are good and take a look at the drive circuits (which so far are looking pretty alright as well  )