Wierd science and testing Mosfets

[Zapped]

Here is a situation which has me baffled. I'm (still) testing some IRF231's that were to go in my A-75 but experienced reverse supply voltage during initial fire up. For control purposes, I also have tested an IRF231 which was not so abused (never in the amp) , but only previously tested for Vgs.

If I connect the Q in the Vgs matching circuit in the A-75 article BEFORE turning on the 15v testing voltage, I get a Vgs reading in the 6-6.75 volt range (too high), and my supply indicates about 221 ma of current at about 13.7 volts, using 2 100ohm R's in parallel.

And I get the same results when I make the connections with voltage ALREADY on and using either of two sequential orders where the G connection is made LAST (ie, the order is DSG or SDG): namely about 6.6.75 volts (too high), and 221 ma current, at 13.7 volts

BUT, If I make connections with voltage ALREADY on and any other sequential connection order (GDS, GSD, SGD, SDG, etc) I get a "correct reading of around 4.25 Vgs, at about 150ma current and 15 volts.

These results are consistent whether I am testing the "abused" Qs from my A-75 or a totally "unmolested" Q never in the amp. WHY does the order matter? And what can I conclude about damage, if any. This is driving me nuts.

Larry Wright
Seattle area

[Nelson Pass]

It is unlikely that you abused the Mosfet by reversing
the voltage in the test rig.

Maybe you should get an Analog meter.

[Zapped]

Mr. Pass: thanks for responding. The reverse voltage mistake actually occured while testing the output stages in the amp itself, but in any case I don't understand how I can get differing results depending on the order in which I make connections (or apply voltage) in your Vgs test. And why would an analog meter have a bearing? (I have one, for that matter, I can dust if off and use it, but I am not sure what I am looking for.)

Larry Wright
Seattle area

[Circlotron]

Maybe the fet is breaking into oscillation? Put a scope on it if you can and have a look-see. Failing that, put a couple of microfarads from gate to source and try it again.

[Nelson Pass]

Mosfets can oscillate, which is why we often still use a Gate
resistor on the test rig.

But that doesn't explain the voltages present with the power
supply removed. Or am I missing something?

[Zapped]

Going back over my lengthy description I can see that its a good thing I am not a technical writer, I'd get fired! Let me try to be more clear than I was the first time and correct a mistatement about supply current as well. Resistor in test rig = 50 ohm at 2 watt.

Three scenarios:

A.
1. turn on power supply at 15 volts, THEN
2. make connections to Q in following sequential order: SDG or DSG
3. measured results: Vgs = 6.25 to 6.75 (too high)
4. current draw on supply = 150ma
5. supply voltage shows draw down to 13.7 v

B.
1. make connections to Q in any order, THEN
2. turn on power supply at 15 volts
3. measured results: Vgs = 6.25 to 6.75 (too high)
4. current draw on supply = 150ma
5. supply voltage shows draw down to 13.7 v

C.
1. turn on power supply at 15 volts, THEN
2. make connections to Q in ANY sequential order where G is NOT last connection made (e.g. , GDS,GSD,DGS etc.)
3. measured results: Vgs = 4.18 to 4.25 (normal)
4. current draw on supply=221 ma
5. supply voltage shows: 15 volts

Weird science or occillation? Didn't get these results when I first matched my Q's when building the amp.

Circlotron: I will try the cap. idea.
Mr. Pass: since the output stages use a gate resistor should I try the test rig with the same Ohm value in series with gate?

Thanks to you both for your help.

Larry Wright
Seattle area

[The Seven]

Hi Larry,

It seems that the "current draw on supply" in A and B should be "210 mA" AND that in C should be "150 mA" according to your first post. The second post shows the opposed.

Which post is correct?

I would concure with Mr Pass' suggestion that a GATE resistor should used in the test to avoid any possible occilation. Further, a local by-pass capacitor (say 0.1uF) should be used to by-pass the 15V supply.

Regards,
The Seven.

[Zapped]

Good eye, Seven! You are correct about the difference in posts.

The second post is correct. I realized last night while repeating the tests that I had described results incorrectly in yesterday's post. (Two much data and too few remaining braincells.)

I will see what happens when I try the suggestions for gate R and the capacitor.

Thank you for your help.

Larry Wright
Seattle area