MOSFET testing

[Randall]

I have a Morgan amp in with a power scaler circuit not working, it is stuck on the lowest setting. I think it uses a N channel MOSFET to control plate voltage, but I have no schematic so I'm not sure. Anyway, I want to check this NTE 2973, and from what I read the test is negative lead on Source, postive lead on Drain should read open in diode mode, which it does. Then with negative lead still on Source, touch positive lead to Gate, where I get a beep, then back to Drain where it should show a low reading since it should be turned on. I am getting the same open reading, does this mean this MOSFET is bad? I read that the common failure is to fail shorted, but can they fail open as well?

[The Dude]

Doesn't sound good and yes they can fail open, although not all that common.

[J M Fahey]

Itīs a regular N MosFet so it must read as all other N MosFets.

I think it should read open in all combinations, because Drain to Source , both ways, is Open until Gate turns it On, and that needs some 3V on gate which you are not applying.

And Gate is insulated from other terminals by an atoms thick layer of glass, so it must also read Open always.

BUT you have a parasite diode from Source to Drain, opposite polarity (Cathode towards Drain, Anode towards Source) so thatīs the only reading you should get with a standard Multimeter, and showing some 700mV like any other Diode.

Getting less than that, down to 0 or any reading between any other leg combination may mean a dead Mosfet.

So

I am getting the same open reading, does this mean this MOSFET is bad? I read that the common failure is to fail shorted, but can they fail open as well?

actually points to a healthy MosFet.

And you donīt need much of a schematic: Power Regulators (including Power Scaling) are basically crude voltage regulators.

So basically you need:

* N channel MosFet which stands 500V or better.
* raw +V to Drain
* (poorly) regulated supply to Amplifier: from Source.
* Variable reference voltage, in the crudest form a 500k pot from +V to ground and with wiper connected to Gate.

add a few extra elements, such as gate resistors, maybe a couple capacitors, maybe a diode, not much more, my point being that you will be able to draw it by hand if you have no other option.

* rather than an open MosFet I guess itīs not getting the drive/reference voltage on its gate.
Measure at its contact or the pot wiper and check it sweeps 0V > +V .

Sometimes it canīt go lower than 60V or so because any lower amps just fart.

* They offten add a second section to the pot, to simultaneously vary bias for a smoother control. IF you have fixed bias that is.

* a few guys get fancy and avoid the second potb section and use an inverting transistior following +V and offering a negative copy to drive bias in unison.

But check whatbyou have and post a drawing, even if approximate, to know what are we dealing with.

[Randall]

I don't know, according to this test at the bottom it lists the Fluke 77 III as being suitable to turn on a MOSFET. I measure my Fluke IV at 7.25v in diode mode.

http://www.4qdtec.com/mostest.html

[J M Fahey]

To turn that Fet ON you need to apply some volts, say 4 or more, between Drain and Source, BUT ALSO some 4V between Gate and Drain; not sure how Fluke might do that.

You actually need to build a small rig for that test.

But thatīs not the point: you are testing parts instead of testing the system.

Even without further testing I sort of trust that MosFet and instead am not sure about it getting the variable drive voltage ... please test that.

[Randall]

Well. we shall see. Juan, you seem to be opposed to what I am reading and understanding how to test this. As I have posted the Fluke applies 7.25 volts. I will report once the fresh MOSFET arrives.

[J M Fahey]

Yes.

I took time to search, download and read the Fluke manual.

I found it has a simple Diode scale (which you mentioned using); with only 2 probes: Com(mon) and V/A/Diode, and works like any other Diode metering : applies a certain Current and measures Voltage drop across the junction.

That is enough to measure junctions so good to test Bipolar Transistors and Diodes, which have them.

Useless , emphasis is mine, to test MosFets to be "good" because they do not have junctions available to you.

Drain to Source and opposite way has no junctions involved, itīs a single resistive path which happens to be open unless triggered by some 3 V or higher at the gate.
Since your meter does NOT apply such voltage to Gate, while measuring Vds, itīs not meeting all required conditions to the test so itīs not being done.

Gate to Source and Gate to Drain does have a junction, but itīs isolated by a glass layer so your meter will show nothing either.

All you can measure is a backwards, parasite diode, not the MosFet itself.

IF MosFet is shorted because of abuse, meter will show that; in a nutshell: your diode test can tell you your MosFet is shorted, can not tell you if it works properly.

So far, all we know is: it does not look shorted. Nothing else.

Now from another angle: MosFet regulates down +V to weaken amplifier output in a controlled way.
Please reread my description on how it works on post#3 , after "basically you need".

IF Mosfet were shorted, youīd have full +V present at OT and tube plates all the time.

But you donīt; quite the contrary.

Since raw +V is (supposedly) present but regulated +V is not , you have 2 possibilities:
* Mosfet has failed open which is the only one you consider
* Mosfet is not getting the required control voltage at its Gate ... which you stubbornly refuse to measure, for unknown reasons.

Not only that, you take offense for me suggesting you to read a voltage.

[Dave H]

To turn that Fet ON you need to apply some volts, say 4 or more, between Drain and Source, BUT ALSO some 4V between Gate and Drain; not sure how Fluke might do that.

You actually need to build a small rig for that test.

Perhaps it would work using two DMMs? Connect DMM1 set on diode test to source and drain. Connect DMM2 to source and gate. To turn the FET on set DMM2 to diode test. To turn the FET off set DMM2 to current (200mA).

[vintagekiki]

http://electronicsbeliever.com/how-to-know-if-mosfet-is-defective/
How to Know if MOSFET is Defective

https://www.androiderode.com/how-to-test-mosfet/
How to test MOSFET

https://www.engineersgarage.com/contribution/testing-mosfet-part-1617
Testing MOSFET - (Part 16/17)

http://www.firstwatt.com/pdf/art_mos_test.pdf
Practical Mosfet Testing for Audio

[Randall]

"
Since raw +V is (supposedly) present but regulated +V is not , you have 2 possibilities:
* Mosfet has failed open which is the only one you consider
* Mosfet is not getting the required control voltage at its Gate ... which you stubbornly refuse to measure, for unknown reasons.

Not only that, you take offense for me suggesting you to read a voltage. "

Please calm down Juan, no one is stubbornly refusing, nor taking offense here. I have great respect for your knowledge and talents, but this is a practical matter for me. I do not want to build a test jig, nor write an engineering paper, I just want to test a MOSFET so I can get this amp out the door and get on to the next one. Making the measurements you suggest in this case, would mean more time and work than I really want to spend, putting things back together that were not easy to get apart, and then trying to get test leads on buried points. I have read several sites, including a few included by vintagekiki, that show an N channel MOSFET being tested with a multimeter, turning it on by touching the gate with the positive lead, but you say no, not enough voltage. I measure 7.25v on my meter in diode mode, isn't that enough? Others thinks so. Heck, I even tried the test using a battery to turn it on, but still I see open drain+/source-.

I think it is bad. When the replacement arrives I will perform the same test in the same way, and will report back. Now, I have to repair another amp and get paid.