MOSFET help

Hi Guys,

I like to tinker with fixing computers, which seems to often be ripping out a psu and putting a new one in or swapping ram etc. Replacing caps etc I want to be able to test more at the component level and MOSFETs seem to come up a lot.

So far the best description of a MOSFET is what I have seen at

formatting link
- although maybe on a very simple level

I have read the sci.electronicsrepair faq on MOSFET testing but that has me confused as : (a) the tests are all different (b) some of them tell you to put a lead here and a lead there, but not which lead on your DMM (c) I brought a new MOSFET to experiment with - yet I still dont get the same results as some of the tests????

So I have some questions that I hope the experts here can clear up. They all relate to MOSFETs as used in computer PSUs and motherboards

1) Can you test them in circuit 2) I see some are though hole mounted and some are, "surface mounted".Is there a special word to describe teh surface mounted ones? Why are they surface mounted (like lieing flat on PCB) 3) I have been thinking of MOSFETs as a current gate that is turned off/on by voltage - so with respect to all the MOSFETS you see on a computer motherboard near the VRM circuitry - that are they actually doing? how does a bunch of gates help?

Any help apreciated in trying to get my head around these things

Cheers

-Al

Reply to
Al
Loading thread data ...

You can do a lot with a battery, LED and resistor.

If you apply reverse voltage from drain to source, the "body diode" will conduct and light the LED. (can be useful to identify leadout on small TO92 MOSFETs).

Swap the leads around so that the correct polarity is applied to the MOSFET and the degree of conduction will depend on any static charge on the insulated gate.

Keep the negative lead on the source and touch the positive lead to the gate to charge the gate capacitance positive, this will make the channel conduct and light the LED.

Keep the positive lead on the drain and touch the negative lead on the gate to charge the gate capacitance negative, then any channel conduction means a faulty MOSFET.

HTH.

Reply to
ian field

Crude basic test for checking MOS Fets , in isolation The following is for n type mosfets and results do vary so it is probably best to try on known good matching FET before falsely discarding a suspect one of same type number. Starting by identifying G,D,S if unknown and assuming starting in non-conducting state. Using a DVM on "diode" check with red and black probes. One pair of terminals only one way round should show a forward voltage x . The black is on the drain (D) the red is on the source (S) leaving the third as the gate. Put red to G and black to D then swap probes to D-S . There should be conductance less than x slowly increasing back to x in the limit. Putting black on the G and red to D then should return D-S to full x value. Or setting the D-S "voltage" to the low setting and just touching the G may reset to the higher "voltage" state.

-- Diverse Devices, Southampton, England electronic hints and repair briefs , schematics/manuals list on

formatting link

Reply to
N_Cook

So this test must be done out of circuit?

on

formatting link

Thanks - will try this.

Incidentially what do you use for MOSFET removal from PCB? The type of MOSFET's I am trying to deal with are all mounted "flat" on the PCB of the motherboard.

Cheers

-Al

Reply to
Al

I have all these parts....;-)

I am pretty new to this, can you explain what you mean by reverse voltage? +ve terminal to ? and -ve to ?

I assume the resistor is just for limiting the current to the LED?

Will try that, as soon as I know what you mean by reverse volage.

It sounds like there are really no useful test that can be done in circuit?

What I have been doing is getting a lot of motherboards and in diode test mode measuring between gate and source. One way it has open circuit, the other way it has a reading of about 790 ohm. This is for the MOSFETs the are around the edge of the CPU as per here (they are

60T03H):

formatting link

All mosfets with Black on G Red on S measure 790 Ohm. With Red on G and Black on S they all measure open.

The two other MOSFETs are here, near the Vdimm for RAM. They are APM2054N:

formatting link

One them ith Black on G and Red on S it takes about 2 seconds to climb to open. With Black on S and Red on G it just measures open.

The other with Black on G and Red on S it takes about 2 seconds and climbs to 1466 Ohm. With Black on S and Red on G it just measures open.

Does the fact that these two test differently between G and S mean anything or could it be because this is being done in circuit?

Cheers

-Al

Reply to
Al

If you need that explaining, you shouldn't be messing with power supplies! - However, what do I care if you kill yourself so here goes.

Its best to use a 2mA ultra-efficient LED, the battery should be a MN21/A23

12V car keyfob battery, calculate the resistor to limit the current to suit the LED.

Assuming an N-channel MOSFET connecting the negative lead to the drain and the positive lead to the source, the body diode will conduct and light the LED.

Next, connect the positive lead to drain and negative to source - the LED may partially light due to static charge on the gate capacitance.

Keep the negative lead on the source and touch the positive lead on the gate, this charges the gate capacitance positive and makes the N-channel conduct - so if you then connect the positive lead to drain the LED will light.

Next, keep the positive lead on the drain and touch the negative lead to gate, this charges the gate capacitance negative and should make the N-channel completely non-conducting - so put the negative lead on the source and the LED should not even glow the slightest bit - if it does the MOSFET is leaky and will go bang in a PSU!

Reply to
ian field

Assuming you can get the MOSFET at a reasonable price, why not just replace it, and not worry about testing it? It isn't worth the trouble.

I mean it only as a neutral, impersonal statement when I say that you don't know much about electronics, and this sort of time-wasting trouble-shooting is not likely to help you learn.

Reply to
William Sommerwerck

ll

! -

23

it

e

rce

Hi Ian,

Thanks - and to put your mind at rest its on a motherboard, so its all DC.

I stay away from AC power supplies :-)

I assume all this must be done out of circuit?

-Al

Reply to
Al

probably

.

third

value.

on

formatting link

test out of circuit, unless comparing known good board to questionable, from the same model, even then depends on circuitry

For small ones I tend to score around and angled under a bit, with a scalpel and prize off, pad by pad , with a small jeweller's screwdriver. For larger ones ,I loop some fine nichrome wire under a couple of pins and then ( not owning a micro-jet hot air system) I shroud around the item (to protect surrounding devices) and blast with paint-stripper hot air, tugging at the wires to reduce the heating/loosening time. Blasting the device with freezer spray beforehand, if wishing to salvage undamaged from heat.

-- Diverse Devices, Southampton, England electronic hints and repair briefs , schematics/manuals list on

formatting link

Reply to
N_Cook

ll

! -

23

it

e

rce

Thanks - have done all those tests on a known good mosfet and got the expected results.

Now to try some other mosfets whose status is unknown.

Cheers

-Al

Reply to
Al

What a complete and utter load of cobblers!!!

Experimenting with how components behave in the real world is exactly how to learn.

Reply to
ian field

trouble.

I'm not sure that applies when you're trying to _fix_ something.

My experience with repairs -- which I've been doing on and off for 50 years -- has generally been that the only thing I learn when fixing things is how _stupid_ I am, for not catching the "obvious".

In my opinion, you learn more from a combination of reading books, building projects, and experimenting with known-good components.

Reply to
William Sommerwerck

I never experimented with other people's property. I have done so with my own. And learning is an ongoing process not unlike how a physician or surgeon learns as time passes. A lot of my learning came with experimenting on building circuits from the junk bin. But a lot of it also was from being mentored on the bench and ultimately from trial and error.

Reply to
Meat Plow

Which is bound to happen if you don't appreciate the importance of understanding how components behave!

Reply to
ian field

But that isn't necessarily something you learn by performing repairs.

I had a problem fixing an electronic crossover that used a op amp and current-driver chip in its output stage. There was severe DC offset at the output, and it wasn't clear whether the op amp or current driver was defective. (The circuit board was almost impossible to remove from the unit. Had it been easy, I would have simply replaced both ICs without worrying about it.)

If I'd read the spec sheet for the current-driver chip carefully, the answer would have been obvious. It was John Curl ("Jesus Christ" in the audio industry) who had to explain it to me: "That current driver has a fixed gain of -1." Aha! There was -1V on its input, and something like 5V on its output. So the current driver _had_ to be defective. *

It was. I replaced it and that was that. (The driver had only five, well-spaced pins, so I was able to snip them off, suck out what was left, and solder in a new driver without having to remove the board. I lucked out. Had the IC been bad, the board would have _had_ to have been removed, as I would not have risked "snipping and sucking" fourteen leads from the top.)

A side remark... If you own oddball equipment you can't easily replace, especially stuff made by out-of-business companies, it's a good idea to study the schematic and purchase uncommon parts. The op-amp and current driver were not easy to find. I bought a few of each, "just in case".

  • The possibility that the op amp was blown and somehow pulling the current driver's output to the wrong voltage seemed unlikely. The current driver is the electrical "alpha male" in this circuit.
Reply to
William Sommerwerck

Sounds like you miss the old days when you could just keep swapping tubes and see if anything improves.

Reply to
ian field

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.