VOGONS

MOSFET testing question

Topic actions

First post, by scruit

Posted on 
User metadata
Rank Newbie
Rank
Newbie

I am testing an NPD603AL (N-Channel Logic Level Enhancement Mode Field Effect Transformer) from an old Comapq Proliant 1500 CPU daughterboard. Symptoms are no post / no video, with only the psu / blackplane / cpu daughterboard and CPU in place.

First question: This is not explicitly listed in the datasheet title as a MOSFET, however in the description it states it uses D-MOS technology. I assume this means it's just a MOSFET, and the "logic level" means it only needs 5v gate-source charge instead of 10v. Is that correct?

I am testing this because when the CPU is in the socket there is no POST but this FET get VERY hot. Within a minute the heatsink will burn your finger. With the CPU removed the FET does not get hot at all.

When I tested the FET on the board I found drain-source was shorted with the power off. I removed it from the board to isolate. Motherboard showed no short between the empty source and drain pads after the FET was removed.

Measuring the FET with multimeter in diode mode:

  • Source(+) and drain(-) the source I get the 0.5v diode reading as expected.
  • Source(-) and drain(+) I expect to see OL but got 1.7v.

Follow-up resistance testing showed OL in both directions. So I'm a little confused. I would expect the drain-source to be fully shorted and read close to zero ohms.

Second question: Could this be a partial failure? (Not fully shorted) or is this FET allowed to read 1.7v in this direction?

Maybe it was initially reading a short on the board because the FET got so hot it shorted, and after it cooled down it is not a complete short? Or maybe there was some residual voltage at the gate? (board was out and on my desk though)

Reply 1 of 4, by snufkin

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie

I don't have massive amounts of experience, but it looks like it's FET with a low threshold voltage, datasheet says 1.5V typical, dropping as temperature increases. Perhaps there was some stray capacitance on the gate when it was in-circuit, enough to drop the drain-source resistance. Still sounds broken if it's conducting drain-source at only 1.7V. Absolute maximum breakdown voltage drain-source is given as 30V (body diode would probably fail), so it shouldn't fail and start conducting until then.

Might be worth rechecking the drain-source diode test, but short the gate to the source to make sure it's definitely off. Datasheet says the drain-source voltage should have to be at least 30V before the drain current reaches 250uA if Vgs is 0V.

Reply 2 of 4, by pentiumspeed

Posted on 
User metadata
Rank l33t
Rank
l33t

MOSFET switches on and stays there if you charge up the gate pin will confuse you as shorted untill discharge the gate to any pins by shorting together. I use diode mode or resistance in respect to ground and gate pin. If open or reading diode voltage drop, they are good, if shorted replace the MOSFET.

Cheers,

Great Northern aka Canada.

Reply 3 of 4, by scruit

Posted on 
User metadata
Rank Newbie
Rank
Newbie

And the answer is:

I got some "new" (NOS) NDP603AL Logic enhanced MOSFETs and duplicated my testing. These tests are all with the mosfet on my bench.

Test 1: Short all 3 legs together for a moment, then use diode mode to test between source(+) and drain (-).
Result: Old and New mosfets both showed 0.5v voltage drop, as expected.

Test 2: Short all 3 legs together for a moment, then use diode mode to test between drain(+) and source (-).
Result: Old mosfet showed 1.8v voltage drop. New mosfet showed OL.

Test 3A: Short all 3 legs together for a moment, then use diode mode to test. Plane (-) on source, touch gate(+) for a second, then test between drain(+) and source (-).
Result: Old mosfet showed 1.8v voltage drop. New mosfet showed 0.00v voltage drop.

Test 3B: While holding the test 3a position, touch gate and drain with finger.
Result: Old mosfet showed 1.8v voltage drop. New mosfet went to OL.

Conclusion: Old mosfet has failed in an atypical manner and has become a TO-220 package motherboard warmer.

The typical failure mode for source/drain is described as a dead short. This is not a dead short, and continuity/ohms mode does not detect it. Only diode mode.

Reply 4 of 4, by pentiumspeed

Posted on 
User metadata
Rank l33t
Rank
l33t

Great!

Cheers,

Great Northern aka Canada.

Go to top of page Go to top of page
Back to General Old Hardware
Close window
Close window