General MOSFET Question

Thread Starter

bwilliams60

Joined Nov 18, 2012
1,450
Good evening. I have a MOSFET that is marked as a PH4847 and is used in an automotive application. I believe it to be a 15A, 60V N-Channel MOSFET TO-220 similar to an IRFZ44N.
When I test this MOSFET with my DCA55, it seems to test okay. When I do it with a DVOM, it turns on and off okay but if I short the leads, I am still seeing a 2.9V reading on my meter with my red lead placed on the drain, and the black lead on the source. Everything else seems to read okay.
Is this a problem? If not, what am I looking at. Thank you in advance.
Also, if anyone has any clue as to what the PH4847 is, appreciate that as well.
 

LowQCab

Joined Nov 6, 2012
4,217
40-Volts,
60-Amps-RMS, ~283-theoretical-Amps-Peak,
3.6nF-Gate Capacitance,

You can't get much useful information about a FET with a Multi-Meter,
You need to fabricate a Test-Jig,
with a Bench-Power-Supply, a Function-Generator, and a 'Scope.
.
.
.
 

Attachments

LowQCab

Joined Nov 6, 2012
4,217
I don't "Test" FETs,
I install a brand new one.
Why take a risk on a part that could be partially damaged ?

Generally, testing is not necessary, and is of little usefulness.
Just read the Spec-Sheet, and design very conservatively.

A Computer-Simulation is best for determining the effects of Gate-Capacitance on
the Gate-Driving Current requirements,
and finding the ideal Gate-Resistor value for Switching-Stability.

Why do You need to "Test" a MOSFET ?
Testing a FET in an SMD type Package,
requires the FET to be Soldered to a Circuit-Board with adequate Heat-Sinking capability
in order to determine the extremes of it's performance.

De-Soldering a FET can be a hazardous thing to the FET, OOpsies do happen,
and the Gate will be subject to static-damage until it can be properly Grounded.
And then You have to Re-Solder it to another Printed-Circuit-Board, which brings additional hazards.

FET Gates can be partially damaged, or completely destroyed, in the blink of an Eye.
Gates must remain Shorted to the Drain-Pin at all times when not Soldered into a protective-Circuit.
.
.
.
 

Mullins

Joined Dec 31, 2021
179
I apologize for intruding... In my case some LPG injection ECUs Sometimes some function fails. Since these ECUs use MOSFETs are you saying it doesn't make sense to test the responsible MOSFET? For example, I cannot simulate the signal coming from the MCU on the bench so I can only test the MOSFET to determine if the problem is in it or not.
 

LowQCab

Joined Nov 6, 2012
4,217
I apologize for intruding... In my case some LPG injection ECUs Sometimes some function fails. Since these ECUs use MOSFETs are you saying it doesn't make sense to test the responsible MOSFET? For example, I cannot simulate the signal coming from the MCU on the bench so I can only test the MOSFET to determine if the problem is in it or not.
.
In your case the best thing to do is to test the FET with an Oscilloscope while the Circuit is operating.

If you have a particular FET that is known to fail,
then it is because of poor Circuit design, or poor Board design, ( over-heating ).
It may be possible to replace the FET with a different one that has higher ratings.

It may also be possible to replace an SMD FET with a trough-hole-FET
which can then have a Heat-Sink attached, ( if overheating is the cause of failure ).

If overheating is the cause of failure,
it may be a good idea to mount the entire Controller
in a different, cooler, location, with good Air-Flow.

If the FET controls a Solenoid,
make sure that the Solenoid has a Freewheeling-Diode,
and/or a Snubber-Circuit attached to it.
.
.
.
 

Wendy

Joined Mar 24, 2008
23,464
At the risk of stating things you already know a normal mosfet requires at least 10 volts on the gate to turn fully on unless it is a logic level mosfet, Are you sure this device is turning on all the way?
 
Top