mosfet matching wide variations/swings

[vdi_nenna]

Purchased 20 irfp140 and 20 irfp9140 mosfets.
Matching them has been less than fun.

Setup:

Mastech 30v 10A power supply
Large-ish blank anodized heat sink 4x6x1 inch
Sheathed alligator clip leads
Silicon iso pad.
15v, 60 ohm power resistor .13A

I thought I was getting good readings but if the dmm lead are touched or if I reach over the setup, the voltage swings but never settles to it was originally.
If i re-test, the Vgs is different.
If i touch the heat sink the voltage changes.

I don't trust the readings so I can't continue with the build. I don't remember it being this difficult.
I tried 4 meters and different meter leads.

How is matching done properly?

 

[Zen Mod]

pics

 

[vdi_nenna]

Gate to drain
15v to drain
Ground to source pin

 

[flg]

Are you using a gate stopper resister?

 

[Nelson Pass]

First, don't put a lot of trust in the meters which come with the regulated supplies.

Second, you may be seeing some parasitics - use a 470 to 1Kohm Gate resistor.

If this doesn't work, try an analog meter.

 

[flg]

From http://www.firstwatt.com/pdf/art_mos_test.pdf :
"P4 doesn’t matter here , as the DC is blocked by the
capacitor, but it is always nice to have a few ohms in series
with the Gate of any Mosfet to prevent high frequency
(parasitic) self-oscillation, which will alter the voltage
reading."

 

[flg]

Sorry, I'm a little slow sometimes...

 

[vdi_nenna]

Thanks everyone.

So the output reading on the regulated supply is not to be trusted. Check.
Use a gate stopper. Check.
Will recheck tomorrow. Check.

 

[vdi_nenna]

Appears to be solid now. Moving the cables doesn't move the meter.

Used 1/2w 750 ohm.

Thanks!

 

[AndrewT]

You must keep Tj at the same value for every test/DUT, that is really difficult !

 

[vdi_nenna]

You must keep Tj at the same value for every test/DUT, that is really difficult !

Yes, I imagine it's difficult keeping accurate temps. I left the setup out over night with the mosfets on the same table, Closed a vent that pushed air over that work area. That's the best I could do. I wasn't going to measure Tj.

the readings were completely different for the N-Ch fets. Some read 7.xxV other times 5.xx, but with the gate resistor, every fet was in the 5.1xxV range.

Interestingly, the P-ch fets had closer readings with or without the gate resistor, even a day later and with different room temps.

Thanks.

Vince