Mosfet testing

[frags]

Originally Posted by Bksabath

Please tel me more
Al

I'll draw a schematic tonight and upload it with more details.

[AndrewT]

Originally Posted by Zen Mod

- you can make differential tester - with one mosfet as "etalon" one and another as DUT, but you need controlled temperature for both etalon and DUT

There's another way of matching FETs if you have two matched resistors (0.1% or better).
In this case voltage regulation and FET temperatures are not a problem at all.
Using this method you will match two FETs connected at the same time with same conditions.

Originally Posted by Bksabath

this sound very good
Please tel me more

differential test, compare DUT to REF.

[frags]

Originally Posted by AndrewT

differential test, compare DUT to REF.

Exactly, and, if you place both of them at the same heat sink with the same thermal pads close enough, the temperature drift will now be an issue.

[Nelson Pass]

You have to do a lot of testing to assure perfect matching, and
this is not usually practical. That's why there are ballast
resistors on the Source pins - to improve the equality of
behavior.

For Mosfets I simply test Vgs under typical idle conditions
and use parts from the same lot code. In actual output
stages with lots of parallel devices this works fine.

[frags]

Nelson, you are 100% right.
When I bought 20 IRFP240s from mouser I got maximum Vgs difference 6mV at 1-1.5A Iq.
It means that the transistors manufactured under same conditions (some equipment, timestamp, etc) are very very close to each other and do not require an extra effort to match them.

[Zero Cool]

Originally Posted by Nelson Pass

You have to do a lot of testing to assure perfect matching, and
this is not usually practical. That's why there are ballast
resistors on the Source pins - to improve the equality of
behavior.

For Mosfets I simply test Vgs under typical idle conditions
and use parts from the same lot code. In actual output
stages with lots of parallel devices this works fine.

For a conventional amp i agree but what about matching for rail switches where there are no source/drain resistors?

[frags]

Originally Posted by Zero Cool

For a conventional amp i agree but what about matching for rail switches where there are no source/drain resistors?

For this it does not matter at all. Not big deal if one FET has ON resistance 18mOhm and another 19mOhm.

[frags]

Bksabath,
Here is the schematic for differential FET testing

You need to measure the voltage difference between FET drains which should be zero in case of two absolutely equal FETs.
With the potentiometer you can set desire current via FETs.
R5 and D1 for FET gate protection and can be omitted.
Temperature drift will equally affect both FETs if are they mounted at the same heat sink.
Again, you need to match R1 and R2 as close as you can.

[AndrewT]

Frags,
I can't open your link, not even in a new window.

[Bksabath]

Hi Frags

Sorry Mate
The link does not work

Al