Pass Mosfet Matching Help

Jump to Latest
#41
really?😱 Then I really don't get it... N-ch MOSFETs look like this (from Vishay datasheet):

Screenshot 2023-08-16 at 20.59.36.png
So gate and drain should go to R1 and source ground, or do I get something wrong?
 
#45
well, that looks better! thanks a lot!!

lab_OK.jpeg

I will create a 3-pin screw terminal directly soldered to the probe wires, in which I can screw in the mosfets (after all, I have close to 100 to test...), it might be even better.

so now, basically, I should lower Vcc until I have Iq=0,170A, and then I match the devices by the Vgs displayed? That means, all devices will have different Vcc, whereas in the real circuit, they will be fed by the exact same Vcc rail. Doesn't it matter?
 
#48
Crank it up to the intended Iq, no reason whatsoever to not do that. And also, consider a second measurement of all devices with a higher Iq, for example 350-400mA. That will enable you to get an indicator of transconductance too 🙂
 
#50
No, but you know, sometimes my Viking spunk forces me into maxing bias beyond safe levels. So, one measurement for stock Iq, and one to ensure good matching when going frying pan loco, is always good. And hurts much less than a spindle up my butt. And avoids screaming :rofl:
 
#51
The Schematic Zen Mod is great. I noticed heatsinks for the mosfets under test. How long do you test them for?

Here is a link that can come in handy:

https://sites.google.com/bucknell.edu/ericshometheater/erics-home-theater/aleph-x-clone-amps

Press Control + F and enter: "Matching Mosfets" and note that there are instructions for TO-220 mosfets and TO-247

Here is another site:

http://www.diamondstar.de/transistor_matching_mosfet.html

Also, a little indepenant research. When I first started matching mosfets, I matched some IRFP9240's with the circuit you would use for a IRF9610. I matched a whole set of around 100. About a month later, I realized my mistake. So I lined up all the mosfets from lowest to highest. I then set up a new circuit with a higher voltage, resistance to hit the appropriate IQ. I tested each for 45 seconds a piece. Just long enough to get them pretty warm. Only a few mosfets changed their locations maybe one place lower or higher with the new testing. Technically they would have been well within tolerence. Also, that could have been a slight error in my testing.
 
#53
Is see. So probably irrelevant. I may have overdone it then. But I swore never to solder in poor matches after my run-in with two bad sets from the store a few years back. Practically, I just changed the lab PSUs voltage after the initial measurement @ stock Iq to see if they were consistent or not. All were. But even though it from an objective standpoint made no difference, it reassured me. At the time I needed that. ❤️ 🏳️‍🌈
 
#54
rhthatcher said:
They don't get warm, and I run them for ~10 sec. So no heatsinks. You'll see the Vgs settle by 10 seconds. I mark the Vgs reading on the MOSFET with a fine tip sharpie and set them aside. Sort the batch by VGS when you've measured them and group them together.

I make sure they are all temperature stable for the matching session. Translation: if they are delivered on a cold winter day or hot summer day, I let them set inside in my temperature-controlled home for a day or so before I begin matching. Simple.
Click to expand...
what you do is good enough to make for some very tightly matched sets for my F4 monsterblocks. Good times!
 
#55
Ones I have to match/measure 350 mosfets irfp240/9240. I learned hard way that there is something called tempco.
Repeted measurement was always different.
Then I used relative measurement, Dut Vgs compared to reference mosfet Vgs.
Dut and reference mosfet are clamped to same heatsink in close proximity.
 
#56
If I match different batches of fets on different days, I retest the ones I want to use before I install them in an amp. My jig is easy to setup. Also, my handwriting is barely legible so.... I need to confirm what I am reading is actually what I wrote. 😬 I haven't noticed any issues so far.
 
You must log in or register to reply here.
Top Bottom