Fuzzy on FET matching?

Whats the problem? You want to know the IRFP244 is a N channel?

Hello dantwomey,

Here it is the diagram for the matching.
Hopeing that it will be more clear to you.
Also, you can find the same procedure of matching
on the www.aussieamplifiers.com .


Best regards,
Kristijan Kljucaric
http://web.vip.hr/pcb-design.vip

How can you get FETs that are all from the same wafer? I ordered the standard package from Digi-Key for IRFP140. I got 25 FETs in a sealed plastic rail with an IRF label on it. The FETs all have the same markings, and the rail has the same markings as well. Does that mean these FETs were all from the same lot/wafer/factory/galaxy?

Yes

Hi Kristijan,

I have got a few questions left:

Is the IRF 9610 the P channel Mosfet and the 240 the N Channel?

Can I use a small piece of Alluminium as a temporary cooling device for the Mosfet during the testing/matching?
Should I wait a few minutes to let it reach a higher temperature?
Can I also use a 12 Volt DC 1000mA for this matching procedure?

With kindest greetings,

Lucas.

Youre right, 9610 is p channel and 240 is n

You can use piece of aluminium as a heatsink for matching. I would advise you that you use clock and read vgs of all mosfets at some time, when i was matching my 100 pcs of irfp240 i readed vgs exactly 30 secs after i connected each mosfet. You dont even need heatsink because mosfet doesnt get very hot in 30secs.

You can test them at 12v and 1000ma.

Hello,

Use regulated power supply when matching the FETs.
The heatsinks are not required if you will properly try
to match the Fets.

Best regards,
Kristijan Kljuèariæ
http://web.vip.hr/pcb-design.vip

Hi Skara,

Thanks! I will try your 30 seconds method.

I forgot to ask whether the BC 550 C (which I ordered as a substitute for the ZTX 450) should be connected as a P channel or N channel Mosfet while matching?

Kindest greetings,

Lucas.

Hello Lucas_G,

You can't match the BC 550 transistors with the method above.
It is only for the IRFP 240 and IRF 9610 transistors (MOSFETs).

Use hfe function on the your multimeter to match the BC 550
transistors.

Best regards,
Kristijan Kljucaric
http://web.vip.hr/pcb-design.vip

Generally you do not match BJTs, because their characteristics are very close already.

Actually, you do match BJT's, but not generally for Vbe.

You match them for current gain. This is useful for any
differential pairs and output stages.

Even if you are not matching, the current gain can vary
quite a bit, so at least we want to confirm the beta
(current gain, Hfe) and usually we want higher values.

Hi Nelson,

How can I match these thing then?
Same method as with the n or p channels?

Greetings,

Lucas

Nelson Pass said:

Actually, you do match BJT's, but not generally for Vbe.

You match them for current gain. This is useful for any
differential pairs and output stages.

Click to expand...

Who am I to contradict you, but unless I got my math wrong
I would tend to agree with Sloane that for a diff pair it is
actually Vbe matching that is important. Matching also hfe may
be beneficial in minimizing offsets. I know you wouldn't take your
opinioin just right out of the blue, so I would be most interested
to hear why you think that hfe matching is more important than
Vbe matching for a diff pair.

NB I am only talking about diff pairs here. In most other cases
hfe matching would be more important.

Usually I match my transistors at the same voltage and current
I intend to use them.

For higher powers this is not so practical.
You must use heatsinks, and heatsinks takes time to warm up
to the maxium heat level.
And you have to mount and demount the transistors.

Anyhow it is best to allow the heat
to warm up all your transistors to the max temp
before you messure their values.

The strive would be to match for the condition
at which the transistors will operate.
And to give all samples the same test condition.

You can test them at 12v and 1000ma

Click to expand...

1000mA for a 610/9610 without a heatsink is rather high...

12 V and 1 A is even a bit much for the IRFPXXX type IMO if using the 30 seconds. I use the current at which the transistor is to be used i.e. biased. Only that way I believe will they be matched or we would need to match both for Vgs and transconductance I suppose to know that they are reasonably matched at the later working point.

/UrSv

Christer said:

Who am I to contradict you, but unless I got my math wrong
I would tend to agree with Sloane that for a diff pair it is
actually Vbe matching that is important. Matching also hfe may
be beneficial in minimizing offsets. I know you wouldn't take your
opinioin just right out of the blue, so I would be most interested
to hear why you think that hfe matching is more important than
Vbe matching for a diff pair.

NB I am only talking about diff pairs here. In most other cases
hfe matching would be more important.

Click to expand...

As a different pair try to match the input on both sides,
they act like a balance, that weigh things.
So this is why both sides have to be identical.
Preferable in all aspects and have the same working condition.
Like current and voltage.
This is very hard to do, even if you have same b-e voltage and hfe.
They can behave different around that messuring point.

It all in the end comes down to your own standard, to what degree
you match.

Nobody should even try to make me believe that most commercial amplifiers have closely matched transistors.
There is no time for that in the factories,
as time for them is money.
And yet those amplifiers does a pretty good job in bringing us music.

halojoy said:

Usually I match my transistors at the same voltage and current
I intend to use them.

Click to expand...

This was the way I went through for my BOSOZ and XSOZV2.

JH