Matching the IRFP240 9240? how many will be wasted in ratio?

[rhythmsandy]

I would like to build an amp using IRFP240 9240 pairs.
so usually how many will be wasted in an avg ratio where we could get a match to acceptable tolerances

1:20? or more?

the reason why is just to know how many to buy?

 

[nigelwright7557]

The 240 and 9240 are quite different transistors so don't match very well.

However this isn't a problem in amplifiers as feedback sorts out any differences.

I have designed a few amplifiers using 240/9240 and they sound very good.

If you want to match say the 240's then I wouldnt bother as they are likely from the same batch anyway.

 

[Bigun]

Perhaps, in a push pull topology, differences between the n type and p type may not be such a bad thing - as the result is mostly 2nd harmonics ?

 

[rhythmsandy]

its for pushpull config...

are they better than conventional Toshiba mosfets? in terms of sonic qualities?

 

[rhythmsandy]

how to match? just measure the RDS with multimeter any ckt for matching these?

 

[Ian Finch]

http://www.diyaudio.com/forums/pass-labs/220834-f5-output-mosfets-matched-not.html

 

[KatieandDad]

When I was matching for sextets, I needed 4 x 6 sets. I achieved this with a batch of 50 IRFP240's.

 

[sajti]

how to match? just measure the RDS with multimeter any ckt for matching these?

No resistance measurement. Nelson Pass has some pictures how to match Your FETs. Try to search for that.

Sajti

 

[sajti]

Check it: http://www.passdiy.com/pdf/matching.pdf

Sajti

 

[triplej]

h_a sells them already matched over in the swap meet forum.

 

[destroyer X]

Mismatch is positive.... harmonics goes much better this way

Strange isnt'it?.... try it... assemble one channel matched and other mismatched then you gonna see.

This is one of the main reason that academics, usually, do not produce good sound..they follow the rules...and mismatch is unaceptable for them... they think these things are blasphemy, heresy... and we... poor devils with low know how and practice knowledge, go making better amplifiers than they can even believe exist.... and they will never know..they do not believe and do not try... thinking they are much clever they stay stucked into the ignorance.

regards,

Carlos

 

[KatieandDad]

If the FETs are unmatched they will not "load share" correctly.

At the extreme one might be doing all the work so will fail prematurely. If they are matched then both are doing half of the work.

It is not necessary to match the 240s with the 9240s.

 

[JMFahey]

how many will be wasted in an avg ratio where we could get a match to acceptable tolerances

It all depends on what you define as acceptable tolerances
FWIW I commercially build Musical Instruments amplifiers, my two workhorses are IRFP250 in the >300W heads and IRF640 in the 60/100W combos.
Yes, I use quasi complementary outputs.
The only widely available "complementary" pair is IRFP240/9240 .... and it's quite unsymmetrical anyway, so, why bother?
Sometimes I also used IRF530/9530 with good results.
Rather than "matching them for a single amp" , which is the typical DIY task, I buy 25 or 50 unit tubes (depending on size: TO247 or TO220) , empty them on a table and measure Von , which in my case is Vgs needed to pass 1mA.
The measuring rig is *very* simple: a proper size socket, with G grounded, D to +15V and S to -15V through a 10K resistor , I measure S voltage relative to ground.
Pushing a transistor in the socket takes all of 5 seconds; writing the small stick-on label takes longer .
So in a few minutes I have 50 labelled transistors, which go back into the tube.
I pick and use matching Von ones if needed.
And what did I find?
I don't do it any more
Why?
Because I *consistently* find them very well matched
In my test rig, they usually sit at around 3.75 to 3.78V ... incredible!!
And anyway, I use ballast resistors, which take good care of smoothing any small difference.
*Maybe* if I parallelled them with no Source resistors, at higher currents they might not track .... but that will never happen in my amps, specially since I also make my own high precision, nil temperature coefficient wire resistors, out of Constantan, so, why avoid them?
So, in a nutshell, do measure them to know what you're dealing with, but I think you'll find them very well matched, out of the box.
And of course use source ballast resistors, you can't really leave them out, unless you use lateral Mosfets (and even so, I'd still use them).

FWIW, I buy "good" trusty ones, straight from the ARROW or AVNET distributor.
Maybe the shady EBay sellers do not provide same consistency.

 

[HighTec]

If the FETs are unmatched they will not "load share" correctly.

+1

The 'offended' FET will blow out in couple of seconds if not caught in time. Unfortunately, in most cases he will drag the other FETs with him to the point of no return. Been there, done that.