Power Transistor and pairs

Hello,

May be this subject has already being discussed. I understand that this might be frustrating for some people to go over it again. But here is my question:-

A year ago I was trying to troubleshoot a Denon AV receiver. Denon (as an example) uses Sanken Complementary transistors. This amp is 5.1 output receiver. So I need 5 Complementary types (as an example 2SD2390 Complement to 2SB1560). Apart from being complement to each other, they have to read the same Vbe and same hFe. (Matched) I understand that. So when I applied to a store providing these, the question was " Do you want them Matched pairs ?" I replied yes.

Now I am in the process of repairing a QSC RMX2450 completely burned both channels. Measuring around, I understood that the transistors are fried, drivers, power transistors, etc. QSC uses in this instance Toshiba 2SC5200 and 2SA1943 complementary pairs.

I read in this forum that these are constructed in very tight tolerances so they do not really need to be matched and incidentally the supplier did not ask me whether I want them in matched pairs. Searched a bit on the web for matched pairs, resulted in no hits for matched pairs. However for the pre-drivers MJE15032 and MJE15033, there was a question do you want them matched pairs ?, and there was a hit for matched pairs on the web.

Dazzled !! Can this be true ?
 
Usually only a single pair is used for Drivers so " matching" 15032/15033 is pointless and sounds like a gimmick.
No NPN/PNP is matched some just get closer.

Far as multiple same type NPN or PNP run in parallel BJT power transistors, rarely matched
And doubtful they actually match them, besides tossing them on a Ebay tester.
Which isn't enough voltage or current to be meaningful for a power transistor

Far as darlingtons, again NPN/PNP will never match.
And if only single pair is used, matching is pointless for multiple channels.
Mentioning matching was just a gimmick, or if more money is charged...there you go

The basic transistor testers sold on ebay are helpful for repairs or finding faulty transistors.
Your more than welcome to buy one and " match" yourself.
Again the voltage and current they test at is not useful for a power transistor.
But you will get some sort of Hfe measurement.

Likely why you might get no answer or strange reaction to "matching"
Nobody does it really and anyone claiming so is likely lying.
Or basically again tossed them on a low voltage/current tester
 
Last edited:
Matching darlingtons is damned near impossible because of the sensitivity of the composite beta to the first stage. You can get sensible grouping, but that’s about it. It is possible to match C5200’s to one another, and get NPN PNP reasonably close. Like hFE +/-1 out of 130. In some cases it’s worth trying, others it’s not. The distortion/noise floor on the QSC is sufficiently high that matching other than getting them from the same production run won’t do squat. Different runs can have gross enough differences where parallel units wont share, so you have to avoid that. The better match the drivers are the better, but not critical on the QSC. Those won’t match NPN to PNP over a wide range. If it was critical, they would be biased up (by design) to get them over the hump where matching becomes possible. They aren’t run that high in the QSC. It is compromised toward efficiency and thermal stability because of its typical use case.
 
  • Like
Reactions: karatzdi
Good morning

I think its been pretty clear from both @wg_ski and @WhiteDragon, that matching is probably a "Myth" anymore. I guess that, more money is involved in getting them but technically is a gimmick like you said.

I don't have an easy access to peers so that I exchange these type of questions over here, in this part of the world. So input like yours wg_ski and WhiteDragon are like jewels.

Wg_ski answered another question I had, which was about the drivers to the power amps Transistors. Thank you very much for the BIAS comments on QSC's and details about their work.

Appreciate it . Thank you