TIP2955/3055 matched pair - where to find?

[bobdole369]

Hi all,

4 channel x 50W amp went out last night, wasn't playing loud or anything, just got all garbly, then it stopped. Took it apart and found a TIP2955 shorted. These are the main output transistors. It's a 3055/2955 complementary pair, so I'm looking for a matched set. I figure I would also replace all of them while I have it open.

So while I can find individuals all over the place, having a tough time finding a matched set for sale. Anyone know where to obtain?

[jaycee]

I doubt anyone will be selling a matched set of such a common low-power device. I strongly doubt they need to be matched in a circuit using them.

[AndrewT]

if you have 1pair to each channel inside that 4channel amplifier then you do not need to match the devices.
Matching becomes a requirement when one parallels output devices, i.e. 2pair and more/channel.

[bobdole369]

Ah!!!

Yes, indeed it is one pair per channel. So I'll just pick up a bunch of single units then and not worry about the match. Many thanks guys!

[nigelwright7557]

Originally posted by AndrewT
if you have 1pair to each channel inside that 4channel amplifier then you do not need to match the devices.
Matching becomes a requirement when one parallels output devices, i.e. 2pair and more/channel.

I have seen designs with multiple pairs that werent matched and worked fine. This is why they add emitter resistors to balance out the power.

[wg_ski]

Originally posted by nigelwright7557



I have seen designs with multiple pairs that werent matched and worked fine. This is why they add emitter resistors to balance out the power.

It really sepends how anal you want to be about it. For practical purposes, multiple of the same die type with emitter resistors are good enough for safe operation. Some, however, are convinced it degrades SQ when the voltage across one emitter resistor is 25 mV and the other is 24.

[nigelwright7557]

Originally posted by wg_ski



It really sepends how anal you want to be about it. For practical purposes, multiple of the same die type with emitter resistors are good enough for safe operation. Some, however, are convinced it degrades SQ when the voltage across one emitter resistor is 25 mV and the other is 24.

I have designed numerous high power amps using 2n3055/MJ2955 and IRFP MOSFETs in multiple pairs and they have all sounded very good. In no cases did I bother matching transistors. In fact one of my designs doesnt even use complimentary pairs.

I think its a bit anal arguing over 1mv when outputting 900watts !

[wg_ski]

Originally posted by nigelwright7557


I have designed numerous high power amps using 2n3055/MJ2955 and IRFP MOSFETs in multiple pairs and they have all sounded very good. In no cases did I bother matching transistors. In fact one of my designs doesnt even use complimentary pairs.

I've got a 140W/ch unit that sounds darn fine and uses Sanken LAPT PNPs and 1980's vintage aluminum-case Motorola 2N3773's in one channel, some darned if I know what it is Sanyo 2SA's with the 3773's in the other. When the application isn't critical, I just use what's on hand. Even stretching a 2N3055 to 90 volts on occasion.

One of my PL400's has D555's in the upper bank on one channel, and RCA1B05's for the other three. It came that way - don't know who serviced it or if it was made that way.

I think its a bit anal arguing over 1mv when outputting 900watts !

That argument usually comes up when "It's the first watt that counts".

[AndrewT]

Originally posted by wg_ski
when the voltage across one emitter resistor is 25 mV and the other is 24.

this is an extremely close match of output emitter current.
Matched pairs often give a spread three to four times as much, i.e.
20mV to 24mV.
Unmatched could be as bad as ten times as much, i.e.14mV to 24mV.

Now try to show me how both these unmatched transistors that are dissipating vastly different quiescent power can possibly have similar junction temperatures.

[AndrewT]

Originally posted by nigelwright7557
with multiple pairs that weren't matched and worked fine. This is why they add emitter resistors to balance out the power.

I don't think so.
The emitter resistor is intended to provide feedback and balance of higher currents when the output is delivering current.
The quiescent emitter resistor voltage is so low and the difference between Vre as a proportion of the Vbe+Vre is so small that unmatched transistors gain little benefit in balancing quiescent current.
However, before one puts such faith in the measured Vre one must first match Re to far better than the usual 5% tolerance of these low power resistors.
I select for <0.5% as a matter of course and try for 0.2%.