Precisely Matched Pairs in T0-3

[py...@ix.netcom.com]

I have an output stage comprised of 8 pairs of 2SD424(npn) and
2SB554(pnp) TO-3 transistors (Toshiba). I would like to improve the
sound of this amp and one of the many mod's i'm planning is to replace
the output devices with matched pairs that were designed with
audiophile purposes in mind.
Does anyone know the specs for the existing transistors or know
where I can get them? (Toshiba, the OEM, does not have them).
Can you reccommend any pairs ? Do you know what's used in the
Levinson, Krell, Spectral around 250W/ch?

[Warren Gaiewski]

py...@ix.netcom.com wrote:
>
> I have an output stage comprised of 8 pairs of 2SD424(npn) and
> 2SB554(pnp) TO-3 transistors (Toshiba). I would like to improve the
> sound of this amp and one of the many mod's i'm planning is to replace
> the output devices with matched pairs that were designed with
> audiophile purposes in mind.

> Does anyone know the specs for the existing transistors or know
> where I can get them? (Toshiba, the OEM, does not have them).

MCM, Parts Experss, and the like _used to_ carry them. They're getting
rare, because most consumer electronics don't use TO-3's anymore.
The originals are 150 watt, 16 amp, 180 volt, fT=5 MHz, with a beta
of about 70 at a couple of amps. You could probably use the usual
Motorola MJ15024/15025 or MJ21193/21194 pairs which are readily
available, but they probably won't make it sound any different. Might
handle 2 ohm loads a little better (2.2 amps at 80 volts I s/b for the
15024, 2 amps at 100 volts for the 21194).
Not knowing more about the amp in question, my guess is that other mods
will do more for the sound than replacing the outputs.

Using higher fT outputs like the 2SC3281/2SA1302 pairs could potentially
allow higher open loop bandwidth, but the entire frequency compensation
scheme would need to be redesigned to take advatage of it. If you know
what you're doing, it might be worth trying. But faster outputs are
easier to blow - the second-breakdown capability isn't as good. With
a cascode output stage this may not be a problem, but for straight
parallel banks off +/- 85 volts you need to watch it. Another note -
most of those high speed Jap outputs aren't available in TO-3.

> Can you reccommend any pairs ? Do you know what's used in the
> Levinson, Krell, Spectral around 250W/ch?

Don't know what the really high end suff is using nowadays. A lot of
the smaller pro stuff is using MJ15024's. I use 'em because they're
reasonably cheap and give very good performance, IMO.

[PY...@ix.netcom.com]

On Mon, 10 Feb 1997 23:06:06 GMT, py...@ix.netcom.com wrote:

> I have an output stage comprised of 8 pairs of 2SD424(npn) and
>2SB554(pnp) TO-3 transistors (Toshiba). I would like to improve the
>sound of this amp and one of the many mod's i'm planning is to replace
>the output devices with matched pairs that were designed with
>audiophile purposes in mind.
> Does anyone know the specs for the existing transistors or know
>where I can get them? (Toshiba, the OEM, does not have them).

> Can you reccommend any pairs ? Do you know what's used in the
>Levinson, Krell, Spectral around 250W/ch?

Thanks for the response Warren, I truly appreciate the data and
advice. I am a novice when it comes to engineering solid state
amplification, but I understand the basics and I read up on what I
don't understand.
For anyone that can offer advice, the amp in question is an Adcom
GFA-555 II. I have been advised to replace the outputs with something
better and to tie the Bases together (assuming emitter followers,
which I haven't checked yet) with a .22uF cap to avoid the sluggish
turnoff when the signal swings through 0v. I also was told that the
drivers use a reference from the outputs to simultaeously turn on. I
would remove that reference and tie it to the negative rail with a
larger resistor.(Class A).
If there is anything else you can reccommend, or if there is a
reason I should not do what I mentioned above, please let me know.
Jim Stead

[Warren Gaiewski]

PY...@ix.netcom.com wrote:
>
> On Mon, 10 Feb 1997 23:06:06 GMT, py...@ix.netcom.com wrote:
>
> > I have an output stage comprised of 8 pairs of 2SD424(npn) and
> >2SB554(pnp) TO-3 transistors (Toshiba). I would like to improve the
> >sound of this amp and one of the many mod's i'm planning is to > replace

[snip]

> For anyone that can offer advice, the amp in question is an Adcom
> GFA-555 II. I have been advised to replace the outputs with something
> better and to tie the Bases together (assuming emitter followers,
> which I haven't checked yet) with a .22uF cap to avoid the sluggish

> turnoff when the signal swings through 0v I also was told that the

> drivers use a reference from the outputs to simultaeously turn on.

Don't know about that particualar amp, but most of 'em are pretty
similar.
What they're probably talking about is the "Vbe multiplier", which is
a thermal reference. It's usually good practice to bypass it with a
good high-frequency cap, which would be connected to the bases of the
driver transistors.

I usually see the bases of the outputs tied together through about
a 10-200 ohm resistor. This bleeds off the stored charge when one
cuts off and the other cuts on. Also partially sets the bias current
in the drivers.

>I
> would remove that reference and tie it to the negative rail with a
> larger resistor.(Class A).

I'd leave it where it is for now. If it is the vbe multiplier, you
need it. If it's something else, I'm not familair with it and would
need to consult the schematic or take a peek at it.

----------|>|-----------

[CyberBlob]

In article <32ffa6d8....@nntp.ix.netcom.com>, py...@ix.netcom.com
whines...

>
> I have an output stage comprised of 8 pairs of 2SD424(npn) and
>2SB554(pnp) TO-3 transistors (Toshiba). I would like to improve the
>sound of this amp and one of the many mod's i'm planning is to replace

>the output devices with matched pairs that were designed with
>audiophile purposes in mind.
> Does anyone know the specs for the existing transistors or know
>where I can get them? (Toshiba, the OEM, does not have them).
> Can you reccommend any pairs ? Do you know what's used in the
>Levinson, Krell, Spectral around 250W/ch?

If you upgrade the transistors, there are many other things to consider.
First out, I would talk you out of the thought, and I will explain why.
Think of the amp as a "chain", with links. The amp is likely a "DC" amp,
meaning "direct coupled", meaning :) each stage of amplification is
directly coupled to the next. When you repair or design an amp circuit,
you need to consider many things about the transistors you use in the
amp, and the way you are using them. If the previous stage, the stage
that drives these output transistors, can't provide the signal level
needed to drive your amp to full power, full power usually being limited
by the power supply, then not only will you lose some power, you lose
reliability, the previous stage may have a hard time keeping up, and you
know what happens next. In a DC amp, when one stage goes, what usually
happens is it takes the outputs with it, along with some resistors and
possibly every stage on the in between. The only reasonable upgrade
would be to replace the TO-3's with one of the same type, with a
higher power dissipation and maximum current from the E to C or the
C to E, and it is VERY important to keep the same gain. The chain is
only as strong as it's weakest link. Scream if I can help ya with
anything else.

[Jon Risch]

PY...@ix.netcom.com wrote:
> I have been advised to replace the outputs with something
> better and to tie the Bases together (assuming emitter followers,
> which I haven't checked yet) with a .22uF cap to avoid the sluggish

> turnoff when the signal swings through 0v.

>Warren Gaiewski <dio...@airmail.net> wrote

>What they're probably talking about is the "Vbe multiplier", which is
>a thermal reference. It's usually good practice to bypass it with a
>good high-frequency cap, which would be connected to the bases of the
>driver transistors.

Jim,
Warren is essentially correct, and you generally want to put as big a cap
in this position as you can afford, and as high a quality capacitor too.
This cap bypasses the DC bias circuit (resistors, diodes, transistors or
some combination)at audio frequencies, and you want the semi's and
resistors out of the way as soon as possible into the audio band.

On some amplifiers with a basically good circuit and parts, adding a good
big cap here can make a wonderful difference.

Jon Risch