Re: a single output transformerstereo tube amp ??

"Boborann" <trottier@xxxxxxxxxxx> wrote in message
news:RcednV1qQN96Od_YnZ2dnUVZ_v2dnZ2d@xxxxxxxxxxxxxx
I bought a Columbia680 stereo
[snip]
What looked like a traditional Push Pull design is
actually fed with a channel per 6L6 and in phase .
The output winding is Center tapped and has a
speaker attached to each side of the winding.
Curious whather anyone has ever seen this and
could comment on how common it
was


Here are some comments from amplifier group....

mykeymykey wrote ..
-------
sounds like a cool amp, can you post us a link to the schematic?
this type of output was class A, about 12 watts per channel and
popular among budget hi fi buffs in the 50s-60s. the output
tranny was built like this just to save money and space.
its not very unusual but its uncommon to find one in good
condition that works. But yes tube amps and questions about
them are always welcome here.


po excuse 4 me wrote ............
----------
Yeah, it's an old trick to get low $ 'stereo' when 'stereo'
only ment 2 speakers, 1 L, 1 r.. I love the way the '3rd'
tranz works the OT..Hate to see what it would cost today
against $$$ design..hehehehe.

It's almost like old McIntosh tube amps with weird NFB based OT.
I got from Ned years ago, a donut OT with a 'screen' tap. NOT
a tube screen, but an xtra winding BETWEEN the pri/sec, that
can be...controled. Also multi-UL taps, just a cool hunk of
tranz.


Ned wrote .......
----------
It's a cheap (but clever) piece of ***, like the rest
of the stuff Columbia (apparently) gave away as premiums with
record-club memberships.

OK, think about this. Having a single output transformer
is cheaper and means no core gap. If one 6L6 is amplifying
something the other isn't, the signal appears out of phase
on the opposite channel speaker. Feed that back to the input
tube of the opposite channel, it cancels itself out, mostly.
Not perfect stereo separation, but it's not intended to be,
it's intended to give "stereo effect" at the lowest
possible cost.

flipper wrote .........
-----------
I'm not real familiar with the topology but it looks like an old
matrix stereo amp, a technique that was experimented with back in the
1950's with the idea being 'cheap', er 'economical'. You get a
'stereo' amp from not much more than a PP 'mono'. Saves iron in the
OPT even though still SE.

I say 'looks like' because the matrix circuits I've seen inverts one
channel, and that one doesn't, so maybe it's either a variation or an
even cheaper attempt at a similar concept.

In the ones I've seen it's analogous to analog stereo radio
multiplexing the L and R channels for transmission and then demuxing
in the receiver. In the 'amplifier stereo' case you have R and -L (the
inversion on the ones I've seen) amplified through the amp and then
summed (subtracted, actually, since on opposite sides of the CT) at
the OPT primary for R+L (this, btw, is 'normal' P-P polarity). Since
the OPT center tap is a common mode null it is R-L, ah hah!, so if you
then feed that R-L back (the purpose of the second transformer on the
CTs) into the main OPT you subtract it out (or add it in, take your
pick), opposite polarity on each side of the CT. So on one side you
have R+L-(R-L) for 2L and, conversely, R+L+(R-L) for 2R on the other
(if all your components were balanced, flawless, and without parasitic
capacitance, leakage inductance, etc)

It's likely that the missing 'inversion' on the front end, compared to
what I've seen, is taken care of by proper phasing of the OPT and CT
feedback but I'm not familiar enough with the topology to be
completely sure if it's 'done right' or a dodge.

But, what the heck and winging it here, that amp looks like both are
in phase, at the input, so it'll be R-L in the OPT primaries. Now,
that sounds bad, like there's 'nothing' with a common mode (mono)
signal, but the second (matrix) transformer is on the primary side CT
and will have the signal (and idle) current from both sides going
through it, in phase, like a double tube SE, so it's R+L. Ah hah!
Taking again the simple case of a common mode (mono) signal, the OPT
secondaries 'on the ends' would, in a normal P-P amp, null each other
(R-L) but the matrix transformer couples the primary CT R+L signal to
the output CT 'adding/subtracting' (depending on which side of the CT
you're on) it in. So, bingo, we have R-L+(R+L) for 2R, again, and
R-L-(R+L) for 2L on the other. Dreadfully crude analysis but it
illustrates the basic principle.

From what I've read the matrix amps didn't work terribly well, though
(components aren't flawless), with poor separation but did have the
interesting characteristic of sounding louder than normal for the same
power rating.





.