Another output transformer feedback solution.

[Paul Barker]

Now that I would like to wind, when I get started again.

Shishido 800 push pull.

[Cressy Snr]

Now that I would like to wind, when I get started again.

Shishido 800 push pull.

Hmmm interesting indeed.

[Paul Barker]

Hmmm interesting indeed.

Although after thinking about it there is no difference from the Quad II. It is a seperate winding in both cases. Just that it is diagramtically portrayed differently. Except that people have reverse engineered the Quad II so we know the amount of feedback. Perhaps the 800 one has more feedback. But anyway.

But there is something else interesting. The zobel network on the secondary of the phase splitting transformer.

[JamesD]

Nice amp!

Its similar to the Quad and to the Mcintosh, Hafler, etc. but the use of interstages changes the game somewhat as keeping the feedback loop stable with all the transformer phase shifts is not trivial at all and, I guess, that is part of the purpose of the zobel across the phase splitters secondary.

The other part is that, I suspect, would be the dominant pole inside the feedback loop and defines the HF roll off at 132KHz...

Damn PP amps and their extended bandwidths!

J

[BobWeaver]

Would work much better, I think, if there was power to the filaments on the output stage.

[IslandPink]

It's well-hidden facing left on the mains transformer above the 100V AC input . Shame to put a crappy DC supply on there, would probably be better to run clean AC into the 800's .

[BobWeaver]

Ah okay, I missed the DC supply, but it still brings up a question of what happens to the 10 ohm pot functionality when the filament supply is tied across both of the filaments. The pots end up being tied in parallel, and the feedback signals will mix together and mostly cancel out. It seems to me that the filaments need to have completely separate transformer windings.

[Paul Barker]

Ah okay, I missed the DC supply, but it still brings up a question of what happens to the 10 ohm pot functionality when the filament supply is tied across both of the filaments. The pots end up being tied in parallel, and the feedback signals will mix together and mostly cancel out. It seems to me that the filaments need to have completely separate transformer windings.

Yes I totally agree. Possibly simply a mistake in the drawing, not transferred to the build.

If Ongaku schematic can have errors so can this one.

As Mark noticed the schematic is from a period before Guido made the first clever filament supply. Another similarity to Ongaku.

[BobWeaver]

When I first saw the schematic, I wasn't sure whether this was an actual project that someone had built, or whether it was just a proposal. But yes, if this is an actual project, then no doubt there must have been two separate filament supplies.

It's well-hidden facing left on the mains transformer above the 100V AC input . Shame to put a crappy DC supply on there, would probably be better to run clean AC into the 800's .

I'm always skeptical when I see a DC filament supply especially in the output stage, especially push-pull. I can't see why it couldn't be done with AC. The 10 ohm pots must be for hum cancellation. I can't see any other purpose for them.

A couple of things in particular that I do like about this circuit.
- The local feedback from the secondary of the OPT back to the output stage.
- The output stage is operated with positive grid bias to get lots of plate current with only modest plate voltage. (I tend to be a bit squeamish running the plates at lethal kilovolt levels.) In fact, I have a project of my own (in progress) that uses positive grid bias to get high plate current at modest voltage.

On the other hand, I think the idea of using a transformer (especially two of them) for the phase inversion is archaic, not to mention expensive. I can't get too excited about that part. Just my opinion.