understanding why 2 Fender AB763 preamp circuits "out of phase" with each other.

[mikepukmel]

Ive seen comments on blogs that some of the marshall amps, the users can take a short guitar cord, plug it into channel 1 and channel 1, then plug their instrument into channel 1 and both preamps will push the power amp. More gain?

Ive also seen many comments that you can't do that with Fender amps because the two channels are "out of phase". Im trying to understand that comment. I understand that going through one triode, e.g. on a 12AX7 from pin 2 to pin1, the output signal on the plate is 180 out of phase with the input grid.

To understand how to get one channel out of phase, is the difference only the number of stages each channel goes through?

Both clean and tremolo go through 2 stages of V1. SO, after V1, both channels are in phase with each other and with the input?

Then the reverb channel goes through a ganged 12AT7, (so flips phase "out" here), one side of another 12AX7 (reverb recovery), so another phase flip so flips back in phase, then the other half of the 12ax7 which looks like just another gain stage to boost some of what was lost in the reverb circuit, so back out of phase again.

is this right:
clean: -> out -> in
tremolo: -> out -> in ( -> out -> in -> out).


But don't all of those capacitors in the signal path cause phase delays as well? How are those accounted for? Would you need a simulator, like a spice or something like that to determine accurate phase relationships?

If someone put another gain stage on the clean side, not that I want to do that, but would that put the two channels in phase or is it more complicated than that?

Thanks, MP

[J M Fahey]

Ive seen comments on blogs that some of the marshall amps, the users can take a short guitar cord, plug it into channel 1 and channel 1, then plug their instrument into channel 1 and both preamps will push the power amp.

yes, itīs very common.

More gain?

Not much, only 2X or 6dB
But both channels are equalized very different, so much so that plugging your guitar in both and tweaking both volume controls, you can get more tonal variation than with Tone controls, go figure.
Typical is to set gain/distortion with the bright one, often straight to 10, and then some bass with volume 2, usually up to 2 with humbuckers and up to 10 with Strats.

Ive also seen many comments that you can't do that with Fender amps because the two channels are "out of phase". Im trying to understand that comment. I understand that going through one triode, e.g. on a 12AX7 from pin 2 to pin1, the output signal on the plate is 180 out of phase with the input grid.
To understand how to get one channel out of phase, is the difference only the number of stages each channel goes through?

Thatīs it, 1 extra gain stage, 1 phase inversion.

Both clean and tremolo go through 2 stages of V1. SO, after V1, both channels are in phase with each other and with the input?

Yes.

Then the reverb channel goes through a ganged 12AT7, (so flips phase "out" here),

no, only reverb signal goes through ganged 12AT7, main signal goes through a resistor which does not change phase.

one side of another 12AX7 (reverb recovery), so another phase flip so flips back in phase,

no, this one inverts main channel signal.

then the other half of the 12ax7 which looks like just another gain stage to boost some of what was lost in the reverb circuit, so back out of phase again.

this one also affects only Reverb signal, not the main one.

is this right:
clean: -> out -> in
tremolo: -> out -> in ( -> out -> in -> out).

not sure what you mean here but in any case Tremolo does not affect signal phase, itīs a (varying) resistor .

But don't all of those capacitors in the signal path cause phase delays as well?

they are "all pass" bypass caps, affect little or no phase within the Audio band.
Phase change happens at subsonic frequencies an is not perceived.

How are those accounted for?

they are not.

Would you need a simulator, like a spice or something like that to determine accurate phase relationships?

you can use pencil, paper , a slide rule or graphs, thatīs the old way.
FWIW even now in the 21st Century I still do it that way, just replacing slide rule with pocket calculator, because I was trained that way, never learnt to simulate.
Oh, we did have a computer at the University, the mighty IBM1620.
The whole room had to be air conditioned, we had no screen and no keyboard.
You punched your problem on a perforated paper tape (cards were expensive), let it at the Computer room, picked up results a couple days later, in the form of bedsheets of paper printed on the IBM electric typewriter you see there or some line printer.
Wonder why there is still a "print screen" keyboard on most modern machines?:

If someone put another gain stage on the clean side, not that I want to do that, but would that put the two channels in phase or is it more complicated than that?

Yes, they would be back in phase.
Normal channel would also turn into a gain monster.
FWIW, thatīs exactly what Randall Smith did in Mesa MK1 and Mk2: "a Fender with an extra stage".

[mikepukmel]

Thanks Juan!

[mikepukmel]

Re the 1620, one of my co-workers who just retired said his university had a similar machine (can't remember the model) that had "core memory". When he retired last year, he gave me a panel of core memory from that machine he had kept all these years a s a desk ornament. I haven't found a way to integrate it into a guitar amp yet.

[Enzo]

When you get there I have some tube-based circuits for core memory.

[Steve A.]

If someone put another gain stage on the clean side, not that I want to do that, but would that put the two channels in phase or is it more complicated than that?

Yes, they would be back in phase.
Normal channel would also turn into a gain monster.
FWIW, thatīs exactly what Randall Smith did in Mesa MK1 and Mk2: "a Fender with an extra stage".

Modern amps will sometimes use an opamp to sum up two out of phase signals. Yes, many people insist on absolutely no transistors or ICs in the audio signal path but opamps are great for summing two out of phase signals.

With all of the words posted so far I want to repeat the important point so it doesn't get buried in the discussion: the typical gain stages we see in guitar amps with the output signal taken from the plate will invert the signal so if both channels have the same number of typical gain stages (like the early Marshalls) the sum of the two will be in phase. If one channel has an even number of typical gain stages and the other has an odd number (like BF/SF Fenders) then the sum of the two will be out of phase. Cathode followers do not invert the signal so they aren't counted.

While there are many different integrated circuits, opamps will have two inputs, one in phase with the output and one out of phase, along with with a +/- power supply which can serve multiple sections in a chip with 2 or 4 opamps.

Steve A.

[Justin Thomas]

If one channel has an even number of typical gain stages and the other has an odd number (like BF/SF Fenders) then the sum of the two will be out of phase. Cathode followers do not invert the signal so they aren't counted.

And THIS is what makes the AB165 and later Bassmans so friggin amazeballs. Two stages per channel, then they're mixed, then the third gain stage... And the Bassman 10 & 100 also are mixable, but are missing that shared third stage... I added it to my Bassman 100. And it's only those models.

SUPER FAT amps, Bassmans can be...

Justin

[Chuck H]

There are mods for the ab763 circuit to enable reverb in both channels. Some are better than others. Most "correct" the channel phase issue.

[Justin Thomas]

I say correct the phase issue, stick the Ice Cube in the rca jacks, disconnect the negative feedback, and whack it on 10. If you're REALLY adventurous, turn the trem tube into a couple cascading gain stages, at least one of which should be the AB165 "compressor." Ah, hell, no I haven't done it, but it should make for a "fun time!" The Trash-Luxe De-Verb.

Okay, I'll shut up now and let the real amp builders help.

Justin