Reduce gain of 12ax7 to 1

[Henrik]

Hi all!
I've got this idea for a new amp running around in my head. I'm trying to figure out a way to make the gain switchable without the complexity, noise and increased tube count of two switchable channels. The fist stage would be a pretty standard 1/2 12ax7. This then feeds the next stage through a coupling cap and volume control. What I want to achieve is the effect of plugging the guitar directly into the second stage for cleans, without having to switch the input from one stage to the next, as this is bound to add noise. My thinking was that maybe I could make some components around the first stage switchable, so that one setting is the standard gain stage and the other is with greatly reduced gain. I suspect that this cannot be done without that first stage heavily impacting the sound, but I thought I'd throw the question out there. Can a 12ax7 stage be wired so that it has next to no gain, but remains fairly transparent?

Cheers,
Henrik

[Helmholtz]

The input triode essentially acts a clean voltage amplifier having a gain of up to 60.
The easiest way to lower gain is to use a (switchable) voltage divider after the input stage.
Or use a second vol. pot and switch between both.

But in any case there will be a jump in volume/sensitivity. So you might want to add a switched second MV to compensate.

[Henrik]

The input triode essentially acts a clean voltage amplifier having a gain of up to 60.
The easiest way to lower gain is to use a (switchable) voltage divider after the input stage.
Or use a second vol. pot and switch between both.

But in any case there will be a jump in volume/sensitivity. So you might want to add a switched second MV to compensate.

Thank you. I thought about that and it's probably what I'll end up doing, but it will add components in the signal path. What I was fishing for was whether I could switch something in/out at the input stage cathode to basically cut the gain of that stage dramatically, without too much of an impact on tone.

[Helmholtz]

.., without too much of an impact on tone.

A resistive voltage divider by itself doesn't change the frequency response.
But together with the input capacitance of the second stage some HF rolloff may result.
A fixed voltage divider can be perfectly compensated for that effect.

Another method would be using a split plate resistor for the first stage, e.g.1.8k on top of 100k for a gain ~1, and taking the lower gain signal from the 1.8k resistor.
(That's actually a voltage divider as well, but the low 1.8k source impedance will avoid audible HF rolloff.)

[J M Fahey]

it will add components in the signal path

Well, that´s inevitable .
That is not the real problem, but whether it affects sound or not.
Up to clipping, the stage will be "transparent", at least in the Guitar amp World.
And a normal passive Guitar will not clip a single 12AX7 stage.

What I was fishing for was whether I could switch something in/out at the input stage cathode to basically cut the gain of that stage dramatically, without too much of an impact on tone.

Sadly not that much range available there.
From unbypassed cap to fully bypassed (say 10uF cap) we are taking around 6dB difference, not enough for what you need.

Best i "helmholtz"´s suggestion.

Your stage will have gain about 50-60X ; apply same attenuation after it, say 100k>2k2voltage divider, and you switch 2k2 to ground or not.
Connect 220k across switch contacts so DC does not build up and POP.

And a switchable Master or 2 selectable Master pots so overdriven does not jump uncontrollably loud.

[Henrik]

Something like this? Sorry about the size, but you can zoom in.

[Helmholtz]

Something like this? Sorry about the size, but you can zoom in.

Yes, that's what I meant. Adjust the 1.8k to desired low level.
BUt you should add 1M charging resistors to ground for the coupling caps.

[eschertron]

Aren't there some Marshall designs that put any extra triode stage into the preamp using contacts on alternate input jacks? I don't recall offhand which amp that was. Somewhere in Marshall's history between being a Bassman clone and a high-gainer.

edit: Or is this the solution what you wanted to avoid?

[trobbins]

One option would be to use a self-biased cathodyne stage, and switch the cathode end arm resistor to a low/zero level to make a typical first stage. But that would require cap coupling and the switching may cause a signal bump as coupling caps change their dc bias levels.

[Chuck H]

A more complete schematic of the project might help. There may be some options that are being overlooked.

[catalin gramada]

I think is possible switching minimal components around as an plate follower or simple load the output in excess if the swing required in first stage is minimal. Why a voltage divider didn't help ? If yo were more consistent put a buffer at input and that.s why the guitar pot was made for.

[Henrik]

Thanks everybody, for your suggestions. I will try the two suggested methods of attenuation - at the plate or post coupling cap - and see how it goes. I'm working on the design now and hope to do some building during the holidays.

Cheers

[J M Fahey]

Something like this? Sorry about the size, but you can zoom in.

No, that will pop big time.

You are switching between points separated by 100V or more, not properly discharging DC to ground, using more caps than needed, etc.

Here, I drew it in full for you to avoid doubts.

This one is always referred to ground, all switched points at same DC potential: 0Volts, can be made with a hand toggle or relay contacts, etc.

Attenuation can be freely adjusted varying attenuator values, I chose these because they will minimize interaction without need to adding buffers.

[Helmholtz]

No, that will pop big time.

You are switching between points separated by 100V or more, not properly discharging DC to ground...

With the extra resistors (post #7) the coupling caps will always stay charged and their switched ends will be at 0VDC.

[Chuck H]

Another option could be shunt feedback. Which comes with 100% local NFB and should make the shunted stage almost invisible in the noise floor.