Reduce gain of 12ax7 to 1

[Helmholtz]

Another option could be shunt feedback.

What do you mean with shunt feedback?

[catalin gramada]

This is how I added a 20db boost stage to an existing circuit.12au7 was used in front. Passive or active pickups inputs.

[Dave H]

What do you mean with shunt feedback?

I'm guessing Chuck meant something like this

[catalin gramada]

The follower is good by principle but impedance care.

[Chuck H]

Thanks Dave. Yes, that's what I meant. I hadn't posted because I didn't have time to draw it up. But you (conveniently) attached an existing image. Thank you.

catalin gramada, Yes, impedance could be an issue. Or not. That's why I wanted to see more of the circuit.

Impedance can usually be managed with this circuit. A bigger concern (for my consideration) would be reconciling any voltage differential at the switched nodes.

[Helmholtz]

I'm guessing Chuck meant something like this

Thanks for clarification. Wasn't sure about the term.

Now, this circuit/wiring wouldn't make a good input stage because of its low input impedance. (It's fine for signal mixing, though.)

Feeding the output voltage back to the inverting input (grid) makes this input a "virtual ground" (aka "zero voltage node").
So the instrument would be loaded essentially by the grid stopper alone.
A gain of 1 with an input impedance of 500k would mean Rf = Rg = 500k.

Another way to look at it is applying the Miller theorem, yielding that the impedance between grid and ground is roughly given by feedback resistor value divided by open loop gain.

[Chuck H]

Agree about the input stage being a bad place to employ this circuit. I was considering the following triode. Ergo my inquiry about more of the circuit.

[Chuck H]

One concern with the split plate load would be the quality and condition of the power supply decoupling caps. Too much series resistance could cause noise when signal loading to the HV rail. Especially at the first triode.

[Helmholtz]

I actually favor the simple switched voltage divider after the first stage (see JMF's example).

[Chuck H]

I actually favor the simple switched voltage divider after the first stage (see JMF's example).

I agree, though...
I get unnecessarily complicated sometimes. Looking for something that might be the best solution rather than the simplest. My only concern with this method was Johnson noise with series resistance. But in this case any noise generated by additional series resistance is probably no big deal because the goal is reduced stage gain.

[Helmholtz]

My only concern with this method was Johnson noise with series resistance.

Resistor noise shouldn't be a problem here.

The noise voltage across the series resistor gets voltage-divided just like the signal.
Only the equivalent source resistance of the circuit (Thevenin) contributes thermal noise and that is something like 4.6k in JMF's example.

[Henrik]

Thanks guys!
you clearly know your stuff.
My amps tend to come together on the bench, so I haven't drawn up the full schematic yet, but the general idea is: 12ax7 input stage - volume - 5879 pentode - 12ax7 cathode follower - tone stack - ltpi - el84 pp output stage, cathode biased.
The 5879 has more gain than the 12ax7 so I expect a lot of overdrive with the volume up.
I will build it first without the switchable attenuation, then I'll try different things and stick with what sounds better.

Thanks again

[Helmholtz]

I forgot to mention that the low source resistance at the divider output also prevents audible treble loss.

[Chuck H]

The 5879 has more gain than the 12ax7 so I expect a lot of overdrive with the volume up.

This might be a misconception. A 12ax7 triode has an amplification factor of 100. And there are two of those in a nine pin bottle. Obviously there are losses due to coupling circuits, but let's just use factor 60 as an actual signal voltage amplification. That's a reliable number for a straight forward resistance coupled cathode biased 12ax7 triode. Now put them in series and you have a total tube amplification factor of 3600.

All the data sheets I found for the 5879 tube indicate the triode operated amplification factor (21) but for the pentode only a dash is shown (?). Regardless, it's not 3600.

[Helmholtz]

For a pentode gain is found from transconductance times total load impedance.

The 5879 datasheet (https://frank.pocnet.net/sheets/079/5/5879.pdf) lists a practical gain of 150.