contemplating shared dropping resistor on LTPI plates

[imaradiostar]

A bunch of the Doctor's amps use a long tailed phase inverter with an unbypassed 10k or so resistor between the phase inverter power supply node and the LTPI anode resistors. Something tells me this is more or less a form of negative feedback to reduce the sensitivity of the PI. Am I thinking about this incorrectly?

It would explain why my EF86 based Dr. Z style amps have too much gain, as I don't recall having a resistor in that position.

jamie

[cbarrow7625]

that sounds like a correct assessment to me. That is a really clever way of providing a small amount of local feedback that never occurred to me. it would be interesting to see how that worked in a hi gain amp between two successive stages to linearize the signal a bit and scrub a bit of gain.

[Bruce / Mission Amps]

A bunch of the Doctor's amps use a long tailed phase inverter with an unbypassed 10k or so resistor between the phase inverter power supply node and the LTPI anode resistors. Something tells me this is more or less a form of negative feedback to reduce the sensitivity of the PI. Am I thinking about this incorrectly?

It would explain why my EF86 based Dr. Z style amps have too much gain, as I don't recall having a resistor in that position.

jamie

You mean like an additional B+ rail dropping node resistor, without a little high voltage filter cap at the junction of the anode resistors and this 10K?

[imaradiostar]

Yeah, that's right, Bruce. There is still a dropping resistor and filter cap upstream and the following node's dropping resistor is connected to that cap. It's almost like a vox style "branch" node but without the filter cap.

I looked briefly for commercial applications of this on hifi circuits but stopped after looking at a few schematics. Perhaps there is a guitar amp example somewhere?

It would be simple enough to put it on a switch and see how it effects to gain and tone of an amp. I must try that next time I've got one apart.

jamie

[guitician]

I don't see feedback developing on the power supply side of the plate resistors. The opposing currents mixing in the 10k would effect the change in voltage at the plate very little. Maybe it has more to do with saving a cap?

[imaradiostar]

No cap saved, just a resistor added.

I'll have to wire it up and compare on a real amp. I'll measure. I'm sure there's a difference, I just don't know how much.

jamie

[Enzo]

The cap is saved by not being there after the added resistor. Generally if you add a resistor to make a new B+ node, you also add a cap to decouple it.

[imaradiostar]

I guess it depends on how you look at it. There is already a filtered B+ node with a cap. This is an added resistor in series with the PI plates. It's normal amp topology with an added resistor.

jamie

[imaradiostar]

Well...no matter. I simulated it and I'm getting the tiniest drop in level depending on the presence of that resistor- less than a dB.

So it does nothing, I guess.

Jamie

[loudthud]

If the PI was perfectly balanced, the drop across the 10K would be constant. With the real world LRPI it would probably add some even harmonic distortion.