Intersound IVP, trimming down to just the Tube Voice circuit... a few questions

[aion]

I'm working on a pedal conversion of the Intersound IVP Instrument Voice Preamp, specifically extracting the Tube Voice portion as a standalone effect.

Original IVP schematic:


Hacked-down version:


Not shown in mine - a charge pump to run it off +/-9V. I figured +/-15 was a little overkill for a pedal conversion.

As you can see, I kept the Pregain, Bass/Treble, Drive, and Master controls in this new unit. A few questions.

1. Is Pregain really getting me anywhere, since the Drive control is really just a gain control? (i.e. it doesn't have any particular control over the amount of clipping except insofar as it drives the following stages into clipping) I'm wondering if I'd be better off re-configuring IC1A for unity gain and dropping the pot.

2. I feel like the number of stages can be simplified even further than what I have. I want to keep the output in phase with the input, so I need either four inverting stages or two. The Baxandall gives me one, and the tube drive stage gives another, but I'm wondering if IC2 (A and B) are really doing anything worthwhile. Originally IC2A (and the Class AB transistor arrangement afterward) was driving a pre-gain effects loop, and IC2B was the buffer for the effects return, so it seems they would not be necessary. But I've never really worked with Class AB transistor stages before and I'm not sure if the IC2A/Q1/Q2 arrangement is contributing anything to the tonal color of the circuit.

Would I lose any of the character of this effect by eliminating IC2, Q1 and Q2, and then changing IC3B toward the end of the circuit so it's non-inverting?

3. Would a TC1044 (20mA current handling) work fine here?  I'm guessing 2-3 dual op amps from the 1980s wouldn't stress it too much, especially since it only has to carry half of the supply.

4. Anything else look odd or potentially vestigial from the original circuit?

[Rob Strand]

My main beef with hacked down version is driving the tone control from volume control which is a high-impedance source.  It will affect the behaviour.  Given EQ is part of the character you probably want to preserve that.

I'd be more inclined to go:

Input Preamp -> Vol Control -> Buffer ->  Tone Control -> Tube Drive gain ->  Tube drive -> Simple Opamp stage.

There's no real need to keep the high-drive stage at the output.
The last stage can be inverting or non-inverting.   You can flip the transformer windings to get the right phase.

BTW,  there was some uncertaintly regarding the 43R resistor.  IIRC it might be 430R.  There's some old posts on this forum regarding the issue and also some stuff about the transformer.    The issue was not resolved.

[Rob Strand]

This pic seems to resolve the 43R vs 430R issue,



Look at the two resistors just above the two transistors.  The 43R/430R is the one on the left.
To me that looks like 430R.   (I traced this thing years ago and my trace had 430R.)

[aion]

BTW,  there was some uncertaintly regarding the 43R resistor.  IIRC it might be 430R.  There's some old posts on this forum regarding the issue and also some stuff about the transformer.    The issue was not resolved.

It's definitely 43R in the schematic, at least comparing the '79 schematic and the '81 schematic ('81 shown in my post above) where it's 43R in both. The '81 schematic appears to have been heavily reviewed & revised, at least based on the notes in the upper right & lower right, but that value stayed the same between versions, so I'd be surprised if that one slipped through the cracks.

Edit: But in practice if the actual units are 430R, that does take precedence! Very good find, those are the type of things that are very easily lost to history even with good schematics.

[PRR]

> specifically extracting the Tube Voice portion



Jazz that up to taste/need.

[aion]

Rev 2 based on Rob's comments...



- Is C8 (after the first opamp) doing anything here, or can it be omitted and the stages coupled directly since it's on a bipolar supply?

- Is the transformer -> buffer transition OK? Does it need a series resistor or coupling capacitor or anything? I've never dealt with transformers outside of a power supply before so this is new territory for me.

- Is R4 (input of final opamp stage) necessary, and a reasonable value? Again, I haven't worked with transformers much on the audio side, but it seems like the secondary (10k per datasheet) could possibly serve as a resistor to ground, giving the opamp its ground return.

[teemuk]

I would replace the transformer with a differential stage. Given limited frequency range of guitar and very low operating voltages that' s the only thing the coupling transformer does.

[Rob Strand]

Getting rid of the input volume pot is a good idea.

- Is C8 (after the first opamp) doing anything here, or can it be omitted and the stages coupled directly since it's on a bipolar supply?

In some cases the input cap does roll-off the lows a tad especially when the bass control is advances.  In this case it is quite a large value so I suspect if you remove the input cap on tone control and perhaps replace the  10uF cap on the input preamp to say 47nF it will work out nicely.

- Is the transformer -> buffer transition OK? Does it need a series resistor or coupling capacitor or anything? I've never dealt with transformers outside of a power supply before so this is new territory for me.

- Is R4 (input of final opamp stage) necessary, and a reasonable value? Again, I haven't worked with transformers much on the audio side, but it seems like the secondary (10k per datasheet) could possibly serve as a resistor to ground, giving the opamp its ground return.

The stage following the transformer has an effective input impedance (to ground) of 47k+4.7k+4.7k.  So a 56k resistor in place of your 100k (R4) would technically match.  In the light of the different transformer the only way to come up truely equivalent R4 would be to measure the original unit + original Tx + 56k and tweak R4 for same frequency response.   I  wouldn't be overly concerned about these fine details.

Perhaps more important are the changes from 15V (14V?) to 9V and the use of a low gain transistor for Q5.   

The current through Q5 is set by R40, R41, R42.  When you drop the rail voltage it lowers the Q5 current because the voltage on the base (set by R41 and R42).   Even more of a problem is the original circuit uses a high gain transistor and the values of R41 and R42 are high-ish.   When you replace the original transistor with a lower gain 2N3904 the Q5 current drops a *lot*. 
So the changes would be:
- The supply voltage change alone would require R42 to be reduced to 270k.   But that doesn't allow for Q5's gain.
- So with 9V and the low gain transistor, R42 needs to drop to 75k and R41 to 10k.
- Given you are playing around with the values anyway it is possible get rid of odd values like 430R,75k etc. as well.

[aion]

Perhaps more important are the changes from 15V (14V?) to 9V and the use of a low gain transistor for Q5.   

Ah, I apologize - ignore the labels on those, the 3904 and 3906 were just the default values for the NPN and PNP footprints. I haven't hunted through the datasheets yet to find good modern alternatives, but was planning on sticking as close as possible to the specs of the original parts.

Assuming transistors identical to the original, which resistor values should change in consideration of the 15V->9v conversion?

[Rob Strand]

Assuming transistors identical to the original, which resistor values should change in consideration of the 15V->9v conversion?

This one:

The supply voltage change alone would require R42 to be reduced to 270k.   But that doesn't allow for Q5's gain.

The idea is to trim the voltage on the base voltage divider.  If the base voltage is the same, the emitter voltage is the same and then Q5's collector current is the same.

[cnspedalbuilder]

No expert opinion's here, just wanted to say I can't wait to get the PCB for this!

[aion]

Assuming transistors identical to the original, which resistor values should change in consideration of the 15V->9v conversion?

This one:

The supply voltage change alone would require R42 to be reduced to 270k.   But that doesn't allow for Q5's gain.

The idea is to trim the voltage on the base voltage divider.  If the base voltage is the same, the emitter voltage is the same and then Q5's collector current is the same.

Thank you - this is a huge help. I think I have all my questions answered. I appreciate it!