Tube screamer... but dual supply positive and negative

[andreaagnole]

Goodmorning to everyone! my first message here even if I'm following you since long time... this place is really full of information, I just love it! I searched everywhere but it seems that my request is so strange that no one ever asked something similar :-)
I'm DIY building a mixer/multieffect board and I would like to place there also a distortion/overdrive stage: since I love the sound and quality of distortion of the tube screamer, I'm starting from that schematic. I'm sure you all know this link: https://www.electrosmash.com/tube-screamer-analysis
Here's my problem... everything that I already built is based on a dual power supply system positive and negative +/-15v because this allows me more headroom for my opamps. The tube screamer circuit is based on the classic "dirty" way to supply the opamp that is 0v - 4,5v - 9v using the 4,5v as a reference instead of the ground. The 4,5v are generated just with 2 resistor between 0v(ground) and 9v. I don't want this for many reasons: the first one is that I already have a good (better) power source (+/-15v), the other one is that I can't have part of the circuits with 4,5v as middle reference and part of the circuit with ground as a reference... these two things are just not working together :-).
So, my question is, looking at the schematic of the tube screamer I linked (section "clipping amp" and " tone/volume stage), do you have suggestions how to adapt this circuit to my voltages?

Thanks!!!

[ElectricDruid]

It'd be a standard conversion from single supply to dual supply. So the general rules are:

1) Anything that goes to +V stays on +V
2) Anything that goes to Ground moves to -V
3) Anything that goes to Vref moves to Ground

There's always a few exceptions, generally where things that *would have been* 4.5V in the single supply circuit have been moved to Ground instead (this is possible where they're isolated from DC by capacitors). In this case, that applies to R4 in the clipping stage, and R8, C5, and P3 in the tone stage. All of those can stay attached to Ground.

That's pretty much it, and in this case, there's probably more exceptions that rules!

Incidentally, it's generally easier to go this way around, rather than from dual supply to single, which tends to be worse.

[intripped]

It'd be a standard conversion from single supply to dual supply. So the general rules are:

1) Anything that goes to +V stays on +V
2) Anything that goes to Ground moves to -V
3) Anything that goes to Vref moves to Ground

There's always a few exceptions, generally where things that *would have been* 4.5V in the single supply circuit have been moved to Ground instead (this is possible where they're isolated from DC by capacitors). In this case, that applies to R4 in the clipping stage, and R8, C5, and P3 in the tone stage. All of those can stay attached to Ground.

That's pretty much it, and in this case, there's probably more exceptions that rules!

Incidentally, it's generally easier to go this way around, rather than from dual supply to single, which tends to be worse.

thanks for the clear explanation.
just a couple of questions:
you wrote R8 but you meant R10, is it correct?
and what about RC? Does it go (stay) to ground as well?

thanks!

[ElectricDruid]

Sorry, I can't find R10. Where's that? (I found two R11's though!!)


R8, C5, P3 can all stay connected to ground. R9 should move from Vref to Ground.


In the output buffer, yes, RC would stay connected to ground.

[ElectricDruid]

Ah, I think I've got it...you're referring to this image, right? Where they've "helpfully" changed some of the reference designators...sigh!



So yeah, R10 is R8, same thing.

[intripped]

Ah, I think I've got it...you're referring to this image, right? Where they've "helpfully" changed some of the reference designators...sigh!



So yeah, R10 is R8, same thing.

thanks, yes i was referring to this image. Sorry i didn't notice the reference designators issue...

[andreaagnole]

Thank you very much for your help.
I was aware of this rule:

1) Anything that goes to +V stays on +V
2) Anything that goes to Ground moves to -V
3) Anything that goes to Vref moves to Ground

But I didn't know about the "exceptions"... at the end of the story, applying all the "exceptions" I will find myself in the situation where, the only point where I can apply the negative voltage is the supply of the opamp... and even if it will probably make sense, it looks weird. You can also take a look at my schematic (only the clip stage and stone stage is implemented in my circuit). Thoughts?

[antonis]

R9 is retundant..
(I do know it exists in original circuit, but..)

R5 should be bigger and placed before R90..

[duck_arse]

..... and, welcome to posting, obviously.

[andreaagnole]

Thanks for all your suggestions! I made a board and I'm going to try this configuration soon!
Thanks!

[andreaagnole]

R9 is retundant..
(I do know it exists in original circuit, but..)

R5 should be bigger and placed before R90..

Can you please explain the reasons for these modifications?

[ElectricDruid]

But I didn't know about the "exceptions"... at the end of the story, applying all the "exceptions" I will find myself in the situation where, the only point where I can apply the negative voltage is the supply of the opamp... and even if it will probably make sense, it looks weird.

Yeah, perhaps it's a bit weird at first, but you'll get used to it! In a dual supply circuit, there's little need to connect anything to the power supplies except for the supply pins of whatever chips you're using. That's a *good* thing!

You can also take a look at my schematic (only the clip stage and stone stage is implemented in my circuit). Thoughts?

The circuit looks good, but Antonis' comments are sensible.

Can you please explain the reasons for these modifications?

R5 sets the input impedance of the clipping stage, and since you're using that as the main input, you need a higher impedance than just 10K. Change it for 1M.
Moving it in front of R90 prevents it from acting like a voltage divider, but if RA=1K, and R5=1M, I don't actually care much about that. The voltage divider effect will be minimal.

R9 acts as a voltage divider with R8. Since you don't really need to reduce the level by 9.1%, it doesn't actually do anything useful, so it could go.

The circuit looks good. Go for it!

[fryingpan]

You know you aren't going to gain much from increasing the voltage supply right? Anyway, a good modification to the original could be swapping the 4558 opamp with at least a 4580, or better yet, an NE5532. You have the voltage to do it. You'd get lower noise and... well, that's it really. But since you can...

[antonis]

R9 acts as a voltage divider with R8. Since you don't really need to reduce the level by 9.1%, it doesn't actually do anything useful, so it could go.

Thanx Tom but, as I might guess, OP doesn't get why it doesn't do anything useful..

[fryingpan]

Isn't R8 there to prevent opamp oscillation because of the capacitive load?

[ElectricDruid]

Isn't R8 there to prevent opamp oscillation because of the capacitive load?

R8 and C5 are a lowpass filter, so no. C5 wouldn't be there unless R8 was there, not the other way around.

It's R9 that we're thinking of getting rid of, not R8.

[fryingpan]

I suppose R9 is there to get rid of any possible offset (amplified by the previous stage).

[ElectricDruid]

I suppose R9 is there to get rid of any possible offset (amplified by the previous stage).

Sorry, you've lost me. How does it do that?

[m4268588]

Isn't R8 there to prevent opamp oscillation because of the capacitive load?

Absence of the R8 also affects the tone control. It will only be boosted and will not be cut.

I suppose R9 is there to get rid of any possible offset (amplified by the previous stage).

Output impedance of IC11A is very low (NFB is applied), it's only 0.91 even if the R9 corrects the offset.

In the stock TS (single supply), R9 can be harmful.
If the diode clip is asymmetric, R9 will vary the Vref according to the input signal. In reality, diodes can be asymmetrical due to tolerance errors.

[fryingpan]

Isn't R8 there to prevent opamp oscillation because of the capacitive load?

Absence of the R8 also affects the tone control. It will only be boosted and will not be cut.

I suppose R9 is there to get rid of any possible offset (amplified by the previous stage).

Output impedance of IC11A is very low (NFB is applied), it's only 0.91 even if the R9 corrects the offset.

In the stock TS (single supply), R9 can be harmful.
If the diode clip is asymmetric, R9 will vary the Vref according to the input signal. In reality, diodes can be asymmetrical due to tolerance errors.

Varying the Vref according to diode asymmetry, while introducing some low frequency modulation of Vref, should rebalance the diode clipping to symmetric. I don't see how it can do any harm.