Op amp topology question

[bushidov]

Using the attached schematic part as a reference, I am trying to understand what is going on here. I think I got a bit of it figured out, but there is a couple things I am not sure about.
R1 is the anti-pop resistor.
R2 is the bias.
C1 and R2 create a high pass filter of around 34Hz which blocks low-frequency parasitic oscillation and some really low frequencies.

The part that does confuse me is the negative feedback portion. I believe the Gain Voltage is 1 + (R5 / (R3 Parallel to R4)), which comes out to about a voltage gain of 12, which calculates out to roughly 20dB. But...

That's assuming that R3 and R4 are going to ground. In this circuit, it is going to bias, (4.5V approx.) Also, how is the R3 and C2 related? It appears to be an RC high pass filter, but if so, it is at a whopping 16KHz, which is that to say it is only boosting the highs and not much of the lows? Also, the RC high pass filter is going to bias, not ground, so how is that effecting things?

I am a bit of a noob on audio electronics theory, but I am trying to learn it, if anyone is willing to help me on this one!

[PRR]

"Going to bias" is, for *audio*, normally the same as going to ground. (The bias is normally a big capacitor or other low-Z audio source.)

C2 is a capacitor. Infinite impedance at DC and zero impedance at infinite frequency.

You already computed that it is 1k at 16KHz. At lower frequencies it is higher impedance. At 160Hz it is like 100K. It has "no effect" next to R4=10k. Gain for all low frequencies is about "2".

What is C2 0.01uFd at 1.6KHz? About 10k. Now it has a significant effect next to R4=10k.

So we have flat gain=2 from 34Hz to 1.6KHz, rising from 1.6KHz to 16KHz with gain like 12, then flat.

Then R6 C3. Which figures as 111KHz. Perhaps pointless. Perhaps there is more stuff not shown in this snippet.

[bushidov]

Thanks for the response, PRR!

Alright, that makes sense. However, then my question would be in regards to the bias vs ground. Why does this circuit have the RC network go to bias instead of ground? On distortion circuits like the ProCo Rat or MXR Distortion+, overdrives like the Tube Screamer, or boosts like the MXR MicroAmp it goes to ground instead. What are the advantages or disadvantages of going to bias instead of ground?

As of the RC filter of R6 and C3, I too, thought that was a little ridiculous of a high pass filter. The attached image is the full schematic. It's an EQ pedal. I built it, so I know it works and even am able to verify the frequency cuts and boosts with an analyzer. I have a pretty good understanding of the gyrator circuits that make up the EQ portion and how to calculate the frequency and Q variables, so it feels pretty embarrassing to understand all of that, but not the simple biasing and filtering that is going on.

[merlinb]

Why does this circuit have the RC network go to bias instead of ground?

Partly it may be convenient for layout since the RC can go directly in parallel with R4. Or it may be because the designer was more used to seeing textbook opamp circuits that use bipolar power rails, where the RC would naturally go to ground which is of course half-way between the rails, and he copied that concept here.
However, in a unipolar pedal it is usually better to go to decouple things to ground (rather than to Vref) since ground is more reliably low impedance than Vref and is less likely to cause unwanted feedback/feedforward effects elsewhere in the circuit.

[pruttelherrie]

Or it may be because the designer was more used to seeing textbook opamp circuits that use bipolar power rails, where the RC would naturally go to ground which is of course half-way between the rails, and he copied that concept here.
However, in a unipolar pedal it is usually better to go to decouple things to ground (rather than to Vref) since ground is more reliably low impedance than Vref and is less likely to cause unwanted feedback/feedforward effects elsewhere in the circuit.

THIS has puzzled me for a looooooong time. "Where to decouple to"... Thanks for shining a light on this! So functionally it does not matter where you decouple to, as long as it is a stable and a low impedance return path?

[antonis]

So functionally it does not matter where you decouple to, as long as it is a stable and a low impedance return path?

In theory, yes..!!! 
(but there always are non-ideal capacitors and phase shift "garbage"..)

[bushidov]

(but there always are non-ideal capacitors and phase shift "garbage"..)

So, in regards to those non-ideal scenarios, is decoupling to bias better or worse than ground? I am aware of the "non-ideal" capacitor situation as a capacitor has ESR and such, but didn't think decoupling one way or the other would help/hurt for even those scenarios.

[merlinb]

So, in regards to those non-ideal scenarios, is decoupling to bias better or worse than ground?

Usually ground is the better choice since most of the opamps in our pedals are actually driving ground-referenced loads. Vref is, for the most part, only for biasing and is not a logical 'dump' for circulating audio currents. Decoupling to Vref is needlessly adding a little more impedance to the decoupling path; a path that leads ultimately to ground.

[bushidov]

Thanks merlinb!

Any idea as to the RC network of R6 and C3? It appears to be a simple high pass filter, but only high-passing 100+K of frequencies seems a little ridiculous unless the gyrator circuits that come after it are what makes it make sense?

[merlinb]

Any idea as to the RC network of R6 and C3?

It doesn't look like it is there for stability reasons (the logical place for that would be in parallel with R17), so my guess is it is a relic from an earlier stage of development. R18 is also a strange value; why dump 6dB of signal when there's a perfectly good level control right after??

[PRR]

Any idea as to the RC network of R6 and C3?

Clearly it mirrors the junk  around U1A. U1A pre-emphasizes the top of the audio band. R6 C3 de-emphasize to result in "flat" response.

This fights whatever garbage may creep-in between these two ends of the chain.

DC issues neglected. R3 limits boost above the audio band, a great idea, but is beyond guitar range.

[Rob Strand]

It doesn't look like it is there for stability reasons (the logical place for that would be in parallel with R17), so my guess is it is a relic from an earlier stage of development. R18 is also a strange value; why dump 6dB of signal when there's a perfectly good level control right after??

It's there for stability.  IIRC, there's two mechanisms:  The first is there's lot of stray capacitance around the bank of boost/cut pots.   The second is the rising output impedance of the preceding opamp.   Most units will squeal without that cap.  Look a few designs you will see it is there.  Sometimes you also see a cap across the feedback resistor of the boost/cut opamp for similar reasons.

[bushidov]

Thanks again guys! I also see where I was making a terrible mistake and I am not even sure where I got it in my head. I kept thinking R3 and C2 is a high pass filter when it is really a low pass. I don't know why I was thinking that. That's why I "perplexed" as I was thinking "No, why would someone want to cut all frequencies below 16kHz and even more "perplexed" as to why someone wanted to cut frequencies below 106kHz after that. Yeah, low pass filter. Dang, I'm a dummy.