Determining [unity] volume for parallel fuzzes to be blended.

[ElectricDruid]

Would increasing the blend pot give more control over the variance between both sides?

Maybe, but I'd be looking at an active control, like described here:

http://geofex.com/Article_Folders/panner.pdf

(The circuit you want is on the box on the left)

Another usage example, blending between dry and wet signals in a FV-1-based digital effects:

https://www.pedalpcb.com/docs/Arachnid.pdf

The 15K resistors with 10K lin pot is standard. It's a very handy building-block circuit for doing this sort of mixing where you can choose all of one, all of the other, or any mix in-between. Since you've already got two op-amps at the end of your schematic, I reckon this circuit will save you an op-amp into the bargain (it does the inversion that you've added, and it can have extra gain if necessary).

[Rob Strand]

I'd drop any attempt at the constant-volume malarky and just implement a basic blend control that can give you enough range to cover any variations.

+1,  Constant volume schemes are a headache for transient signals like a guitar.

One scheme is to use a dual-gang pot for the drive control.  As you up the drive the second gang is used to throttle-back the volume.

[Andon]

Would increasing the blend pot give more control over the variance between both sides?

Maybe, but I'd be looking at an active control, like described here:

http://geofex.com/Article_Folders/panner.pdf

(The circuit you want is on the box on the left)

Ah yeah, I actually referenced this design in my initial post. I feel like I had volume loss toward noon from either side (as mentioned), but I'll give it another go. I used a 50K linear pot instead of the 10k and calculated the respective resistor values per the notation too.

One scheme is to use a dual-gang pot for the drive control.  As you up the drive the second gang is used to throttle-back the volume.

I had actually considered this too, but for a different idea - good to know it would be applicable here!

[ThermionicScott]

Would the volume loss in the middle be due to phase?

[ElectricDruid]

Would the volume loss in the middle be due to phase?

Quite likely. Didn't the first circuit have one fuzz that inverted and the other that didn't?

[Andon]

Correct, which I knew was an issue with certain blending designs but I didn't know that that was due to phasing issues (which I know now!), so with the added buffer and phase inverter it should be good to try again:



Full size image available here: https://i.imgur.com/c5snlD1.png

For the panning circuit with a 50K pot the resistors should be about 71K and the feedback resistor should be 241K so I went with 68K and 220K respectively - does that seem alright? I know things don't have to be exact when it's not a super sensitive/timing circuit, but I have some 75K resistors I could use in lieu of the 68K (though that's further from the target resistance). Is it better to be close but under or close but over, or does it not matter either way?

[ElectricDruid]

One simplification would be to swap the inputs over on the comparator so that it is non-inverting, and then you could drop the inverting op-amp stage before it.

In fact, if you did that, you'd be down to five op-amps, which means you'd have an op-amp you could use as a comparator, and that'd save you a whole chip.

There's one caveat, which is that we've just added a "blend" mixer which is inverting, so the circuit is inverting overall (in its current form). If that's important, we'd have to remove one inversion from each of the two fuzz paths. That's easy for the top one (just get rid of that inverting stage) but a bit more involved for the bottom path (you'd have to combine the gain of the first inverting op-amp into the filter, which would involve some recalculation since it will affect the Q).

For the panning circuit with a 50K pot the resistors should be about 71K and the feedback resistor should be 241K so I went with 68K and 220K respectively - does that seem alright?

Running through the numbers quickly it looks close enough to me. The feedback resistor only sets the overall gain (there's a loss through the previous resistor network) so its value is only crucial if you really need unity gain overall. In your situation, that's not going to matter much.

T

[Andon]

One simplification would be to swap the inputs over on the comparator so that it is non-inverting, and then you could drop the inverting op-amp stage before it.

In fact, if you did that, you'd be down to five op-amps, which means you'd have an op-amp you could use as a comparator, and that'd save you a whole chip.

Good idea! However, I'd like to keep the LM393 per the advice of others in recent comparator-related threads seeing as how it's made specifically for that function and I've run into less issues using it than when I was using say a TL071 trying to do the same thing. BUT, I will switch it over to the non-inverting input, and use the other half of the input buffer IC (originally use for phase inversion) as the op amp for the blend circuit at the end.

There's one caveat, which is that we've just added a "blend" mixer which is inverting, so the circuit is inverting overall (in its current form). If that's important, we'd have to remove one inversion from each of the two fuzz paths. That's easy for the top one (just get rid of that inverting stage) but a bit more involved for the bottom path (you'd have to combine the gain of the first inverting op-amp into the filter, which would involve some recalculation since it will affect the Q).

I'm totally fine with the pedal ultimately inverting the phase, since it won't matter when it's played in series with other pedals. Ha ha, and actually to quote you from a couple of years ago:

https://www.diystompboxes.com/smfforum/index.php?topic=120584.msg1129789#msg1129789

For the panning circuit with a 50K pot the resistors should be about 71K and the feedback resistor should be 241K so I went with 68K and 220K respectively - does that seem alright?

Running through the numbers quickly it looks close enough to me. The feedback resistor only sets the overall gain (there's a loss through the previous resistor network) so its value is only crucial if you really need unity gain overall. In your situation, that's not going to matter much.

T

I also suppose I could add a 22K resistor in series with the 220K resistor in the feedback loop to get even closer to the calculated value (70.7K x 3.41 = 241K).  I could even add some 3-3.3K resistors to the 68Ks to bring those closer to what they should be too, but that's also adding a few more parts.

[ElectricDruid]

BUT, I will switch it over to the non-inverting input, and use the other half of the input buffer IC (originally use for phase inversion) as the op amp for the blend circuit at the end.

Fair enough.

I'm totally fine with the pedal ultimately inverting the phase, since it won't matter when it's played in series with other pedals. Ha ha, and actually to quote you from a couple of years ago:

https://www.diystompboxes.com/smfforum/index.php?topic=120584.msg1129789#msg1129789

Yeah, I've never been convinced it's required. "Nice if you can" is my opinion, in short.

I also suppose I could add a 22K resistor in series with the 220K resistor in the feedback loop to get even closer to the calculated value (70.7K x 3.41 = 241K).  I could even add some 3-3.3K resistors to the 68Ks to bring those closer to what they should be too, but that's also adding a few more parts.

I doubt it's worth it. Or rather, it's certainly worth trying it without first. If it's a tragedy, then we can fix it, but I very much doubt that.

[ThermionicScott]

https://www.diystompboxes.com/smfforum/index.php?topic=120584.msg1129789#msg1129789

Some great stuff in that thread!