hi all,

i'm working on my first unique(ish) design -- it's an adaptation of the keeley rotten apple, which is itself an adaptation of the EHX op amp big muff. the input, sallen-key and clipping sections are largely the same, then i've added a tone section based on the neve 1073, a clean blend, and as a bonus i used the clean blend path as an optional buffered bypass. it works and sounds good on the breadboard but can you give it a once-over and let me know if i've made any mistakes or if you have any suggestions for how i could do this better or smarter?

just to preempt questions about this: having the clean level control after the distortion volume is deliberate. from a usage standpoint i would've preferred it to be before the volume but doing it this way prevents me from having to boost the clean signal ~22x and clipping the op amp to get it to a level where it can compete with the distorted signal.

any and all advice is appreciated.

thank you!

By wiring CLEAN & VOLUME pots to Vref you save two caps but you put garbage in Vref..
(if you insist on this wiring, raise C103 value..)

As for CLEAN / DIRT mix, I'd wire VOLUME pot wiper on CLEAN pot lug 1, replacing it with a linear one..

P.S.
Not enough coffee yet, so plz proceed at your own risk..!!

i have spent the last day trying to reduce crosstalk and i've had success by either
a) raising the VREF filter cap to 100uF, or
b) wiring the CLEAN and VOLUME pots to 0V GND, each across their own dedicated 100uF caps.

so i'm happy to hear someone confirming that.

i'm going to try upping the VREF cap to 220uF and see if that helps even more.

The general rule of thumb is that the input resistor values should be "significantly" larger than the pot values, if you want to avoid loading the pots. At x10 larger, you can mostly ignore it. At x3 larger, you can mostly get away with it.

You've got a 10K pot into a 10K resistor on the CLEAN path, so the pot will be loaded and the pot response will not be what it says on the tin.
The VOLUME path is worse, with a 50K pot into a 10K resistor, so that 50K pot isn't going to act anything like its designed response. You've specified "Log" (Type A) pots, but they probably won't feel much like that in this situation.

None of this has anything to do with the separate issue of the VREF and the DC bias, which Antonis has been helping you deal with. For that, you need to think about the lowest input frequency and the 1/2PiRC cutoff point of the pot and the cap. What I'm really saying is that a 100u cap sounds way too big and you'd get a perfectly acceptable result with one a lot smaller:

https://electricdruid.net/rc-filter-calc/?f=&r=10K&r_series=3&r_errors=1&c=100u&c_series=1&c_error=10

1u would probably do it:
https://electricdruid.net/rc-filter-calc/?f=&r=10K&r_series=3&r_errors=1&c=1u&c_series=1&c_error=10

HTH

i was getting a weird taper from the pots, good to know what the cause is! i'll swap the resistors out for larger values while keeping the ratios the same.

as for the caps to ground, i was going with 10uF on the breadboard because it seemed redundantly large, just so i didn't have to think about it too much, but i was still getting significant crosstalk. 100uF made it better, even though theoretically just in terms of cutoff frequency it shouldn't have made a difference.

but while we're at it, why did antonis recommend a larger filtering cap for VREF in the power section, if the 10uF with the 10K voltage divider resistors already gives a corner frequency of 1.6Hz?

Quote from: thunderaxe on March 18, 2025, 10:42:14 PMwhy did antonis recommend a larger filtering cap for VREF in the power section, if the 10uF with the 10K voltage divider resistors already gives a corner frequency of 1.6Hz?

Actually, corner frequency is at 3.2Hz (filters "resistive" part should be considered the equivalent resistance of divider's resistors parallel combination.. ) but this isn't the point here..
(BTW take a look here for further bias & decoupling info..)

Tom above deals with individually AC grounded pots (no connection to Vref hence no C103 involvement) where I was talking about utilization of C103 for both pots AC grounding..
The bigger the C103 value the lower the path to GND resistance hence the more effective signal grounding..
It has nothing to do with filters and time constants - just less disturbance on Vref..

There is nothing saying you can't make more than one VREF divider in a circuit.  You can avoid VREF feedback problems by having one for the input and clipping stages and another one for the tone control and output stages.  This gives you isolation without having to go to large values of C103.

Quote from: amptramp on March 19, 2025, 08:11:16 AMThere is nothing saying you can't make more than one VREF divider in a circuit.  You can avoid VREF feedback problems by having one for the input and clipping stages and another one for the tone control and output stages.  This gives you isolation without having to go to large values of C103.

that actually is mostly how i have it set up on my breadboard -- the original circuit (the first three stages in the dirt path -- input, sallen-key, and clipping) is on the pedal's PCB with its own power supply section, then i'm pulling power from the pedal's DC input to the breadboard where i've set up my own power section with 9V filtering caps -- 100uF & 100n, and a VREF with a filter cap. these are used for the sections i've added to the circuit -- the clean input buffer, buffered bypass, tone section, and output mixer.

if i were to set up two different VREFs in the final circuit, what would be the best sections to isolate from one another? should i do one for biasing and one for dumping the clean and dirty pot signals to ground? at that point why not just give each pot its own dedicated cap to ground? or is there crosstalk between the different sections even without the pots dumping signal to ground? i do find there is some even when i've disconnected the pots entirely and there should (to my understanding) not be any signal going into the mixing and output section, it still shows up somehow and i can only assume it's leaking in through VREF being used to bias the op amps in the three breadboarded sections

If you continue to experience crosstalk with the below configuration, you might have to but the blame on breadboard wiring..



edit: Make boths pots 100k (preferably linear taper..)

Quote from: thunderaxe on March 18, 2025, 10:42:14 PMi was getting a weird taper from the pots

I have a question about your GAIN pot - is it maent as a rheostat or a voltage divider? either way, that hanging lug should be connecting to either wiper or "ground".

Quote from: duck_arse on March 19, 2025, 09:32:59 AM

Quote from: thunderaxe on March 18, 2025, 10:42:14 PMi was getting a weird taper from the pots

I have a question about your GAIN pot - is it maent as a rheostat or a voltage divider? either way, that hanging lug should be connecting to either wiper or "ground".

it's a rheostat, sets the gain of the stage in combination with R10 and a high pass in combination with C6. the lug is disconnected in the original circuit and i trust that robert keeley knows what he's doing. that said i can connect it to the wiper. what's the potential downside to keeping it disconnected?

Quote from: antonis on March 19, 2025, 09:08:17 AMIf you continue to experience crosstalk with the below configuration, you might have to but the blame on breadboard wiring..

i was putting the caps on the ground side of the pots, are there advantages or disadvantages to doing it this way?

Quote from: antonis on March 19, 2025, 09:08:17 AMIf you continue to experience crosstalk with the below configuration, you might have to but the blame on breadboard wiring..

actually, since the next stage is also biased to +4.5V, wouldn't i need caps on both sides of the pot to be able to tie lug 1 to GND? and if so, why not just put a single cap between lug 1 and GND?

Agrrrrr...!!! 

Lost posts, again..

okay, here are my results from some of the suggestions in this thread:

-increasing the value of the summing amp resistors: the pot taper is smoother, but the signal is quieter and has lost a lot of treble. i used 560K resistors for the dirt side and the feedback loop, and a 390K for the CLEAN side, which currently has a B5K pot on it. 560/390 should give me a 1.43x boost but i need to turn the pot all the way to 10 to get unity volume with bypass, and with a much darker signal.

-lowering the value of the capacitors across lug 1 of the (dirty) VOLUME and CLEAN pots to GND:
-1uF: signal is still very present when the pot is turned to 0, but audibly low-passed. also there's a hum.

-10uF: hum is almost gone, signal is much quieter, but the very low end is still coming through

-100uF: signal is practically silent until i turn the pot up.

Quote from: antonis on March 19, 2025, 09:08:17 AMedit: Make both pots 100k (preferably linear taper..)

i gave this a try, linear didn't have the right sweep, audio-log felt much better. to keep with electric druid's 10:1 rule i upped the dirt path and feedback resistors to 1M and the clean path to 560K to try and get ~2x gain out of it. same problem as before but even more so: the signal becomes quieter -- clean pot at 10 is about unity with bypass -- but significantly duller/darker.

You'd have to adjust C15 to match the new feedback resistor. If the resistor goes larger by x5, the capacitor needs to be smaller by x5. Otherwise, you'll get heavy lowpass filtering in that stage, which would account for the darker tone, and might also be where the gain disappearred to (you cut a load of signal away with the filtering).

If you went from 10K to 560K, that's x56, so that cap comes down to only 10p - virtually not there.

Quote from: ElectricDruid on March 19, 2025, 06:27:40 PMYou'd have to adjust C15 to match the new feedback resistor. If the resistor goes larger by x5, the capacitor needs to be smaller by x5. Otherwise, you'll get heavy lowpass filtering in that stage, which would account for the darker tone, and might also be where the gain disappearred to (you cut a load of signal away with the filtering).

If you went from 10K to 560K, that's x56, so that cap comes down to only 10p - virtually not there.

ahh, that makes perfect sense.
the 560p cap was there "just in case" to filter out any hypothetical noise above 30kHz. to adjust for the higher resistors i'd have to go down to a smaller value than i even have so i'll just skip it

Why large vref cap are needed
https://postimg.cc/mPPKbN4

Quote from: thunderaxe on March 19, 2025, 07:01:05 PMahh, that makes perfect sense.
the 560p cap was there "just in case" to filter out any hypothetical noise above 30kHz. to adjust for the higher resistors i'd have to go down to a smaller value than i even have so i'll just skip it

... and that seems to have solved all my problems. A100K pots for both, 1M resistors for all three positions (i decided unity gain for the clean signal is loud enough), and lug 1 of both pots wired to VREF.

can someone let me know what would be the factors in deciding the value of the volume pots, or of volume pots in general?