Messages

If I see it correctly, R17 and R16 mix wet and dry signals.
You could try using a 20k trimmer or a pot to adjust how much of wet signal you like.

You can try adding it before the two low pass filters that go to the 4024 (after C5).

I'm not sure what the function of R5 actually is. Input capacitance of the clock input is only 7.5 pF, so any low pass filter would be at a really high frequency. I don't think the signal is attenuated there, so reducing it will not help much.   



   

Based on your description, my suggestion was in the wrong direction.
I thought the octave was too low in volume. What you actually want is more stability.

Maybe you can achieve that by adding a transistor or jfet gain stage at the input.

My guess is that R5 is part of some tone shaping filter.
Since the gain of the 386 is already maxed I would play with R9.
Both R9 and R10 are 220k, so that the gain of the following OpAmp is 1.
Try reducing R9 to 100k, and you have twice as much gain for the low/sub signal.

I played a lot with the VCO and the DC voltage of the LFO.
You can get weird noises if you go too low. That would get more pronounced with the Depth maxed.
Not sure if it applies here, but checking the voltages might give a clue.

It gets faster with the LFO it seems. I would start looking around there.
Is there any noise in the matrix mode?

What about discrete OpAmps?
https://hackaday.io/project/176860-homemade-operational-amplifier
It seems there is a 4558 tube screamer kind of pedal
https://www.parasitstudio.se/stripboard-layouts/discrete-tube-reamer

I've found it helpful to gently press down on the part to make sure it's not "floating".

Yes, otherwise some weird stuff happens. When the parts are really small the have their own will, and sometimes that means 90 degrees to the solder pad.
It must be something with the wettability of the solder, because if you don't press them down they will just turn up or get stuck to the soldering iron. I will try using glue to position them first.

Nice one. Thanks for the video!
Much easier than suffering while holding the chip in position with tweezers.

Do you have a suggestion of how to solder SMD small-ish resistors?
It's a nightmare. Too much pressure on the tweezers and they fly away and I have to look for them on the carpet.
If you know of a better way of handling such parts it would be very helpful!

Cool!
Took me some time to see Loki!

Nice one!
The longest time always gets some clock noise, right?

Interesting,
it is probably the external wiring catching up some noise. Positioning is very important. I bet a shielded wire would cure it.
But, glad you found it!

This one has the negative power supply. So pin 4 is VD or VC, depending on the IC, and pin 8 is GND.

Some whine is "normal". Can you measure the clock frequency or the delay time?
That would help identify if you are above the common value of 550 ms (as in the DMM).
The tolerances of the components can also affect this, as in the original DMM.
The filters are quite on the limit of the clock frequency, so that variations in components could led to higher cut frequencies, leaving more of the whine in the remaining signal.
Check if you used the right resistors and caps at the filters located after the second MN3005.

Is there any off-board wiring? how good is it?
Is the pedal in a grounded box, or still on your desk?

Depending on how much delay time you are getting, some whine is always expected. Unless you cut a lot of delay time.
The MN3005 are only advertised for 205 ms, aren't they? Two would give you 410 ms without whine. The DMM goes way into whine territory. Without being as dark as other delays, it also lets more noise trough.

I am not sure if changing the calibration (how much signal is amplified where) improves this.

Have you checked the voltage on both sides of r4?
It should be roughly at VB.

I checked the voltages and values but that volume drop is why I even considered buying the Collo. I'll go back through but I don't have many expectations, if it's 1 of the chips I'm getting out the dremel because those are not socketed.

Before you do that, check the parts values, and if possible measure them.
Most of the time it is a wrong cap or resistor. 10k instead of 100k, or 2R7 instead of 2k7, and so on. With caps it is more difficult to check everyone, but I would also suggest that.

Sometimes your supplier just mixed some values. With 1% (blue) resistors red and brown are hard to tell apart.
With SMD parts it's even worse, caps in the nF and pF range have no number printed on them. You have to actually measure them to be certain.