PT-80 with mods - bypass bleeding

[MrStab]

Hi,

I'm trying to use a 4066 switch CMOS to try and implement 2 mods to the PT-80, with the use of latching SPDTs. Tails and Wet-Only. It seems to work overall, but I can't completely block the 571's input with the 4066, as there's always a tiny bit of fresh delay coming through in bypass.

here's what i have (please assume that power supply pins and unused control pins are properly terminated!):



When activated, the "Bypass" switch passes signal to the delay stage and allows the clean signal to be toggled by the "Wet Only" switch. When off, the Bypass switch forces the clean signal on.

Manually, it takes two 2M2 resistors in series at pin 11 of the compander to fully block the sound, but I figured a CMOS switch would be a lot higher than that. at points i've wondered if it's actually something coming back to pin 12 (via. the mixer & Repeats pot), but there was zero bleed with that 4M4 of fixed resistance. whenever i try to ground pin 11 of the 571, the delay does disappear, but the tails cut out in too sudden a way. Weirdly, grounding the "outside" of the 1uF cap before pin 11 doesn't do anything to prevent the bleed in bypass (using a resistor to avoid shorting the clean signal ofc).

i'm not using those exact switches on the 4066, i'm using the two on the other side, but it should be the same. all i can think of at this point is crosstalk on the 4066 die, not that i know loads about that. anything i might be doing wrong?

FWIW i'm not using the charge pump or 12V regulator, but ofc that's not relevant here.

cheers!

[MrStab]

well, i went back to stock, completely removed the connection to pin 11 of the 571 and i'm still getting the bleed. either i'm chasing a really tiny solder bridge, or the clean signal's going from the mixer's inverting input --> the Level & Repeats pots --> 571 pin 12. is that a thing?

i'd stick a buffer between the Delay pot's wiper and the mixer input, but i'm all outta steam for tonight

[PRR]

Tack 47K or 33K across the 510K resistors. Is the leak less?

The switching used can only be "infinite" if the CMOS went to infinite. It gets real darn high but not infinite. Also at this level the pFd of switch (and layout!) capacitance will leak a lot into 510K. Less for a few dozen K.

The downside is that when ON the CMOS distorts more for lower loads. But studio consoles were clean enough with CMOS into 100K. I'm sure 31K won't suck, and will give you an idea if the 510K is the problem.

The next step would be to use the un-used switches, with inverters, to hard-ground the line you want to make silent, a SPDT switch.

[MrStab]

turned out it was a dumb error as usual, Paul - thanks a lot for the suggestions, though. i was off by one row on stripboard and the clean signal was feeding into the Delay pot. gah!

works great now, exactly as shown in the diagram. Tails bypass and Wet Only functionality are fully independent. Negligible micro-pop at high volume, which could probably be removed altogether by decoupling/biasing the output of the first op-amp.

Tangent:

while we're at it, here's another mod, which seems to work well. Somewhere on DIYSB, Mark Hammer suggested an LPF in the feedback path to prevent runaway feedback without reducing the number of repeats. originally this was done just before pin 12 on the 571, but i moved it upstream to avoid the Repeats pot affecting the corner frequency when turned. 1.1KHz was about the most i could preserve, hence the values.




thanks again!