I pounded on this experiment for some more hours today and this evening and, in protest, the pedal gave up the ghost or went on strike or some such after too much poking and prodding. I think after this report I will move on to other things and come back some day to resuscitate the pedal as vibrato/chorus and leave it at that.
The most recent highlights:
(1)
Quote:
You could try replacing the V3207 with a V3208 if you want longer delay times. Might be easier than swapping pots.
This was an excellent suggestion, which I followed. But by the time I'd tracked down a V3208, my passion for the longer delay times had faded as I wanted to chase down the flanging. But some day I may go back to this (and if I ended up destroying the 3207, now I have a chip to replace it with.)
(2) I settled on injecting the signal from the emitter of Q4 into the base of of Q2 and have not recently tried any other spots, having rejected some others as stated in a previous post (months ago by now). Perhaps another spot is better for injection. So the fed back signal goes from Q4E through a 0.1 uF cap and 10k pot to the top of a C50k pot whose own foot goes through a 47k fixed resistor to ground. (All the interesting action is in the top part of the top half of that ~100k resistance to ground.) The wiper signal goes through another 10k pot and 0.1 uF cap to Q2B. The 10k pots allowed me to adjust for a good balance, allowing a noticeable effect while suppressing runaway oscillation. That was the theory, anyway. I always needed at least a few kilo-ohms in one spot or the other, but as I had guessed earlier, more than that seemed needed in the end to properly tame the response. It's perhaps a matter of taste; I had about 5k on each of the two pots when I thought it sounded ok.
(3) I added some more flying wires to the PCB, to allow me to put a resistance in parallel with the series combination R48 (47k) and "TRIM2 (DELAY)" (nom. 470k). That series pair determines the clock frequency that in turn sets the delay time through the BBD. Conventionally this is lower for flanging than for chorus. If one were to try to set up this box for chorus OR flanging, it would be better to have the DELAY setting on the outside of the box. Another way would be to switch in a lower resistance in parallel with R48+TRIM2, to enforce a lower equiv resistance and hence shorter delay. I thought perhaps this could be a fixed resistance. To find the "best" resistance I put a 1M linear pot in parallel and tried lots of settings. Of course, there IS NO one best setting; that's why flangers have "manual" knobs on them. And with the resistance too low, there's no delay (there is an upper limit to the clock frequency that the pair of chips can use) so I then added a 5.6k resistor in series with the 1M pot. Even this 5.6k was too low, but before I got around to replacing it with say a 33k, the box went on strike. But I heard enough to know that I would never have been happy settling on just a single resistance (that is, single clock freq; that is, a single delay setting) for flanging. It would really need a knob on the outside of the box I think.
(4) I've never looked at the shape of the LFO signal on a 'scope but once I got the RATE slowed down below 1 Hz by implementing the "slow" mod (increasing C19 and C20 by 0.15 uF each in my case) I got the sense that it's more like a sine function than a triangle. The filtering pauses at each end rather than immediately turning around. Modding this pedal that way is not in the cards for me ... but after I get this running again I'll probably have a look at the shape of that signal, just to know.
(5) When the delay is low (in flanger territory rather than chorus territory) even the chorus mod makes for an interesting filter, which is not surprising. If I were implementing both mods permanently I would make it possible to have both "extra signal paths" available at the same time.
(6) Nothing to do with flanging, but in making the "slow" mod I found that every time I switched in a larger capacitance for C20 the modulation would die out in a few seconds or even stop immediately. I was always able to restart the modulation by turning the RATE knob full CCW and then turning it back up, so it's not really a problem. I am assuming that this is a reflection of the details of how that oscillator works.