Next Kosmodrome module is a mixer. Three inputs, DC coupled, with regular and inverted outputs and a bias control. It’s intended mainly for mixing CV signals.
I looked at some designs online but ended up coming up with my own. The MFOS Quad DC Mixer was my starting point but I made lots of changes — not the least of which was not trying to jam four of them behind a single front panel. Yeesh, crowded knobs, anyone? Besides, if I want to use four mixers, I probably want them in four different places in the synth case.
I added a general level control, to make it easier to use large input signals without clipping. Turns out having inverted output, bias control, and general level control is a trickier combo than you might think — I had to split Wilson’s first gain stage into two.
Inverted output is tricky, that is, if you want it to work nicely with the bias control. That is, if you want a positive bias to raise both the non inverted and inverted outputs. That way you can take a ±5 V sawtooth for instance, cut its amplitude in half, and add a 2.5 V bias to make it into a 0–5 V sawtooth… and the inverted output will be a 0–5 V ramp, instead of 0– -5V.
I put a trimmer into the first gain stage’s feedback loop. That allows calibration for reasonably accurate unity gain, with the general level control full up. It’s probably not something you’d normally want to use for 1 V/oct signals, but you could try it.
This was the first time I’ve built something with slide pots. What can I say, I’d been thinking about the ARP 2600. They’re soldered to the back of the main board — which means correcting assembly errors is difficult. And unfortunately, I know. First I discovered my 120k input resistors were actually 120Ω. Didn’t work too well! I had to unsolder and replace them from the top of the board only. Not as hard as I was afraid it would be, but not easy. And then I realized I’d forgotten to do something I knew I should do, which was to check the total resistances of the slide pots and use the one closest to 100k for the general level control. Pot tolerances are loose, and evidently the one I used was somewhere south of 95k, so adjusting the trimmer couldn’t get me up to 120k in the first stage feedback, which meant I couldn’t get it to unity gain. I had to unsolder the 15k fixed resistor and replace it with 22k to compensate. Again, from the top of the board. After that it all worked.
Schematics, Gerbers, and docs are on GitHub: