I’m now thinking about whether to change my plan for the (first) Kosmodrome VCO. What’s prompted this is Sam saying he’s going to consider a Mark 2 LMNC VCO. Sounds like it would address the output levels, add a sine output, maybe ditch the 3340 pulse wave in favor of creating one from the triangle wave, and possibly add a sub or wave shaping.
That would put it considerably closer to what I have in mind. My remaining issues are the octave switch and the tuner. I’d like more than five positions on the octave switch — I was thinking nine. It’s possible I could do that as a mod, though. The tuner’s a clever idea and I can see how it might be useful in stage performances, but for me I think it makes more sense (when tuning by ear isn’t good enough) just to use an external tuner and save the complication/part count/panel space. Then again, if I were to build the LMNC Mark 2, I presumably would repanel it to fit the look I want, so I could just omit the tuner and not have a big rectangular hole to show for it.
Question is, though, when would this Mark 2 be available? Not for several months at least, I’m sure, and then the next question is, when would I want it? Probably… not for several months? Maybe?
Well, I’d hate to go too long without an oscillator in Kosmodrome. Patching one over from Kosmogenesis (the first case) would get old fast. So maybe I should consider Plan B: Build a different oscillator and save the LMNC Mark 2 (or my own 3340 design, if I decide on that) for the second oscillator.
Which different oscillator? I’ve looked at several and the main contenders are the MFOS VCO and the YuSynth VCO. The latter’s a simpler design. I wonder if it’s too simple? For instance, the YuSynth tri/sin converter is this:
and the MFOS is this:
There are three trimmers there — bias, shape, and trim — versus none for the YuSynth! Does it really need three trimmers? I don’t know. On the other hand, I’d think at least one would be in order. (One trimmer’s labeled “sine-trim” on the YuSynth PCB, but in the schematic it’s “sine/tri symmetry” and comes before the tri out buffer. There are three other trimmers, “init freq”, “V/oct”, and “saw bias”. On the MFOS, besides the three sine trimmers, there are “1V/oct scale”, “high freq. compensate”, “ramp offset”, and “tri offset”.)
The differences in the pulse converter are interesting. The YuSynth just puts the saw through a comparator, with the level set by a 25k pot between two fixed 7.5k resistors connected to ±12V. In the MFOS the same sort of thing happens, except that the voltages for the two ends of the pot are apparently obtained by sampling and holding the positive and negative peaks of the ramp wave! I guess Wilson’s concern, or his experience, was that the ramp wave amplitude isn’t exactly constant — perhaps it varies a bit with frequency? In which case the MFOS design in principle should work better, but is it overkill?
Another consideration is that the MFOS is more thoroughly documented.
Both designs have coarse and fine tuning knobs, no octave switch. Well, one could be added, I suppose, or use a separate octave switch module — I have an initial design for one. The MFOS panel lacks the level controls for PWM, linear FM, and exponential FM CV which the YuSynth panel has, but those would be trivial to add.
If I do build one of these I’m leaning in favor of MFOS. But that’s not decided, and neither is it decided not to go the 3340 route for the first VCO.
Another VCO that caught my attention was Haraldswerk’s Trapezoidal Quadrature VCO and its associated waveshaper. Instead of a ramp or triangle core, this has a core that generates a trapezoidal wave. In and of itself, that’s not very useful: little harmonic content and what there is is odd harmonics only. But it generates those waves in four quadrants — at phases of 0°, 90°, 180°, and 270° — and by combining different quadrants you can create triangle and sine waves in four quadrants, and ramp and pulse waves in two. The main selling point is that the trapezoidal core lends itself well to through-zero modulation — it can go down to zero Hz and on into negative frequency. Why would you want to? The discussion linked above gives some applications. Not sure they’re really something I’d be interested in, but they might be.