The PCB was pretty straightforward. The most novel part I did first, soldering the SOIC SSM2212. It went smoothly enough, didn’t even need my solder braid and all the connections were good. The tempco resistor went on top of it, with some heat sink compound to join the two.
(The tempco is R24; the empty R24 footprint is for when you are using a normal, non tempco resistor instead.)
Wilson specified ceramic for all the caps except the 10 µF (electrolytic) and 1 µF (tantalum) ones, and the main integrating cap, C6. But I chose to use polyester film for the 1 nF to 100 nF ones, aside from the two 100 nF bypass caps (only two, despite five op amps and an OTA…); for C6, polypropylene. That one turned out to be too wide for its footprint and ended up half sitting on top of a resistor… but the leads went through the holes and so I soldered it there.
Clear sailing otherwise. I also made one of my Euro/Kosmo power breakouts, with diodes but no caps since there are caps on the board.
Then I put pots and jacks on the panel, and mounted the main board using a sheet aluminum bracket.
The board’s long enough that it had to be set back to clear the jacks, so the module ended up pretty deep — not too deep for the Kosmodrome case, though.
I wired everything up, omitting, to start with, the pulse width knob voltage mod. I wanted to verify my breadboard observations first. Checked for shorts, plugged in, turned on, looked at outputs — working! All waveforms there, after some quick and dirty trimmer setting.
The pulse width knob indeed behaved pretty much as I saw on the breadboard. So I wired up the mod. I initially tried building it free form, no board. I wasn’t very happy with the result, but it sort of worked, but it didn’t. It functioned as a pair of amplifiers, as it should, but for whatever reason, despite my thinking a gain of 1.24 would allow a good margin for error and having that work on the breadboard, this time it wasn’t enough gain — shutdown wasn’t reached before the pot limit.
Not wanting to try to swap out the 24k resistors in the precarious bug I’d made for 33ks, I instead built the mod over again on stripboard, which I hot glued to the back of the pulse width pot after verifying I could now reach shutdown with pulse width range from 0% to about 98%.
The knobs went on.
The wave shapes could use a more careful job of setting, and I haven’t done the V/Oct calibration yet. But basically it’s done.
Read the rest of the MFOS VCO build series: