Just a follow-up to my previous posts. I have not yet completed this project, but am well over half way there. Since my last post, I've done a lot of designing and redesigning of various ideas, and started work.
I've built a board containing eight additional note blocks (to expand the keyboard to 44 keys and implement high C properly), forty-four 12dB/octave RC filters to add a flute voice (i.e. more "Thingz"), and a master clock that can be switched between three crystal-controlled octaves or a variable control as the original had.
I built a second board containing two AD envelope generators, a wave-folder (controlled by one EG and/or the 1550's LFOs), two VCAs (one controlled by the other EG and one by the LFOs), and some additional buffering (for the flutes and also to allow the piano to be split just like the strings).
The new case is finished, and all the electrolytics have been replaced on two of the three boards, and the old ones removed from the third board.
I haven't written the project up yet, since it's not finished, but I have put together a page with some photos of the progress so far:
http://www.stefanv.com/electronics/paia_stringz_n_thingz.html
I found one peculiarity, which was that
both voltage regulators were 7805s instead of one each 7805 and 7905. Considering that the pinouts off the two are different, and the 7905 is a negative regulator, I was amazed that the synth worked at all (it seemed to pretty well before I started disassembling it). I sketched up what the schematic would be by wiring in the wrong regulator with the wrong pinout, and came to the conclusion that what I was getting was a nicely regulated +5VDC, but relative to the unregulated -17VDC input (i.e. I was getting an unregulated -12VDC). That explains why it worked, and also why I was getting a fair bit of AC hum on the output.