The schematics were provided by Mark Glinsky of Manual Manor.
I'm not sure what the exact differences are between the various software versions. They're straight binary files. Many thanks to Tom Collins of GTRI for making the images for me; few people around the department still had equipment that could read them! Tom included some text files comparing the different versions; I'm not sure how useful the comparison files are, but here they are anyway.
Here is the source code as a 7-zip file. (Some quick googling will find something that will decompress 7-zip files on your particular system.) Recompiling this will be quite a challenge. Lynx suggested that the macroassembler AS might be worth trying out.
See the following notes Lynx included in an e-mail:
You have my permission, as the copyright holder, to publish any of the source code or documentation, so long as it's not for profit. If you or anyone else wants to make a profit on it, contact me first and we'll come to some sort of mutually acceptable arrangement.
I never got a chance to properly complete the OB-Mx code, since they released it while I was in the hospital recovering from a heart attack, so I can almost guarantee that it's not up to my usual standards for a completed project, but it will give anyone studying the OB-Mx a good look at what the code looked like in an "almost finished" state. There is some debug code still active that slows down processing, particularly the LFO, and it should be removed, or at least disabled, if you should happen to want to actually run the code.
In any case, I hope you find this stuff useful.
[I'd also note that, although Lynx won the rights to the source code he created, as far as the whole OB-Mx goes, other parties, such as Gibson and Don Buchla, probably own rights to various bits of IP, such as the front panel designs, so I'd recommend treading carefully.]
Anyone have patches they want to share?