Well, about the role of batteries, partially right. I also used Gamba's schematic.
Yes, one keeps RTC going. No, not only the second brings the computer up, but both of them. Something must power the 5V flipflop and that's why there are two batteries.
By using a NuBus connector, some latches and logic analyzer I made a makeshift diagnostics tool. I found many more interesting things, such as the worst video board for early Mac I encountered: The video board, which is not original (marked "SPECTRUM 24"), talks with NVRAM all time during boot. Right after boot, it reads an undocumented area looking for a byte. The byte then determines the type of display used with the board. Because f..k Apple's general standards of monitor pin-outs and control panel settings... and display switches. Now when you zap your PRAM or replace battery you land without display, as it's some unearthy resolution by default. But booting further with a "default" setting goes worse. During the system boot-up, the board is probably enumerated. So it goes again, totally confusing 7.5 and making 7.1 boot or not. It takes control and pokes around ADB for something, then it reads the byte again. If in this case the byte is not in the blessed range of supported resolutions, it throws with sad mac. Too bad that I can't read the code as the screen is garbled by, yes, wrong resolution.
I recapped the machine and it is still not stable, I'll try to re-solder RAM sockets as first one is white an rest is black - and it looks like it has been made in factory, so there could be some problem with it.
And what have I bought?
A few books. Mostly about electronics and switching power supply design, but some look like original instructions for PDP computer programming.