The
I wound up with one piece of the R2, even before it was decommissioned.
IIRC it was an interface to an IBM typewriter and that circuitry was abandoned ~1972. I asked if I could have it when I started working on the project. It brought back some not-so-good memories. The ECL used was flat-packs because they were the fastest gates that could be acquired at the time (1ns?) But the selection was poor so most of the logic was quad NAND gate chips. (Mistake 1 ? too many wires.) These things was then placed into these little plastic trays. Undereath the chips was a plastic runner-thingy that had flat conductors that you could trim to fit. IIRC that was only used for power and ground. Once that was in, the chips were put on top. Mistake 2: any mistake in the runner was permanent. Mistake 3 as I was told (as a freshmen, so long ago ?) is that solder rotted the runners. So over time you had to go back and rework everything. That mistake made the machine totally fragile. e.g. if you bumped into one of the metal frames that had the dozens of trays in it, you were guaranteed to break at least one solder joint, leading to wasted time. By comparison the cache storage unit was made up of a handful of either MOS or TTL DIPs and that was on a piece of breadboard and never needed to be fiddled with.
Here are the pictures of the interace circuitry:
