I was an IBM brat, we never saw much Data General stuff. When I moved to Florida in 1984 one of my first tricks was to find and fix a broken weld in a core storage array in a 3890 check sortor (to get one of those million dollar "captures" in on time). It got me some attention.
Nobody had an array in stock, it would have ended up coming from old equipment.
The core was the same as what they used in the 360 mod 25 and 30 and I was region specialist on those boxes although I had never really seen a 3890 before. Core is a very structured matrix and as soon as you develop the failing pattern you can usually point to the bad weld, even if you can't see it. I ended up using a paperclip on the end of a fine tip soldering iron to dab a bit of solder on it. 5 minutes later they were running checks.
I started on 1401s and "greasy gear" machines in 1966. By the 80s the writing was on the wall that actually fixing computers was a dead end job and I got into building computer rooms, then that became a dead end job when computers were shoved in closets or under desks.
I still like the days when it was 4 gates per card and a computer had thousands of cards in them. THAT is fixin'.
I/O was fun in those days too. No electronically controlled stepper motors. They used hydraulics to move read heads and paper.
