Interesting that they did 100% incoming part inspection on their ICs going into the arcade machines. I’m not sure if ICs were less reliable then, or they just built a higher quality product.
The recently-departed Chuck Peddle [1], designer of the MOS 6502, caused me to fall down a wiki hole for a bit and I learned that IC failure rates, right off production back in the 1970's, was something surprisingly dismal like... 70% were good? Can't find the original article but it illustrated that a LOT of silicon was melted back down and reused until MOS Tech came along, refined some of the processes to increase reliability and the industry followed behind.
Edit: found it... [2] tl;dr: MOS's advancements resulted in IC production success rates of 70% or more, and were markedly worse before (lows of a 30% success rate would still be considered profitable or successful).
> Interesting that they did 100% incoming part inspection on their ICs going into the arcade machines. I’m not sure if ICs were less reliable then, or they just built a higher quality product.
If memory serves, it was a little of column A, and a little of column B.
They also had a much higher part count in their systems than is common today, so even a low rate of bad parts would result in a lot of malfunctioning systems.
You might be surprised. Modern electronics are highly integrated; a modern smartphone is probably no more than a few dozen ICs. Vintage arcade hardware tended to rely heavily on discrete logic; even a simpler game like Asteroids had somewhere around a hundred ICs on its main board.
That semi-automatic pick and place was neat.