Well I finished up a few more vector + dumps. There's 2 new D553C games and one HD38820 game. I have 2 more to finish (another D553C and HD38800) and then I'm through all my games, except dumping the COP411 on turtles.
This one game I have, tandy kingman, has an HD38800B on it. This is the first time I have come across the B revision of this chip- it appears to be a die revision and is most likely a die shrink. It does not dump properly, and upon inspection the first 10-15 words are OK then it just sort of falls apart.
I am guessing it's because I am clocking it too slowly, so I will adjust my clocking doodad see if it behaves then. This would lend credence to a die shrink since the gate capacitance will go down making clock speed more critical. I am running it at something like 20-50KHz which is way below datasheet spec.
These chips needed a special circuit to clock them because it only uses an RC oscillator, so you have to "simulate" an RC circuit for it to work properly. You cannot just stuff a clock in because the oscillator pin is both input AND output. (it lets the resistor charge the cap up to a certain voltage, then it turns a transistor on to dump the capacitor to a lower voltage, then the cycle repeats).
Tomorrow I will try to get the last two done (kingman and astro command which is the D553C game).
Looks like I am up to 19 games done!