After initial setup the game ran for a little while but then no longer booted. Swapping in the spare game PCB resulted in the same - still no boot. Since the Galaxia game PCB uses 2708 EPROMs with +5V, -5V and +12V supplies I suspected there may be power issue on the +12V or -5V. A meter confirmed that the -5V was low at -1V. Fitting the spare 1B1126 regulator PCB brought the game back to life.
Show day 2 encountered graphics problems. It was already running on the spare game PCB but both game PCBs had developed graphics issues :(
Regulator PCBs don't fail very often, it has large heat sinks and most cabinets (though not this one) have a fan blowing across it. I didn't have any special test harness setup for them and used a JAMMA power dummy load board and a couple of test leads to hook up the +12V DC lines to the AC input. Polarity doesn't matter - the bridge rectifier adapts it and I tested both polarities to make sure the bridge rectifier was working. The DC input to the 7905 measured -10.73 VDC and output -1.72 VDC. I suspected a bad 7905 and replacing it fixed the -5V supply.
On the bench the game PCB displayed repeated characters across the whole screen.
Attaching the Arduino ICT found that both ROM and RAM tested OK. Using "RAM Write All AD"
to set a known pattern of address == data across all the RAM and inspecting IC 3D (2708)
found:
Returning to IC 1C (2114) didn't reveal anything obviously wrong on the scope, even with "RAM Write All AD". The counter ICs 8D,10D,8E,10E outputs were all active. Still suspecting another Signetics 74LS161 was bad, all were tested with an HP 10529A logic comparator that flagged IC 10E (LS161) as bad. Replacing it completely fixed the graphics.
This Galaxia PCB had very similar graphics issues to #004 fixed earlier - doubled screen. Jumping in with the HP comparator on ICs 8D,10D,8E,10E (LS161) didn't find any issues this time. The game display showed the missiles were working properly (not doubled) confirming that the problem was slightly further downstream in the character graphics. Checking IC 1C (2114) discovered that the scope lead touching pin 16 (A8) affected the graphics and the scope showed a poor noisy signal. Tracing this back to IC 5C (LS157) output pin 4 found the corresponding inputs looking OK. I suspected IC 5C was bad and a quick piggyback check fixed the graphics. Replacing IC 5C (LS157) fixed the game.
Temporary fan assembles were built for all three cabinets in this style (Astro Wars, Super Galaxian & Sea Scare), described on the Astro Wars maintenance page.