mcandrewsoun
Well-known member
Ive repaired a handful of these so far, and want to post some observations.. Most of this applies to BW/Space Duel also. Not sure about MH/Quantum and Tempest, as those are somewhat different implemetations.
Vector boards can bring a whole new level of suck to repairing. If theres a fault in the vector logic, the board will boot, run (sometimes), but display nothing..and have no logic activity to debug. Having a reference board to verify roms and rams and avg is verfy helpful. A signature analyzer is a HUGE help. When I had my fluke it was pretty useless other then RAM/ROM tests.
The most useful tool debugging these boards is the signature analyzer. A logic comparator is pretty useful, but I think sometimes you get a false negative. I found 3 bad 194s in the data shifters using signatures.. I really thought it was somewhere else, what are the odds of 3? Well, I replaced them 1 at a time, testing between each, and turns out all 3 were bad. To be fair the comparator also tagged these chips as bad also.
IF your signatures are unstable, make sure the ROMs are good, and legs are clean.
Heres some tips:
Im assuming the Monitor, power supply, and ROMs have been verified.
If your getting watchdogging:
1. Pull the pokeys. A bad pokey can cause the game to not boot, test mode has no video and a solid tone.
2. Swap the AVG.
3. Vector timer circuit. If the timers arent running, I think it can trigger a watchdog. (This happened on a AD, so not sure if it applied to color vectors).
If no watchdog, but video:
1. Pull the vector ram, then boot to test. If you get the beeps, then program code is running, and busses are communicating.
2. If the rest of the board seems like theres no activity, ground the ~SA, test point. This should generate traffic in the logic circuits. Usually this is sufficient to look for dead or stuck logic. However, there are lots of normally high or low logic points,so you have to be careful. Look at the schematic.. Anywhere you see 8P54 as a signature, that means logic high.. 0000 means logic low.
3. If your getting crazy or no output, Look at the vector data shifters. They are the link between the data bus and the DACs. They also feed back to the AVG, but in SA mode, the AVG ignores the feedback.
Other:
DC offset on the output:
You should be able to center the signal around a 0v origin (using a scope). If the signal goes negative immediately, its a bad op amp.
BW Crashes when a game starts.
As soon as you get the web, the game goes into a black hole.. This was a bad pokey. I think the pokey generates the random numbers that decide where the enemies are, and if it cant, it gets confused. The board wont watch dog in this situation. FYI my board always reports bad pokeys in test mode, even with brand new ones.
Wierd vertical or horizontal smearing
(like this) http://www.flickr.com/photos/mcandrewsnd/8101429777/in/photostream
Check the digital switch E10
Here is a Tempest nugget:
Game would reset, or shut down intermittantly after a while, ~Reset held low or goes low. This ended up being a bad big blue. The board uses the 10.3 to trigger the POR. The regulated +5 looked fine, but the 10.3 had enough ripple on it to prevent the power supply stabilization circuit from taking the POR signal high.
Vector boards can bring a whole new level of suck to repairing. If theres a fault in the vector logic, the board will boot, run (sometimes), but display nothing..and have no logic activity to debug. Having a reference board to verify roms and rams and avg is verfy helpful. A signature analyzer is a HUGE help. When I had my fluke it was pretty useless other then RAM/ROM tests.
The most useful tool debugging these boards is the signature analyzer. A logic comparator is pretty useful, but I think sometimes you get a false negative. I found 3 bad 194s in the data shifters using signatures.. I really thought it was somewhere else, what are the odds of 3? Well, I replaced them 1 at a time, testing between each, and turns out all 3 were bad. To be fair the comparator also tagged these chips as bad also.
IF your signatures are unstable, make sure the ROMs are good, and legs are clean.
Heres some tips:
Im assuming the Monitor, power supply, and ROMs have been verified.
If your getting watchdogging:
1. Pull the pokeys. A bad pokey can cause the game to not boot, test mode has no video and a solid tone.
2. Swap the AVG.
3. Vector timer circuit. If the timers arent running, I think it can trigger a watchdog. (This happened on a AD, so not sure if it applied to color vectors).
If no watchdog, but video:
1. Pull the vector ram, then boot to test. If you get the beeps, then program code is running, and busses are communicating.
2. If the rest of the board seems like theres no activity, ground the ~SA, test point. This should generate traffic in the logic circuits. Usually this is sufficient to look for dead or stuck logic. However, there are lots of normally high or low logic points,so you have to be careful. Look at the schematic.. Anywhere you see 8P54 as a signature, that means logic high.. 0000 means logic low.
3. If your getting crazy or no output, Look at the vector data shifters. They are the link between the data bus and the DACs. They also feed back to the AVG, but in SA mode, the AVG ignores the feedback.
Other:
DC offset on the output:
You should be able to center the signal around a 0v origin (using a scope). If the signal goes negative immediately, its a bad op amp.
BW Crashes when a game starts.
As soon as you get the web, the game goes into a black hole.. This was a bad pokey. I think the pokey generates the random numbers that decide where the enemies are, and if it cant, it gets confused. The board wont watch dog in this situation. FYI my board always reports bad pokeys in test mode, even with brand new ones.
Wierd vertical or horizontal smearing
(like this) http://www.flickr.com/photos/mcandrewsnd/8101429777/in/photostream
Check the digital switch E10
Here is a Tempest nugget:
Game would reset, or shut down intermittantly after a while, ~Reset held low or goes low. This ended up being a bad big blue. The board uses the 10.3 to trigger the POR. The regulated +5 looked fine, but the 10.3 had enough ripple on it to prevent the power supply stabilization circuit from taking the POR signal high.
