Battlezone
Active member
Not the most exciting fix log but I'll take the time to write it up anyway.
Symptoms
An Atari Warlords 4P cocktail (colour) machine. Some of the paddle sprites drift one way on the screen even though the controller is still. They aren't jittering back and forth between a couple of positions, they are determined to go walk-about by about a pixel at a time for 10-20 pixels. They often return back to about the right place when the controller is moved.
Fixes and red herrings
a) The 4 paddle positions on Warlords are read using analogue to digital converters (ADCs) inside the sound chip (C012294-01 a.k.a Pokey). So I tried swapping this socketed chip for some others I have. This didn't appear to improve the problem. Also I have no easy way to test the ADC logic in the Pokey chips independently. Either all my Pokey chips had bad ADCs or the problem was elsewhere.
b) The only other semi-conductor that might break just some paddles is the LM339 @ A5. So I socketed this and replaced it, but it wasn't the problem.
Pokey is in the lower-middle of this picture and the quad op-amp LM339 is in the lower-right.
c) I noticed that the drift reduced when I held the smaller edge connectors (P19). I'd already cleaned all the edge connectors, but P19 wasn't making robust connections. The video comes off P19 as well and that was also playing up if I touched the connector. So I replaced the old original with a new connector in the loom and that significantly improved the problem on some paddles but not all. It's so awkward re-wiring looms inside the cab.
d) Swapping the control potentiometers around showed that the remaining problems followed the pots. Also I had one working pot that was rotating over it's end-stop. It wasn't cheap but I thought the old machine deserved 4 nice shiney new pots. Replacing them solved the remaining drift.
As a side note: I think Atari did a pretty poor job in the software damping these paddles when you enter your initials. It's easy to accidentally end up entering the letter before or after the one you are targetting. I would have implemented it differently by showing the character you are near when very close to the correct position, and showing the previous character viewed when in the no-mans land between two characters. I might also time how long the character has been stable and freeze it after a while until there is a big movement. Maybe I'll tweak the ROMs at some point.
Symptoms
An Atari Warlords 4P cocktail (colour) machine. Some of the paddle sprites drift one way on the screen even though the controller is still. They aren't jittering back and forth between a couple of positions, they are determined to go walk-about by about a pixel at a time for 10-20 pixels. They often return back to about the right place when the controller is moved.
Fixes and red herrings
a) The 4 paddle positions on Warlords are read using analogue to digital converters (ADCs) inside the sound chip (C012294-01 a.k.a Pokey). So I tried swapping this socketed chip for some others I have. This didn't appear to improve the problem. Also I have no easy way to test the ADC logic in the Pokey chips independently. Either all my Pokey chips had bad ADCs or the problem was elsewhere.
b) The only other semi-conductor that might break just some paddles is the LM339 @ A5. So I socketed this and replaced it, but it wasn't the problem.
Pokey is in the lower-middle of this picture and the quad op-amp LM339 is in the lower-right.
c) I noticed that the drift reduced when I held the smaller edge connectors (P19). I'd already cleaned all the edge connectors, but P19 wasn't making robust connections. The video comes off P19 as well and that was also playing up if I touched the connector. So I replaced the old original with a new connector in the loom and that significantly improved the problem on some paddles but not all. It's so awkward re-wiring looms inside the cab.
d) Swapping the control potentiometers around showed that the remaining problems followed the pots. Also I had one working pot that was rotating over it's end-stop. It wasn't cheap but I thought the old machine deserved 4 nice shiney new pots. Replacing them solved the remaining drift.
As a side note: I think Atari did a pretty poor job in the software damping these paddles when you enter your initials. It's easy to accidentally end up entering the letter before or after the one you are targetting. I would have implemented it differently by showing the character you are near when very close to the correct position, and showing the previous character viewed when in the no-mans land between two characters. I might also time how long the character has been stable and freeze it after a while until there is a big movement. Maybe I'll tweak the ROMs at some point.