Centipede Audio Intermittent Start & Low -30v

jasonsmith01

Member
Joined
Feb 13, 2014
Messages
506
Reaction score
5
Location
British Columbia, Canada
When the audio is on it stays on but, sometimes when I turn the game on there won't be any audio, and sometimes after it sits on for a while the audio won't start on coin up. A power cycle always makes it come back on. If I start the game with all the others from the main breaker the audio almost never works so I've been starting it on it's own after all the other boot up. This to me screams a voltage issue but they are good, although I haven't continuously monitored them through game play and boot up, just spot checks.

Does this issue sound familiar to anyone? A wiggle to the connectors doesn't help either. My gut says the audio circuit is failing to start maybe due to voltage. Could it be a failing pokey?
 
Last edited:
Also, go into test mode and see if any audio chip failures are indicated (aka Pokey chips). See the manual for details.

My money would be on a bad Pokey (or audio amp on the game board), but checking the AR is a good idea as well.

Also, check all of the DC voltages on the game board (+5, -5, and +12V), using the appropriate test points, and the GND on the game board. The +12 and -5 are used for audio.
 
Had time to do the test routine and it checks out. Paid particular attention to the audio tests obviously and they seem to be doing what they should be doing. No beeps on test start either.

I'll check all the voltages again and maybe watch them for a bit, then I'll do a AR swap. I do have a cap kit for the main board which I haven't done... worth doing?

Thanks fellas
 
Had time to do the test routine and it checks out. Paid particular attention to the audio tests obviously and they seem to be doing what they should be doing. No beeps on test start either.

I'll check all the voltages again and maybe watch them for a bit, then I'll do a AR swap. I do have a cap kit for the main board which I haven't done... worth doing?


No, capping the game board (or AR) is not worth doing, unless the caps are physically ripped off, stepped on, or otherwise damaged. Atari used very good caps, and they don't go bad.

Note that it's also possible to have a bad Pokey and not have it flag a Pokey error in the test screen. This is particularly true with audio issues. But if the aural audio tests were working, that is a good sign.

The only other audio chip on the game board is the LM324 at K10, and they do go bad. You may want to try swapping that out, if the AR and voltages check out ok.
 
So voltages...

+5v = +5.004
+12 = +11.91
-5v = -5.082

The above from what I can tell feed the Pokey and the LM324.

The 22's and 30 are a bit wonkey:

+22v = +24.51
-22v = -24.91
+15v = +14.88
-15v = -14.97
-30v = -25.51

ARII

+12v = +11.99
+22v = +24.59
-5v = -5.000
-22v = -24.83

I think I may replace the pokey and play for a bit and see what happens or hopefully it'll happen again soon and I can run the test mode. I'm currently researching the 22's and the 30's to see what they're out.
 
The 22's are fine. They normally run closer to 24, so no issues there.

The -30 is used by the score saving EAROM. -25 is a little low, but probably not enough to cause issues. But the worst that would happen would be the scores wouldn't be saving right.

I'd have to check the schematics to see which type of circuit it has (as there are two), but there's either an LM324 or 555 in the -30v circuit. Change that and it'll get you back to -30. But change the Pokey first, as sometimes a bad one can drag the voltage down too.
 
Thanks for that, I did actually have the scores reset once on me but only once. I think there's a triple 5 in the 30 circuit, I got a bunch of those fun little chips. I'll swap the pokey and check the voltage, if it's still low I'll swap out the triple 5.
 
Well this was interesting and smelly. I changed out the Pokey, need to wait for a bit to see if I lose audio again.

The -30v however didn't change and was still at -25. I swapped out the LM555 for a NE555 and the -30v test point dropped to -16v and pin 8 (Vcc) on the 555 slowly went up in smoke with one beep (H2 Ram Fail) on start up. Stuck the old LM555 back in and back to normal with -25v and no smoke and no beeps.

Looking at the data sheets the 555's should be replaceable?? I did use a socket so can easily try another NE555 but figured I'd check in first.
 
Depending on where you got it, you may have gotten a counterfeit chip. (Also, dumb question, but you're sure you didn't put it in backwards, right?) You also want to test your work, and make sure there's no shorts under the socket. But if it works with the old chip, it more likely was just a bad chip.

I do think there are different ratings for 555's, but they differ only by max frequency, and any of them should be fine for this application.
 
I got a bunch a while ago and have used a few for various electronic fun projects but could be a bad one in the batch I guess. Good question, I did make sure it was in the right the way and I'm totally ocd about checking my work.

I did look like someone changed the 555 before I did though, maybe it's a weak point on my board.

The LM555 data sheet states it's a direct replacement for the NE555 so who knows. I'll try another and see what happens.
 
Same thing happened with a known good NE555. -30v went down to -14v and it got hot. Turned it off before the smoke this time. Checked the configuration again as well, pin 1 both the same place.

This is strange.
 

Attachments

  • 555.PNG
    555.PNG
    33.1 KB · Views: 7
Last edited:
Keep in mind that just because a chip worked on another board, it's not guaranteed to work on the Centipede board, unless the other board was the same circuit. It's possible it could be failing at whatever speed/configuration it's running in on the Centipede board, and not on whatever other board you tested it on. (I'm not saying that's definitely what's happening in this case, but it is something to keep in mind, as it is in the realm of possibilities).

Also, take a DMM and check all of the socket pins for shorts, to all other pins, and ground.
 
Okay so checked all the voltages and labeled the schematic image. More importantly I checked all the connections and determined the following:

4, 5, 7 go nowhere as they should.
8 connects to +15v.
3 connects to C86.

Here's the strange part...

6 goes to GND as it should.
1 and 2 are connected together via trace and should also go to GND as per the schematic but do not. There aren't any traces on the board to take 1 and 2 to GND or connect with 6. Is this done internally in the 555? It must be unless the centipede board has more than two layers of traces.
 

Attachments

  • 555.PNG
    555.PNG
    28.7 KB · Views: 8
Last edited:
FWIW, simulation shows some waveforms etc. :D
BTW, most Atari boards are only two layers. Can't find a trace? Likely hidden under the IC. Do trig/thrshld pins ohm-out to GND?

-30V_555.jpg
 
Last edited:
Okay now that ^^ is cool! Makes me wanna dig out the scope and see it for realzies. Curious what the numbers would look like if the GND and Trigger were connected but didn't go to ground?

Off to try and sim this as well, great idea.
 
For some reason the TH (6) goes to ground, just not the trigger (2) or pin 1 GND.

Just building a sim, having an issue with the reset I think. What is in-between that and ground?
 
Back
Top Bottom