Asteroids Cocktail: Resets and flicker

[Tore-Norway]

Hi guys, new member here, located in Oslo, Norway. Electronics engineer.

I am restoring an Asteroids Cocktail. It's firing up normally in test mode. However, when I set the test switch to normal game mode, it runs normally for some seconds and then resets repeatedly, not periodically. It's looking like screen flicker. If I set test mode again, all is normal.
I have changed all electrolytic capacitors and adjusted the main PCB voltage to 5V. The supply voltage is not changing when the game runs/resets.

Any tips on why the game behaves like this?

Thanks and cheers from Oslo,
Tore

 

[Nes4life]

Hey, there's only 1 section I really know of that doesn't get exercised in test mode, and that's the Stack Counter at K5 (LS191) and Stack Register at F4, H4 & J4 (all LS670) in the Vector Generator Program Counter. Test the counter first! Hope that helps.

Pulling out pin 1 on L6 will disable the vector section (no output to screen) but keep the logic section working. It's a good thing to do until you are 100% sure the logic section is working properly in and out of test mode. A faulty vector section often causes the constant resetting (watchdogging). You should be able to start a game and play it blind if the logic section is 100%.

 

[Tore-Norway]

Nes4life, thank you for the tips.
I discovered that the behaviour changed when I touched the ROM IC's.
I have previously thoroughly cleaned corroded ROM and CPU pins, but decided to also change the IC sockets using high-quality sockets. Then the game first started up nicely, but becomes flaky after a while, resetting intermittenly. This looks like a connection problem to the ROMs or faulty ROM chips. It is possible to alter the behaviour if I touch the ROM pins on the ROM IC's, so I assume the problem lies somewhere around the C1, D/E1 and F1 chips.
I have a version -04 PCB with 035143/44/45-02 ROM chips. Do these fail, or should I look for other problems?
Cheers,
Tore

 

[Nes4life]

Yes, Asteroids EPROMs are prone to failure; they're 40 years old after all!
As long as you're happy the new sockets you put in are great and the soldering you did made good connections, then I'd totally try replacing the EPROMs.
2716 EPROMs aren't too hard to burn and are still readily available at a low price (I'm pretty sure you can also use 2732 if you have those but then need to double the image to fill the space). If you've got all the equipment and parts then it's a really quick thing to check.
If you intend to use a multi-game with your Asteroids PCB then these tend not to communicate with the original EPROMs anyway.

 

[Tore-Norway]

Thanks!
I'll double check the sockets.
I have some 2716 EPROMS in the drawer and the good old Dataman programmer at hand, as well as a TL866. Where do I get hold of the content files? As mentioned, I have version 04 of the PCB. I could then also verify the content of the original memory chips against the new ones.
Are all four socketed IC's EPROMS?
Cheers,
Tore

 

[Tore-Norway]

I am still working on this:
-Disabling L6 pin 1 does not affect the operation, so this looks like a CPU/program section fault.
-My new EPROM IC sockets has been tested OK. No problems in both solder and component layer connections.
-I have verified the EPROM contents, all four are identical to ROM images I found in an online MAME ROM file set.
-Disabling the Watchdog circuit by grounding test point does not affect the operation.
-CPU reset pin 40 is stable unless I press the reset button.

Test mode works, and the /WDCLR signal then works fine, so I guess this means that there is some problem in game program execution.
It could be that my TL866 and Dataman S4 reads the EPROM chips OK, and they do not work in normal program execution. This is not very likely, but I have ordered some 2716 IC's in order to eliminate this.
Game seem to be working OK for 5-15 seconds when the board starts from cold. If it is warm, it flickers occasionally and may lock completely up, and need button reset to start again, or the Watchdog wakes it up. I have not been able to spot any likely cause by using freeze spray.

Any tips would be appreciated.
Cheers,
Tore

 

[Nes4life]

You'll know more once you swap out the EPROMs. I'm concerned that pressing on them alters the behaviour. Once swapped out, if pressing them still changes things then you have bad connections somewhere.

As for freezing after 5-15 seconds; once you've swapped out the EPROMs, and this is still a problem, try swapping out the CPU.

Can you attach a picture of your PCB? L6 should be a LS42. Pin 1 is the DMAGO signal, and without that the Vec section should not start up, hence it should play blind.

 

[Tore-Norway]

Thank you.
I was probably fooled into thinking that touching the EPROMs had an effect. The game resets intermittently also without touching anything. L6 is indeed a 74LS42, and the problem persists without the DMAGO signal, (video disabled) so this looks like a CPU/memory/bus problem.
The Self Test runs fine and without reporting RAM/ROM problems. Looking at the disassembly of the program, the Self Test routine is placed on top of the program memory, so I do not think this is a B2/C2 bus buffer problem causing addressing problems. I will order a 6502A CPU to swap this too.

Posting messages here takes some time, my messages are subject for moderator approval before being visible, and I am therefore reluctant to update postings. Maybe the approval procedure is a newcomer necessity of this forum.
All the best,
Tore

 

[Tore-Norway]

It works!
I did a bit of thinking. Since the game was problematic in game mode only, and the video part was working OK, I suspected a fault in the address decoding circuits and was probably right. Since 74xx chips are fairly cheap, I chose to use brute force and change a number of IC's instead of doing more fault finding to determine the actual bad chip(s).
I ordered new IC's for the complete address decoding circuits. While at it, I ordered several other of the uP circuit IC's to avoid having to wait for another delivery if the first parts were not solving the issue. Since it now works fine, I will not bother to try to identify the faulty part(s).
I changed the following: B6:7414, L3/E4:139, D6:02, C6/N5:00, E5:20, L6:42, N11:174, P5:86, M5:32, D4:74, D5:393, B2/C2:244.
The original ROM chips works fine.
I got a tempered glass made at a local vendor. New push buttons, panel overlays and glass underlay from www.arcadeshop.com
New paint on legs and screws, new furniture legs (M8x20) and square plastic end caps from eBay.
Thanks for all help!
Cheers,
Tore