Asteroids Deluxe RAM errors

Grbgemen

Active member
Joined
Mar 12, 2008
Messages
1,167
Reaction score
8
Location
Lincoln, Nebraska
Got a AD pcb that is giving to high beeps and two low beeps. Manual says R3 & M3 are bad. Looks like R3 was replaced or botched desoldering job at some point. Managed to get it out and install a socket. Few jumpers were needed as a few pads were already burnt off. Socketed M3 also.

I have a bunch of 2114 RAM chips from spare Ms. Pac boards and other misc pcbs. I have tried them all and still stuck with the two high and two low beeps. A few times i got three highs and one low which indicates that just R3 is bad but after swapping it out, it will go back to two high and two low.

I cant imagine that all of these 2114 rams that I have are bad but I guess it is possible. Is there something else that can cause false RAM failures?

I have triple checked my work and seems okay.
 
If you're confident that the RAM is good, and you have good continuity from the legs to all of the address and data bus lines...

Maybe I'd look at the chip select logic. Starting from the RAMs, ICs that are in the decoding/select circitry include: M4, L4 & L2.
 
Hi - sorry I'm not at my pc and typing all this on my mobile. If memory serves me, the RAM being reported as bad is vector RAM. These are in a bank of four. Close to these are the octal RAM buffers, depending on your PCB these will either be DP8304BN or 74LS245. There is one for the MPU close to the 6502 CPU and one for the Vector circuit near the Vector RAM. Check that IC for faults, if this is bad CPU cannot access the RAM.

I will post more when I am back home.
 
Bumping an old thread to add some info, for reference.

I was having a similar problem as OP. Was getting 4 high and 1 low beep, but the RAM was definitely good. Turns out M4 was the culprit.

Thanks to DarrenF for the assistance.
 
Bumping an old thread, to add more info like Andrew did! I had an Asteroids pcb just now, giving me two high tones then two low tones signifying M4 and R4 bad. Swapping the ram didn't fix the issue. Asteroids and Asteroids deluxe are very similar and have nearly identical circuits involving these ram chips.

Those are the first two vector ram chips, those 4 use a separate buffer at R2 on this board (not sure on each version, there's at least 4 or 5 different designs) . Logic Probe showed nice fat pulses on all the data lines going in but completely anemic signals going out of the buffer, I believed they probably should be similar going in as going out. Swapped the 74LS245 and now the board plays with no ram errors.

So it wasn't specifically saying R4 and M4 were bad ram, it would have said all 4 of those ram were bad if it had the ability to talk to the other two after R4 and M4.
 
Daaaaaamn, 2015. I was but a baby then, repair wise.

Yeah, the buffers are common issues on these, when you get a bad RAM that isn't the RAM. I've seen many of them since.

The other thing it can be is bad decoder logic. I've seen that since too.

And lastly, shorted bus traces. You'll often see little wisps of solder flake off of the traces on the parts side (where there is no solder mask) and short bus traces together, causing similar symptoms. You can also see it from damage under the ROMs, if people used screwdrivers to remove chips, and unknowingly damage traces. So it's always priority to give them a good visual inspection under a bright light first.

(Just FYI, for anyone else searching and finding this in the future.)

Great job on the bump. Good memories.
 
Those "wisps" of solder might just be tin whiskers. We're seeing them more and more in aging electronics.
 
Those "wisps" of solder might just be tin whiskers. We're seeing them more and more in aging electronics.


Yeah, tin whiskers are real, but these aren't that. (For anyone reading who isn't familiar, google them, they're pretty interesting.) I have seen one or two cases over the years that looked like they could have been tin whiskers, but it was hard to tell if it wasn't just debris. But what I'm referring to here is something different, and more common on Atari boards.

These are just an artifact of the fact that Atari didn't use solder mask on the parts sides of these boards. And if something gets dragged across the bus traces (just from the boards being tossed around over the years), the thin layer of tin that is on top of the copper can flake off pretty easily, and can reach over and short to an adjacent trace, in the tighter spots where the trace spacing is one trace width or less. The tin just isn't bonded to the copper really well in places, coupled with how close the bus traces are, particularly where they run under the ROM and RAM chips, between the chip legs. It's definitely handling-related in this case, as you can see the tin flaked off.

I really don't know why Atari didn't use top solder mask (maybe just to save a few bucks?), but it's responsible for a lot of issues on boards I see. They used to stump the hell out of me, but now that I know to look/test for them, I waste a lot less time.

The trick I use is to hold the board up against a bright light, and look *through* the PCB (i.e., with the light behind the PCB). That makes them a lot easier to spot.
 
Back
Top Bottom