Atari system 1 problems

bonehead

I'll get pix tonight
vacBacker
Feedback
50 (100%)
Credits
1,729CR
Hi, been spending the evening trying to get my cab up and running with a couple of game pcbs. I have Indiana jones, marble madness.
Indiana Jones seems to play ok apart from a few lines through graphics and needs a new main processor socket as it's a bit temperamental. In order to get the game to boot I have to ground the reset line after power or it keeps watch dogging. I am missing the 10.3v from power supply I think but I have a spare I can try tomorrow.

Marble madness cart will run through game test but then says bank switch error. Could this be the slapstick? I have repro from Colind I could try. The mm cpu would try and run once reset line is shorted to ground but said a message like load error or something but now it won't reset and just gives me a garbled up screen image, I guess this means it's not even watch dogging now!!!

Questions is what does bank switch error mean? Google says slap stick, dirty sockets and all sorts.
What could know be wrong with mm cpu? I don't know anything g about repairing pcbs somwill have to send off for repair unless anyone could give me a pointer to try?

Thanks guys,

Steve
 

bonehead

I'll get pix tonight
vacBacker
Feedback
50 (100%)
Credits
1,729CR
Further testing this morning and mm pcb says working ram error on start up and keeps resetting the same when watchdog grounded. I'm getting 15v on audio board but no -15v going to reset line. Hope all this makes sense!!

The cab is pretty jacked up inside but doing my best to work things out.
 

Kingpin

User
Feedback
1 (100%)
Credits
113CR
Bank switch error should indicate faulty ROM(s). Have you verified that the ROMs are OK? There should also be an address displayed which indicates board location if you happen to have the old PCB layout referenced in the manual.

I myself have a Marble Madness LSI PCB which has a "working RAM error" (like your) and garbled graphics and it will not enter self test or attract mode. Previously I've replaced one ROM, swapped all 4 SLAGS from a working RoadBlasters PCB as well as changed the SLAPSTIC to no avail.

Ideas on "working RAM error" are welcome.
 

tb2000

Active member
Feedback
4 (100%)
Credits
3,579CR
I've seen bank switch errors caused by the slapstic before so yeah it could be that. Bank switching is to do with memory usually. As far as I know, it's usually a process used so a CPU can address more ram than it would normally be able to.
 

bonehead

I'll get pix tonight
vacBacker
Feedback
50 (100%)
Credits
1,729CR
An update. I have tried 2 new slap sticks but no improvement. Roma have been verified and still no joy.

Help always appreciated.
 

Kingpin

User
Feedback
1 (100%)
Credits
113CR
In both our cases it could also be bad(E)PROM sockets. Try clean the sockets and legs of the (E)PROMs and heat up the solder joints and if that doesn't work then replacing the sockets might be one alternative and to verify the connections going from point to point (including legs of socketed chips to joints on the solder side of the board) so there aren't any cracks etc. along the paths.

I'll post when any progress has been made.
 

Purity

The Oracle
vacBacker
Feedback
17 (100%)
Credits
2,999CR
Is working ram not work ram? Could you not have an issue with the main CPU ram or the buffers between that and the CPU?

It looks like 6116 ram? Are they socketed? Can you remove them and test them?
Purity2017-03-02 11:16:07
 

Kingpin

User
Feedback
1 (100%)
Credits
113CR
Well, I happen to have another Marble Madness LSI cartridge which I've verified (with success) so I suspect the issue is related to the cartridge. I've also swapped SLAPSTIC from that board as well as some of the (E)PROMS. I've also re-flowed some sockets (error remains) and now I'm using my DMM to verify traces.

It is "WORKING RAM ERROR" which can be seen amongst the garbled graphics from time to time (a bit intermittent actually).
 

GadgetFreak

Active member
vacBacker
Feedback
7 (100%)
Credits
963CR
I am pretty certain the 'Working Ram Error' is related to four 1024x4 static rams on the mainboard.
On the TTL mainboard they are at 1L-4L and on the LSI mainboard they are at 3J, 3K, 4J, 4K.

My LSI ones are in sockets which I put in and am fairly certain it was that error that came up.
 

Purity

The Oracle
vacBacker
Feedback
17 (100%)
Credits
2,999CR
GadgetFreak said:
I am pretty certain the 'Working Ram Error' is related to four 1024x4 static rams on the mainboard.
On the TTL mainboard they are at 1L-4L and on the LSI mainboard they are at 3J, 3K, 4J, 4K.

My LSI ones are in sockets which I put in and am fairly certain it was that error that came up.

I agree and it's what I said 4 posts up
smiley17.gif
. It is certainly where I'd be looking. Just get a logic probe on th ram pins and see if you can see anything odd
 

Kingpin

User
Feedback
1 (100%)
Credits
113CR
But why is my other Marble Madness LSI cartridge working with the same mainboard then? In fact, all my other cartridges is working (all the 5 System I titles) with the same mainboard including the maxed cartridge of RoadBlasters.

Problem is also that I don't have an external harness for testing on the bench which makes fault-finding hard(er).

I forgot to add that I do appreciate your suggestions! :)Kingpin2017-03-04 14:05:42
 

GadgetFreak

Active member
vacBacker
Feedback
7 (100%)
Credits
963CR
Well there is no ram on the cartridge boards and what mainboard ram there is, is only accessed by the 68K cpu.
I would suggest that you have rom errors or problems with the rom sockets though usually these would be detected by the rom test on startup.
 

Kingpin

User
Feedback
1 (100%)
Credits
113CR
OK, sounds like the next step to verify. Again I connected my working Marble Madness LSI cartridge to my mainboard - boots up without issues.

Maybe actually replace all the 28 pin sockets or "steal" them from the unused positions.

I'll also pull and read all the (E)PROMs and verify them against set 5.Kingpin2017-03-04 16:44:02
 
Top