- Joined
- Mar 8, 2002
- Posts
- 3,686
Got three faulty 1-slots (models MV1A and MV1FZ) from an arcade op, all with back up ram errors. When you get a READ 55xx (where x=any number), then the problem is related to the "lower" ram, that is the ram whose data lines (D0-D7) are connected to D0-D7 of the CPU. When you get a READ xx55, the problem is the "higher" ram, with data lines connected to D8-D15 of the CPU. Check Nightmare Tony's many and very useful posts on the subject.
The MV1FZ, which I dealt with for first, gave this error:
http://img.photobucket.com/albums/v65/nassivera/backupramerror1.jpg
At first I thought it was a bad lower ram that needed to be replaced. However after having replaced it I still got the same error which meant the ram wasn't bad at all. A more accurate continuity test revealed that the connection between pin 27 (write enable) of the ram and pin 8 of the 74HC32 (location U17) was broken somewhere (not visible). Connecting these pins with a wire fixed the problem completely:
http://img.photobucket.com/albums/v65/nassivera/neoram8.jpg
So when I set out to fix the other two boards (that were giving similar error messages) the first thing I did was check the same connection and much to my surprise both boards had the same problem:
http://img.photobucket.com/albums/v65/nassivera/backupramerror2.jpg
http://img.photobucket.com/albums/v65/nassivera/backupramfixed2.jpg
MV1A: ADDRESS 00D0000 WRITE 5555 READ 5540
http://img.photobucket.com/albums/v65/nassivera/backupramfixed1.jpg
In these cases too it was practically impossible to see the exact point where the trace was broken, it certainly wasn't due to an accidental scratch either on parts or solder side. I assume it was corrosion in one of the points where the traces change side. It may be noted that the printed circuit on these 1995 boards has the same quality of high serial home systems, a quality definitely inferior in my view to that of the early MVS boards (MV1, 6-slots, older 4-slots) and low serial AES. I think the failures had partly to do with this cheaper PCB quality.
The MV1FZ, which I dealt with for first, gave this error:
http://img.photobucket.com/albums/v65/nassivera/backupramerror1.jpg
At first I thought it was a bad lower ram that needed to be replaced. However after having replaced it I still got the same error which meant the ram wasn't bad at all. A more accurate continuity test revealed that the connection between pin 27 (write enable) of the ram and pin 8 of the 74HC32 (location U17) was broken somewhere (not visible). Connecting these pins with a wire fixed the problem completely:
http://img.photobucket.com/albums/v65/nassivera/neoram8.jpg
So when I set out to fix the other two boards (that were giving similar error messages) the first thing I did was check the same connection and much to my surprise both boards had the same problem:
http://img.photobucket.com/albums/v65/nassivera/backupramerror2.jpg
http://img.photobucket.com/albums/v65/nassivera/backupramfixed2.jpg
MV1A: ADDRESS 00D0000 WRITE 5555 READ 5540
http://img.photobucket.com/albums/v65/nassivera/backupramfixed1.jpg
In these cases too it was practically impossible to see the exact point where the trace was broken, it certainly wasn't due to an accidental scratch either on parts or solder side. I assume it was corrosion in one of the points where the traces change side. It may be noted that the printed circuit on these 1995 boards has the same quality of high serial home systems, a quality definitely inferior in my view to that of the early MVS boards (MV1, 6-slots, older 4-slots) and low serial AES. I think the failures had partly to do with this cheaper PCB quality.