Backup RAM Region error (MV-1 with Unibios)

segasonicfan

Camel Slug
Joined
Feb 3, 2006
Posts
517
Hey everyone,
Got an MV-1 here with an error code on the Unibios (image attached)

Address 00D014CA

Read:
5555

Write:
5515

any idea what chip could be the culprit / good place to start? I don't know the Neo hardware super well.
Thanks!
 

Attachments

  • backup ram region error.jpg
    backup ram region error.jpg
    164.7 KB · Views: 52

HeavyMachineGoob

My poontang misses Lenn Yang's wang
10 Year Member
Joined
Apr 3, 2011
Posts
5,850
IIRC it's the lower backup ram that has gone bad, going by the specific address and the second byte being bad. Fastest way to identify which ram is which is to check which half of the 68k each ram is connected to. Lower backup ram will be in the A0 - A7 range I think.
 

segasonicfan

Camel Slug
Joined
Feb 3, 2006
Posts
517
thanks for the reply. I beeped out A0-A14 and they all check out OK connecting to the 68k. searched around amd found which 2 chips handle backup ram (the 2 sony surface mount ones directly next to the battery). also found schematics(blurry) and a page on backup ram addresses and such:

https://wiki.neogeodev.org/index.php?title=Schematics

https://wiki.neogeodev.org/index.php?title=Backup_RAM

It boots into AES mode cross-hatch now (after removing overclock mod, that helped for some reason) although when I hit "D" for memory viewer it goes black.

so im guessing next step is to just order and replace these chips? is there anything else worth checking before i do that? cause its about $20 and a months wait to get them.

IIRC it's the lower backup ram that has gone bad, going by the specific address and the second byte being bad. Fastest way to identify which ram is which is to check which half of the 68k each ram is connected to. Lower backup ram will be in the A0 - A7 range I think.
 
Last edited:

HeavyMachineGoob

My poontang misses Lenn Yang's wang
10 Year Member
Joined
Apr 3, 2011
Posts
5,850
Thing is, if the error is static and doesn't change, then for sure one of the backup ram chips is bad. It's because the address is very specific and not at a chip size boundary. The read "5515" represents one byte from each ram chip. One is good (55), the other isn't (15). It's been a while since I had to fix an mvs board, but I think it's the lower backup ram that is bad here.

I might have some 32KB surface mount SRAM chips left, I could ship those to you, I'm only in Canada.
 
Last edited:

segasonicfan

Camel Slug
Joined
Feb 3, 2006
Posts
517
So after some poking around I got memtest again and some kinda weird stuff. I basically have been flipping the hard DIP settings and I removed the backup RAM battery. There is no corrosion at all.

1) Half the time it boots into a cross-hatch and freezes if I press D for memview

2) If I leave it off a while and default DIPs it will show the original error and memtest. I once got it to even boot into the logo screen for AES mode

Here's a bunch of pics of the address info from memviewer and such:
 

Attachments

  • 20170113_180836.jpg
    20170113_180836.jpg
    133.1 KB · Views: 39
  • 20170113_181110.jpg
    20170113_181110.jpg
    173.4 KB · Views: 39
  • 20170113_181207.jpg
    20170113_181207.jpg
    174.4 KB · Views: 37
  • 20170113_181249.jpg
    20170113_181249.jpg
    115.4 KB · Views: 38
  • 20170113_181311.jpg
    20170113_181311.jpg
    181.8 KB · Views: 38
  • 20170113_181332.jpg
    20170113_181332.jpg
    175.9 KB · Views: 38
  • backup ram region error.jpg
    backup ram region error.jpg
    164.7 KB · Views: 39

segasonicfan

Camel Slug
Joined
Feb 3, 2006
Posts
517
oh, just saw your reply!

Thing is, if the error is static and doesn't change, then for sure one of the backup ram chips is bad. It's because the address is very specific and not at a chip size boundary. The read "5515" represents one byte from each ram chip. One is good (55), the other isn't (15). It's been a while since I had to fix an mvs board, but I think it's the lower backup ram that is bad here.

That makes a lot of sense (I'm still new to understanding this). Also, this was in storage for years, so not much else would explain the sudden failure (working before being in storage). I was reading that the Backup RAM fails cause the battery is powering it 24/7, so after 25+ years I can see that being pretty hard on a chip.

I might have some 32KB surface mount SRAM chips left, I could ship those to you, I'm only in Canada.
that would be great!!! Are yours CXK58256MM-12? Here's a pic of mine.

-Segasonicfan
 

Attachments

  • 20170113_180751.jpg
    20170113_180751.jpg
    186.7 KB · Views: 38

segasonicfan

Camel Slug
Joined
Feb 3, 2006
Posts
517
Replaced RAM from Neo MV-1B and now I get
Address: 00D00010
WR: 5555 Read 5551

sounds like one of the chips is still bad, right? any idea which one? Im so zonked from reflowing this thing....Ive swaped 6 RAM chips so far x_X
 

HeavyMachineGoob

My poontang misses Lenn Yang's wang
10 Year Member
Joined
Apr 3, 2011
Posts
5,850
Is the RAM error different every time you power on the MVS? If so, check the state of the enable lines, it's possible one of them has a corroded trace. They are pins 20, 22 and 27 last I checked.

It's also possible either the RAM chips you've used are bad or are incompatible. In my time of working on NES/FC games, they can be picky with RAM chips for no apparent reason.
 

segasonicfan

Camel Slug
Joined
Feb 3, 2006
Posts
517
nope error is always the same....i just wish i knew which one was the lower ram... it boots and works im aes mode so at least i can still play stuff
 

HeavyMachineGoob

My poontang misses Lenn Yang's wang
10 Year Member
Joined
Apr 3, 2011
Posts
5,850
The fastest way to know is to do a continuity test between the RAM chips and the 68k CPU. The lower RAM uses data lines D0 through to D7, upper RAM uses D8 - D15. You can just check D0 on both RAM chips and see where they go to on the CPU.

If the MV-1B you got really was voltage damaged, then it's possible the RAM chips all died. It's not ideal to use chips from parts boards, because they will have accumulated some wear already. You should get a fresh new RAM chip (or at least lightly used and tested) for your MVS.
 

segasonicfan

Camel Slug
Joined
Feb 3, 2006
Posts
517
The fastest way to know is to do a continuity test between the RAM chips and the 68k CPU. The lower RAM uses data lines D0 through to D7, upper RAM uses D8 - D15. You can just check D0 on both RAM chips and see where they go to on the CPU.

If the MV-1B you got really was voltage damaged, then it's possible the RAM chips all died. It's not ideal to use chips from parts boards, because they will have accumulated some wear already. You should get a fresh new RAM chip (or at least lightly used and tested) for your MVS.

THANK YOU!!!
I did not realize that was all I had to do to check. The lower RAM uses D0-D7 as mentioned, pins 11 thru 19 which are labelled "I/O" on the datasheet for the RAM. Pin 14 is GND of course.

Turns out there was a bad trace from the 68k to pin 12. My guess is this happened during all the reflowing I was doing with the bad SRAM chips. The ones from the broken MV-1B work OK it looks like :D (that system is super toast, Lol)

soooo happy to have my original MV-1 back from the dead after 8 years in storage - just in time for my birthday :D :D :D (my friend is gonna bring over some Last Blade to play - woot!)

-Segasonicfan
 
Last edited:

segasonicfan

Camel Slug
Joined
Feb 3, 2006
Posts
517
oh and for those curious, the "bottom RAM" is the RAM closest to the metal shielding (or plastic) of the MV-1.
 
Top