What's new

[SOLVED] M92 RAM (Irem Skins Game ) Errors

Pr3tty F1y

Student
Joined
Jul 17, 2024
Messages
28
Reaction score
14
Location
USA
I recently picked up a ROUGH Irem Skins Game (from the roms, it's Set 2 in MAME) M92 board. This is the US version of Major Title II. Eventually I'm going to do a conversion on it, but I want to get it up to speed first.

Some small, fuzzy mammal definitely made a home in between the two board sandwich at some point in the past. There was some very neatly packed dirt in between the entirety of one of the 32pin DIP sockets to protect from the pins, some leaves, and plenty of small bits of dried plant material, spider webs, a few unlucky bugs - including a spider - that died and dried out under some of the ROM sockets, etc. The dirtiness (well, maybe not the packed dirt itself :-D) was clear from the pictures from the seller, but something still drew me to it anyway. I guess I'm a sucker for junk that may be gold.

That all being said, after I gave the board several rinses of isopropyl alcohol (which may or may not be the best choice of solvent, but my thought was that it would, at the very least, dry fast as I was eager to test the board :-D) and scrubbed it down with a soft tooth brush, it cleaned up pretty well.

And, amazingly, it booted without issue on the first try. The graphics were a bit yellowy, but it played fine. I was amazed. There was no sound, but the potentiometer was working as I could hear the background buzz get louder when I turned it up so I figured the amp was OK.

So I removed all of the electrolytic capacitors which all seem to be near the audio portion of the board. C236 doesn't seem to be part of the audio, but replacing it seemed to boost the color balance back to where it should be. And, after recapping everything, still no audio.

I then removed and dumped all of the roms on my T56. All of the roms tested fine and matched Irem Skins Game (Set 2) in MAME except for MT2-SH0- which Xgpro gave a pin error on all but like 2 pins. The legs were straight, I tried cleaning them, but nothing. So I burned a new 27C512 and replaced it and - BOOM - audio is working great!

Which now brings me to the RAM errors. Sometimes when I boot the game, it boots fine and it says RAM OK, other times, I get an ever changing series of RAM errors.

My questions are:
  • Due to the condition of the board, is it possible that multiple RAM chips failed?
    • I know that anything is possible but....
    • Although the board was rough, it still seems wild that I'd have multiple different types of SRAM chips fail when the board overall works fairly well
  • Is there something else that could be causing the RAM issues?
    • Like maybe the interconnects between the two boards not making a good connection?
    • Or a failure of one type of RAM cascading and causing failure of other types of RAM
    • Maybe a LS logic chip fizzling out, but I'm not sure which ones would affect all of the various RAM types
  • Could it be something else?
    • I'm willing to entertain anything as I really want to get this thing working
Here are some examples of the RAM errors that seem to change on each boot. Sometimes I get RAM OK and it just boots fine (although the graphical corruption below still is exhibited when the RAM tests as good):
ram_errors.jpg


less_ram_errors.jpg


The game looks great overall, even with the RAM errors:
looking_good.jpg


But there is clearly some corruption with the layering and either the row across the green being corrupt/transparent or the column down the leader board being corrupt and/or missing:
corruption.jpg


corruption2.jpg


Here's what the board ended up looking like after the initial cleaning (pre-recapping). It's dirty/rough, but the components are not rusted or corroded (the only rust seems to be on the support post screws). It may be on the verge of starting some light corrosion, but the solder on the custom chips is silver and shiny like it was brand new. I'm really perplexed about why some of it looks good and other places more rough. I'm not sure of how it was stored. All I know is that it was hilariously dirty, the green lacquer of the board wipes off green a bit in spots near the JAMMA edge connector with any moist wiping (water, isopropyl, etc.), and the bottom board seemed a quite bit rougher than the top board. The boards came assembled and screwed together tight. I had to break the rust on the support post screws, so I don't think the board interconnect pins were exposed. This thing seems to have been left bolted together for a very long time:
top.jpg


bottom.jpg
 
Last edited:
Check your voltages for 5.0V. If the board isn’t receiving sufficient power from the power supply then it will definitely cause random RAM errors. Those boards tend to draw more power than your average board too in my experience.
 
Thanks for that information. I did get cued in to a potential voltage issue as my test bench setupp in my mini-electronics shop in my garage is powered by a PicoATX power supply with a 19V laptop adapter. It works for the majority of games (although no -5V), but it wasn't enough to get the M92 board to boot.

I have an AxunWorks supergun box and I drove the voltage as high as I was comfortable with at 5.25 volts to be within the recommendations in the Major Title II jamma manual and I still get various RAM errors:

5_22_volts.jpg


ram_issues_2.jpg


ram_issues_3.jpg


5_25_volts.jpg


ram_issues_1.jpg
 
I just swapped out the PALs on the B board with GAL16V8B's programmed with JED's from the PLD archive. Still the same... :-(

I have various RAM chips on order. I'm guessing I"m probably going to socket them/replace them and still have this issue X/ But with the relatively clear graphic corruption, despite getting the error messages or not, I'm figuring at least something is bad RAM wise.

I also pulled the CPU and two socketed customs (one on the A board and one on the B board) and reseated them.
 
5.22V is way too high. Dial it down to 5.0V and retest. I never exceed 5.15V on any board if it all possible.
Agreed. I was just testing to ensure that it wasn't a power supply issue due to my picoatx powered setup clearly being insufficient. The manual allows for a tolerance of 5V+/-5% and 12V+/-10% (although I know that doesn't speak to amperage). I'm pretty sure I'm not suffering from voltage drop causing the RAM issues.

Check the CPU socket
I've already pulled the V33 and re-seated it. I'll check it again.


I've also found this similar-ish repair log. It does look like one type of RAM error will cascade into another (potentially): https://jammarcade.net/irem-m92-motherboard-repair-log/

Being that I'm still waiting on parts and the fact that my T56 can test 6116 SRAM modules to see if they're good, I removed the BR6116 chips to see if my issue was the same as the above issue. I thought I found my problem when I saw that one of the legs of one of the chips was not protruding from the solder side of the board. Once I got the chip removed, that leg was folded flat under the chip! Unbelievable. Not sure if that speaks to early 90's Irem quality control or not, but I found it funny. The board was in an untouched shape, so it had to come from the factory that way. However, after testing both 6116's and mine tested as good (I'm still going to swap them out with replacements just to check) and socketing, everything was the same re: the variable RAM errors at boot.

So I moved on to removing the adjacent MB81C78A's. The first one came out fine. But the second one... I'm not all that experienced, but I've removed a good many chips... and this chip just sheered in half. Almost like it was de-laminating and peeling apart. There was no pop, no snap... it just peeled into two sections as I (in my opion fairly gently) lifted up on it. I could see the fine wires from the die to the pins just cleanly hanging there. I was stunned ;-D So maybe that was the problem (probably not, I'm not that lucky), but now I'm down to just one MB81C78A and (of course) that leaves me with RAM errors.

I'll post follow-up info next week on my progress after getting some replacement modules in.

Thanks for all the ideas/guidance thus far.
 
Last edited:
I'm no closer to a solution and I've socketed all the ram and a handful of 74LS273N's as they seem to have failed in multiple other repair logs from various folks. Still no dice.

Got my crap capture setup out and Youtube compressed the video to hell, but it demonstrates the consistency of the errors I'm seeing. It also demonstrated the board auto-rebooting after failing the ram test the first time.

It seems like the board has problem on a cold boot. It generates the craziest errors/graphical corruption during the POST routine during first boot. After that, it seems to work better. I've been racking my brain about what chips on the board may need to "charge up" to start working correctly, but I got nothing.

While the RAM errors change each time, the in game results are the same:
  1. Leaderboard loses a column when moving left and that column has artifacts after a hole (1:50 and again at about 3 minutes in the video)
  2. The putting green is messed up on the rows that have tiles for the ball and the hole (about 2:30 in the video).
Not sure where else to go at this point. I'm really hoping that the custom chip(s) aren't bad, but I'm starting to lose faith that it is something else:


After letting the board run for like 45 minutes, I used my handheld temperature gun to scan the board. The hottest I measured was 105F on the PLD near the board interface connectors on the B board. The chips directly above that area in the back corner with the two interboard connectors on the A board were slightly less warm (high 90's F), but higher than most everything else. All of the other chips were lower in temp (somewhere in the mid 80's F). So I can't identify anything that's fizzling out. The cpu/large custom chips/audio chips were some of the coolest on the board.
 
Last edited:
OK. I'm exhausted, I'm sweaty (mostly from reflowing the SMD chips and making all sorts of ahhhhhh noises as my brain couldn't process the directionality of my soldering iron drag vs. the orientation of my very limited microscope), but the board is fixed.

Will post shots tomorrow. But no more RAM or graphics errors

It was a broken trace on the A board just above one of the interboard connectors.
 
Thank you for everyone who provided ideas/guidance with my issue!

Here's some photos of the fix. Hopefully it can save someone else some headaches. The scuff with the broken trace looked originally like dirt on the PCB (as it was VERY dirty), but I started scrubbing things down with a very soft toothbrush and some isopropyl alcohol as I was thinking maybe something was being bridged by something that was conductive. After cleaning/straightening the pins on the top of the connector, I noticed how bad that scuff was once the dirt/rust was cleared and a few quick tests with the multimeter made it pretty apparent that there was a connectivity issue.

Trace issue/fix:

trace_fix_3.jpg


trace_fix_2.jpg


trace_fix_1.jpg


Fixed graphics:

fixed_green.jpg


fixed_leader_board.jpg


PCB that had MUCH more work done to it than needed :-D

fixed_pcb.jpg
 
Last edited:
Back
Top