Mortal Kombat 2 (v3.1 4Mbit) - Most ROMs test bad

[blahjedi]

So I've got this MK2 here that does not want to play nicely. Xodarap kindly looked at it several years back with another one and had both working, but the one he posted about was a rev2.1 & this is a 3.1. Since then I've pulled it out of storage and Schrodinger's board came into effect (ie. It's working/alive until you turn it on!).

This is the 4Mbit version with the daughterboard, and with that plugged in a handful of the ROMs on it will test good during the startup - but the remainder all fail. With the daughterboard off, it'll just show red on all the ROMs. RAM, UG12 & UJ12 all test ok.

Things I know:

• It'll occasionally boot into the service mode if the DIP is enabled. Some of the text is corrupt, but it'll allow me to go through and run tests etc
• If the Service mode is disabled, it'll do the test, throw up the "CMOS DEAD" screen and either go into a reboot - or very rarely it'll boot the game to a very corrupt character select screen
• If the sound board is connected, the "bong" noise pops off happily
• I've tested a handful of the ROMs using a comparison tool to match the dump vs. the ones supplied by the owner of the Hobbyroms website. All I've tried compare ok.
• Tested voltages around the board, all are consistent with what I'm measuring at the edge connector
• Both PLCC chips have been pulled, cleaned and reseated. No signs of lifted legs on UE13 either. Have also reflowed the pins on the PLCC sockets & on the daughterboard connectors (both ends). No change.
• If I remove the PAL at UE8, the game boots to a black screen. If I remove the PAL at UF17, the game boots and just displays the ROMs as failing (which I believe is normal behaviour?)
  
  
• The jumpers and DIP are correctly set for 4Mbit
• My workbench setup works happily with my NBA Jam board, so likely rules out power (Although I have tried MK2 in my KI2 cab and it also fails there)

What I don't know:

• If the PAL at UF17 is ok. I'd assume because some of the daughterboard ROMs test ok, then it means it's probably ok as well. But without another MK2 or means to burn a new PAL/GAL, I'm stuck on this one
• If swapping to the 8Mbit version will accomplish anything - especially if there's some other underlying issue preventing the FPGA/Custom from seeing the game ROMs. Gut tells me this is a bad idea purely because attempting the swap when there's already an issue introduces something new to the mix that may complicate life more!

While there's some unhappy looking surface mask on the traces, every dodgy one I've tested comes up fine. likewise a couple of the resistors around have signs of rust on the legs, but I'm unsure if they're in spots that may cause these issues (Look on the PCB at the resistor bank just above P12 at the bottom, and R136 at the top)

Any suggestions on where to look next? Thanks!

 

[Asure]

It'll occasionally boot into the service mode if the DIP is enabled. Some of the text is corrupt, but it'll allow me to go through and run tests etc

That sounds to me like a problem with adressing rom space. From my recompiling this game, the gfx used in diag mode (the font, line bars etc.) are all inside the program rom. None of this diagfont is in GFX space area.

This seems to point to the PAL. You can use the 8M version in a PAL chip with some glitches, probably the 4M version too.
But at least it will show if the problem is there. Try to swap the PAL with one from NBA JAM or other T-Unit to see if it at least boots with borked game gfx yet?

 

[blahjedi]

That sounds to me like a problem with adressing rom space. From my recompiling this game, the gfx used in diag mode (the font, line bars etc.) are all inside the program rom. None of this diagfont is in GFX space area.

This seems to point to the PAL. You can use the 8M version in a PAL chip with some glitches, probably the 4M version too.
But at least it will show if the problem is there. Try to swap the PAL with one from NBA JAM or other T-Unit to see if it at least boots with borked game gfx yet?

Ahh thank you - I didn't realise it was possible to use Jam's PAL in MK2. I'll give it a try and check back in.

Thanks!

 

[blahjedi]

Sadly no dice - I tried swapping both UE8 and UF17 off my NBA Jam (individually) and still get the same outcome (If I put both in everything fails!).

Debating the value in removing both sockets for the PALs and checking under them & replacing with new ones. Otherwise there's a voice in my head telling me I'm about to venture off to trace each address line.

 

[Asure]

I'll try to explain what happens here since the roms are testing bad. What you are seeing is 'correct' for bad rom data.
As the game boots it runs the diag code from program roms (which are green.) and shows the test screen with font data in the program roms.
(So we know that access to that part of ROM is ok.)

After booting it will the unzip a picture it uses for the background of Tsung and Goro. The code that unzips is not robust and it will crash on bad reads.
Then the watchdog will notice it crashed.. (it's running from program rom in a loop) and reset the board.

You say it occasionaly boots to a corrupt select screen. You can probably get it to boot to the select screen by setting hard dips to CMOS COINAGE - FREE PLAY.
Then just keep smacking 1p/2p start as the game boots. You'll 99% sure skip the attract mode image unzip before it can crash.

We also know the reason it crashes, it cannot access all GFX roms properly. I don't know why exactly, but here's some things to try.

- Can you enter DIAG mode at al with the test switch? (if booted to corrupt select screen?)
- Does it show a BAD CMOS alwyas, even on a fresh battery?
- Does the sound work fine on the select screen?
- Can you post a pic of the corrupt select screen?
- Is the DIAG mode corrupt too?
- Can you reset CMOS from the corrupt DIAG/Test mode?
- Are the dip switches for 4M/8M set correctly? (UJ1 dip 8 should be set to OFF for sandwich MK2)
- Can the DIAG/TEST mode display the dips you set correctly? (Most importantly the UJ1 #

Bootup sequence is a but nasty, it tests ROM first but there are also tests for UB21/EU13. These are the chips that handle the GFROM decoding along with the PAL.
From a ROM hacking/code perspective, i cannot tell which one of these might be broken. I'm hoping to deduce some more with the answers.

 

[mecha]

usually suggests that the Custom Chip is bad or best case scenario merely popped some solder to the legs. you definitely don't want to bend or flex those boards in any way.

someone on KLOV had the same problem and their fault wound up being broken parts on the memory expansion board. (which I'm assuming you have based on your POST screen)

try removing the memory expansion and see if any of the CPU board roms come up green then.