Hello,
I know this is too late, but have you solved this issue? What was your solution? Well, in case you did not solve it yet, let me give you 5 cents of contribution.
In my eyes, the CMC50 on this Metal Slug MVS is working perfectly!
Here are the facts:
1- IF the M1 ROM had any damaged bit from data lines or address lines, the CMC would be unable to decrypt anything, not even the part of the Graphics or the HUD Layer of the screen, so the M1 is perfect.
2- The CMC doesn't care what is passed on the bus to the NEOGEO after the decryption takes place, if there is any noise in the address lines or data lines the CMC will keep working unless the M1 fails, or the power fails, and whatever the CMC receives from the C roms it will pass on. You can see some of the decryption on screen (LOGO or Score), so this is another proof the CMC is working, else you would not recognize anything. So I assume some of the Cs are causing the problem. The decryption is diffuse and is split over all C ROMs. For example, the S1 is exactly located at the end of C7 / C8 after decryption, but because the encryption is diffused the data from this S1 is spread over all C Roms to be more precise.
3- I would suggest reflowing all the C ROMS (from C1 to C8)
4- Also Verify that every resistor in front of the CMC chip is measuring approximately 561 OHMs (from R3 to R34) ignore R2 because is for another purpose and has a different value of 161 OHMS.
5- Could also be the case that some pin of the CMC that makes communication with the C Roms gets damaged by something falling over it or the cartridge falling on the ground or even bad welding that decayed over time and needs to be reflowed on the CMC itself, but this would require having technical skills.
Let me know...