MV4FT2 crosshatch of death

[fgm]

Hello, I have a MV4FT2 board that I'd like to repair. The first thing I did was remove the battery; fortunately it didn't leak and there's no visible damage. The board is in pretty good condition, specially the bottom board without track rot.

Using the Neo Diagnostics the board showed a Video Ram error. I replaced the VRAM chips and after that all I got was a white screen that turned cyan after a second. I was really pissed off thinking I broke it, but yesterday I tried again. Poking ​some SMD custom chips the board showed signs of life again finally I found the culprit: the LSPC2-A2 needed a reflow.

Now the board passes all tests, but it's stuck at the crosshatch of death. With a cartridge in slot 1 I can use the memory viewer and looking at address 0x100 I can see "0x4E00 0x4F00 0x4700 0x4F00" instead of "0x4E45 0x4F2D 0x4747 0x4F00"

Putting the cartridge in other slots I can't access the memory viewer, getting a black screen.The top board has been cleaned very well so it shouldn't be a dirty slot issue. I have tried to press the CN10 connector (using the protective foam) with no success.

- What data or address lines can be failing?
- Is the MV4F CN10 pinout the same as MV4FT2 CN10?
- Is the SLOT4FT top board the same as SLOT4F?

Any idea on how to continue debugging is welcomed.

Things I have done since writing this post (work in progress):

- Replaced the two 470µF capacitors with new original Panasonic NHG.
- Check connectivity between d0...d15 Neo-G0 at B1 with CN10 according to Neo-G0 pinout and MV4F pinout
- Reflowed B1 and B2
- Checked continuity between cartridge 1 and B1. Noticed a typo in the MV4F pinout link: pin 39 connects to CN9-13A (the document indicates CN9-13)

I suspect of C11 and D11 (74xx245) since they are the responsible for Cartridge 1.

Just desoldered C11 and it tested bad! replaced it and the board now works with all slots, but it still has some issues with some games and sound, even after replacing all electrolytic capacitors.

Metal Slug 2 misses some melody sounds, and Pulstar doesn't play the melody. Both carts work fine in other machines, and the slots and cartridges are very clean.

One of the two channels sounds a bit less loud in the hardware tests, but both sound. Things I have tried so far:

- Clear the backup ram
- Replaced all electrolytics caps
- Tested 74LS245 at M2, N2, M3, N3
- Replaced Yamaha YM2610
- Replaced Yamaha YM3016
- Replaced Z80
- Replaced Z80 memory (6116)
- Replaced 2066D
- Replaced 2003 (BA12003)
- Extracted and changed order of the three 4558, no change
- Neo Diagnostics all tests passed, even using the M1 ROM keeping +D pressed at boot
- 2M and 2N test ok. In the MV4F I think they are data latches for the YM2610.
- Tried a SM1 rom from a parts board. Same sound problem.
- Tested with stock bios and Unibios 3.3

The board came with a blob of solder and a cutted trace to make it jamma compliant. I've removed the blob and restored the connection to make it stereo as it should be originally.

I'm running out of ideas. Is it possible this is caused by some broken trace or IC at the top board? games CRC checks are fine according to Unibios tests. Maybe the problem is the Neo-D0 custom chip?