MV4 top board issues, only slot 1 works

[neSneSgB]

I have a MV4 that the top board only works on slot 1, the other slots all go to crosshatch. If I try to access the UniBIOS memory viewer on slots 2 and 3, I get a black screen, but on slot 4 I can see memory fine. What I noticed is that past "NEO-GEO" at address 100, the lower byte of each word has D2 stuck high: 00 becomes 04, F9 becomes FD, etc. I can use a working top board on this MV4 and all slots work no problem, so it's the top board for sure. It's possible this issue is also present on slots 2 and 3, but I can't see the memory viewer to confirm. The first picture attached is of MAME, the second is the board.

The M1 test passes fine in all 4 slots and I gave all the slots a good clean with same results. I did try replacing the LS245 at R4 because that's what D2 goes to, but it did not seem to make a difference.

Any ideas?

 

[ic3b4ll]

What version of the MV4 is it? The big one with loads of ICs or the newer with SMD customs?

 

[neSneSgB]

It's the MV4, not the MV4F, so it has no customs on top

 

[ack]

You are mostly there finding the issue. As you have identified in the pics D2 is stuck high on slots 2/3 and probably 4.

The 16 data lines from the program rom for each slot go to the 2x245s to the left of each slots (slot4 is R3&R4, slot3 is L3&L4, etc). They go in pins 2-9 and come out pins 11-18. The indivdual output pins are all tied together across each slots 2x245s (ie you should continuity between each of the pin 11s for the 245s at R3/L3/G3/C3). The outputs all end up on CN10.

The flow of the traces connecting the outputs is slot4's 245s -> slot3's 245s -> slot2's 245s -> slot1's 245s -> CN10

With D2 being stuck high starting at slot2, its pointing to a cut trace between slot1's and slot2's 245s. Which ever one is carrying D2.

 

[neSneSgB]

Thanks for chiming in. I found that pin 13 of the 245 at C4 was not connected to the rest of the 245s, so slot 1's connection was isolated from the rest.

I patched that up, and it seems that all 4 slots are good to go now. I assumed that there might have been an extra issue with 2 and 3 not displaying the memory viewer, but they look good to me.

There's no scratches or scuffs on the board anywhere, so I can only assume the connection just broke somewhere under a chip. The good top board that I used to confirm this bottom board had two invisible breaks on it that I only found through thorough continuity testing.

 

[ic3b4ll]

What @ack said

 

[channelmaniac]

Nice fix!

If you look closely you'll see that there are 4 repeating rows of chips on that top board. They are all connected together on the Neo side of things with the other side if the chips going to the slots. The 74LS138 in the bottom corner is used to select which slot. The 74LS244 chips are always active on their outputs. The 74LS245 latches only output data to the Neo side if that slot is chosen.

Makes for pretty easy troubleshooting. Dead slot, but the other slots work? Start looking between the chips for that slot and the slot connectors. Multiple slots dead? Start looking from the inter-board connectors to each bank of chips for the missing signals.

The video data output is a bit different in that you'll find some 4 or 8-to-1 data selectors depending on if you're looking at a 4 or 6 slot top board. 74LS153/253 on the 4 - if you have lines in graphics look at the inputs and outputs for those chips for a floating or stuck signal and check traces.

Crosshatch = missing signals for the Program ROMs - Address or Data. Crashing in one slot? Look at the upper address lines to see if any are missing.
Garbled sound effects? Look at the data lines for the sound ROMs.
Incorrect sound effects? Look at the address lines for the sound ROMs.
No sound? Z80 appears dead when carts installed? Check the connections to the M1 ROM: address and data.

They aren't too bad to repair. Clean them well and look them over for scratches/gouges or rodent waste corrosion damage. Focus on those areas first. If you find graphics problems and dead traces, one suspect area is under the fishpaper and foam on the bottom side of the board. When the foam gets hard it starts to corrode chip pins and traces on the bottom side of the PCB.