magicianlord77
New member
Hi, just wanted to post a repair log of my House of the Dead board (Model 2C):

I need it for my HOTD cabinet restoration, the board did almost entirely work, it's in great condition, no signs of previous repair or disassembly or rust, always stayed inside the cabinet, it just had an amount of 3d graphics issues – polygons flying around and misplaced textures, also it would get stuck when displaying the high scores table in the demo loop, and display the wrong amount of credits.
Video:
As a first troubleshooting step, took all the roms out of the romboard, dumped them with an XGecu T48, and ran mame romident, it displayed that all are fine, so the problem shouldn't be rom related (would be very unfortunate to have an issue in the romboard itself), but I don't have another one so let's hope not.
Then ran the memory test from the service menu, it tests a lot of stuff but not most of the 3d related rams. Grouped the ICs to later locate them in the board.

It did catch a BAD SRAM, IC79, it is a BR6265BF-10SL (8K x 8, low power), looks related to backup memory since it's close to the supercap. Found that I could get an exact same model one from a MV-1A neogeo board thanks to this page: https://wiki.neogeodev.org/index.php?title=RAM_chip. Fortunately I have a broken MV-1A board, so performed the replacement, and that fixed the issue, now the memory test displays all OK, and the game does not get stuck anymore.

Now about the graphics issues, I wanted to know which ICs from the board were being checked in the memory test, to focus into the ones that weren't, so I highlighted them in a photo of the board:
Yellow groups are tested ones. Blue are untested.

From the untested, clearly group 7 has to be the sound SRAM so it doesn't matter, also group 8 is far from the others, I think it's related to the 3d framebuffer (can't remember where I saw that) so I'll leave it for later.
Remaining groups are 1-6. From those, I wanted to know which ones could be related to the issues the board is having, so I could try replace them.
I've seen a technique could be to carefully probe their data/address pins by grounding them (grounding VCC or a signal wired to VCC, etc. could bring greater problems!), and to check what happens on screen. Learned about that from here: https://www.ukvac.com/forum/threads/sega-model-2a-glitching-poly-repair.63592/ (great repair). Also of course applies to other kind of game boards.
It's quite difficult to probe them since the romboard sits on top, so I had to poke them from the rear of the PCB sandwich or solder a thin wire from the desired pins and ground them in a breadboard. Also other way (seen in the above post) is to find and poke the tiny vias related to the wanted signals from below the board, haven't tried it.
I found group 1-4 affects polygons, poking group 1 alters polygon positions quite a lot. Group 5 seems related to texture colors, so it is not related to the issue. I can't recall what did group 6 did, I think it was polygon related too.
So, began by replacing group 1 (CY7C199-25VC) (32K x 8), because I had compatible (faster) new ICs to put ready at hand (IS61C256AL-12JLI). (photo before soldering them in)

And nothing improved, quite a disappointment
Some time later after thinking about it, and by seeing the results of this other post https://www.ukvac.com/forum/threads/sega-model-2a-fixed-random-missing-polygons.55302/, I just bit the bullet and replaced group 2, which is 8 TC518128CFWL-80 PSRAMs (128K x 8), they're quite odd, had to take them from a good Model 2B board, whose graphics section is quite similar to Model 2C, they are SEC (samsung) KM658128ALG-7.

After the tedious job, something improved for the first time! There aren't untextured polygons anymore, and the glitching looks a bit less chaotic.
Video:
At this point I was also considering replacing the 315-5798 custom, because I didn't believe that many different rams could be bad, fortunately I yet don't have a 40x40 qfp nozzle, so had to continue exhausting the bad ram possibilities.
So I continued with group 3 (CY7C185-25VC) (8K x 8), because they were more comfortable to solder than the others remaining, in this case I've also replaced them with the ones from the Model 2B board.

Tested again, and graphics improved even more! Now besides looking much better there aren't remains of previous polygons in the screen anymore.
Video:
Now I had only group 2 (CY7C188-25VC) (32K x 9) remaining from that part of the board, also replaced them with the ones from the Model 2B board, I think it's an odd IC too. After some work soldering them, clearing solder bridges,..

Installed the romboard back on, started it up, and graphics are entirely fixed!
, fortunately that was really the last group with issues.
Video:
Later I would like to test the replaced ICs, probably a couple should be bad and the rest fine, but it'll be difficult since they're different sized SOP/SOJ.
Also I should locate replacements for the bad ones, for rebuilding the good Model 2B, it looks quite thin now, donated too many ICs.


I need it for my HOTD cabinet restoration, the board did almost entirely work, it's in great condition, no signs of previous repair or disassembly or rust, always stayed inside the cabinet, it just had an amount of 3d graphics issues – polygons flying around and misplaced textures, also it would get stuck when displaying the high scores table in the demo loop, and display the wrong amount of credits.
Video:

As a first troubleshooting step, took all the roms out of the romboard, dumped them with an XGecu T48, and ran mame romident, it displayed that all are fine, so the problem shouldn't be rom related (would be very unfortunate to have an issue in the romboard itself), but I don't have another one so let's hope not.
Then ran the memory test from the service menu, it tests a lot of stuff but not most of the 3d related rams. Grouped the ICs to later locate them in the board.

It did catch a BAD SRAM, IC79, it is a BR6265BF-10SL (8K x 8, low power), looks related to backup memory since it's close to the supercap. Found that I could get an exact same model one from a MV-1A neogeo board thanks to this page: https://wiki.neogeodev.org/index.php?title=RAM_chip. Fortunately I have a broken MV-1A board, so performed the replacement, and that fixed the issue, now the memory test displays all OK, and the game does not get stuck anymore.

Now about the graphics issues, I wanted to know which ICs from the board were being checked in the memory test, to focus into the ones that weren't, so I highlighted them in a photo of the board:
Yellow groups are tested ones. Blue are untested.

From the untested, clearly group 7 has to be the sound SRAM so it doesn't matter, also group 8 is far from the others, I think it's related to the 3d framebuffer (can't remember where I saw that) so I'll leave it for later.
Remaining groups are 1-6. From those, I wanted to know which ones could be related to the issues the board is having, so I could try replace them.
I've seen a technique could be to carefully probe their data/address pins by grounding them (grounding VCC or a signal wired to VCC, etc. could bring greater problems!), and to check what happens on screen. Learned about that from here: https://www.ukvac.com/forum/threads/sega-model-2a-glitching-poly-repair.63592/ (great repair). Also of course applies to other kind of game boards.
It's quite difficult to probe them since the romboard sits on top, so I had to poke them from the rear of the PCB sandwich or solder a thin wire from the desired pins and ground them in a breadboard. Also other way (seen in the above post) is to find and poke the tiny vias related to the wanted signals from below the board, haven't tried it.
I found group 1-4 affects polygons, poking group 1 alters polygon positions quite a lot. Group 5 seems related to texture colors, so it is not related to the issue. I can't recall what did group 6 did, I think it was polygon related too.
So, began by replacing group 1 (CY7C199-25VC) (32K x 8), because I had compatible (faster) new ICs to put ready at hand (IS61C256AL-12JLI). (photo before soldering them in)

And nothing improved, quite a disappointment
Some time later after thinking about it, and by seeing the results of this other post https://www.ukvac.com/forum/threads/sega-model-2a-fixed-random-missing-polygons.55302/, I just bit the bullet and replaced group 2, which is 8 TC518128CFWL-80 PSRAMs (128K x 8), they're quite odd, had to take them from a good Model 2B board, whose graphics section is quite similar to Model 2C, they are SEC (samsung) KM658128ALG-7.

After the tedious job, something improved for the first time! There aren't untextured polygons anymore, and the glitching looks a bit less chaotic.
Video:

At this point I was also considering replacing the 315-5798 custom, because I didn't believe that many different rams could be bad, fortunately I yet don't have a 40x40 qfp nozzle, so had to continue exhausting the bad ram possibilities.
So I continued with group 3 (CY7C185-25VC) (8K x 8), because they were more comfortable to solder than the others remaining, in this case I've also replaced them with the ones from the Model 2B board.

Tested again, and graphics improved even more! Now besides looking much better there aren't remains of previous polygons in the screen anymore.
Video:

Now I had only group 2 (CY7C188-25VC) (32K x 9) remaining from that part of the board, also replaced them with the ones from the Model 2B board, I think it's an odd IC too. After some work soldering them, clearing solder bridges,..

Installed the romboard back on, started it up, and graphics are entirely fixed!
Video:

Later I would like to test the replaced ICs, probably a couple should be bad and the rest fine, but it'll be difficult since they're different sized SOP/SOJ.
Also I should locate replacements for the bad ones, for rebuilding the good Model 2B, it looks quite thin now, donated too many ICs.

Last edited: