MV1F weird error message

[HeavyMachineGoob]

Ok so on a stock BIOS, the board hangs at the garbage screen and there's no click of death. I'd have to research that again but I think that means either the 68k is dead or the watchdog circuit is dead, or both. I'll post again if I find any leads on what can cause that problem.

For right now, get a 68k pinout and check /HALT and /RESET. I.... Can't remember what to check on them. Wish I could be more of a help at this time, click of death isn't something I research much.

 

[Apocalypse]

Ok so on a stock BIOS, the board hangs at the garbage screen and there's no click of death. I'd have to research that again but I think that means either the 68k is dead or the watchdog circuit is dead, or both. I'll post again if I find any leads on what can cause that problem.

For right now, get a 68k pinout and check /HALT and /RESET. I.... Can't remember what to check on them. Wish I could be more of a help at this time, click of death isn't something I research much.

If I can enter the menu with the diag BIOS then the 68k is fine to me. It's an uncommon problem with symptoms close to "stuck on watchdog" but that's not it. It's video related (VRAM?).

 

[HeavyMachineGoob]

I think the diagnostic bios is actually masking the real problem here. Let me explain here.

When you first startup the Neo Geo, you see a big screen of garbage tiles, the video sync may or may not be in the right range for your display currently. Those garbage tiles appear because unpowered SRAM chips always power-on in an unknown state, as in full of random data. During the Power On Self Test (POST), the VRAM is cleared, the video sync is stabilized and various other parts of the system are tested for functionality.

On the stock BIOS, it doesn't POST, nor does the watchdog circuit attempt to reset the 68k. That is a significant problem.

The diagnostic BIOS, being designed just for diagnosis, probably doesn't care about the integrity of the system, aside from the click of death which can still happen.

Going by that, well the quote from channelmaniac I mentioned doesn't seem to be as relevant since there's no click of death. Stuck at POST can be a dead 68k, probably not since yes it does still work with the diagnostic BIOS. The thing is, that VRAM isn't clearing like it is supposed to, neither on the diagnostic BIOS or stock BIOS. In fact, POST doesn't even happen at all. That's bad.

Now, I'm not that familiar with failed POST situations on MVS boards, but if you've gotten Video RAM error messages in the past, then you could certainly try replacing them and/or check the enable lines for proper functionality. A dead address or data line shouldn't cause POST to fail, but the enable lines certainly could.

 

[mikejmoffitt]

We may need to further look into the POST and the order of initialization of the 68k. It is possible the 68k is in the process of running a test and is stuck in a loop, but the watchdog is being pet properly in this loop. If a 68k higher address line is stuck, it could cause the program to be jumping to an incorrect location which contains incorrect (but valid) code, causing such a situation. There are a myriad of possibilities there. I am willing to bet the 68k itself is not the problem.

If you have a logic analyzer, a trace can be generated from the address lines and R/W signal.

 

[Apocalypse]

You really think that board can be saved?
I've spent enough time on it and don't want to be disappointed by finding the LSPC2 is faulty after having fixed everything else on the board...

 

[channelmaniac]

This is why I said to concentrate on the Backup RAM circuitry. If it's not right it will output spurious data on the bus and the CPU will never start properly.

SMH...