MV1F weird error message

[Apocalypse]

I'm trying to fix a MV1F for fluxcore.
The battery has leaked and eaten some traces.
fluxcore has already found 3 broken traces and fixed them but that didn't help.
The screen is a complete garbage but doesn't flicker. The color change each time you power the board.
I've found 2 other broken traces and I can see a little progress: the garbage screen is still here but now I've got this message:



The board has a diagnostic BIOS.
Any idea?

 

[fluxcore]

from the diag bios rom...

FIT IH SETTXR MEIASNH RE.E......HTNES SYET MSIS UTKCI NAWCTDHGO..ETTSNI GIBSOM RIORIRGN

if this text remains here..... then system is stuck in watchdog. (testing bios mirroring)

https://wiki.neogeodev.org/index.php?title=Watchdog

NEO-B1 problems? Bit weird, it doesn't seem like the system is resetting constantly, which I would think you'd expect from watchdog issues.

Good progress though, definitely couldn't read that text before!

Do you have a logic probe? Might need to bring mine around.

 

[Apocalypse]

from the diag bios rom...

FIT IH SETTXR MEIASNH RE.E......HTNES SYET MSIS UTKCI NAWCTDHGO..ETTSNI GIBSOM RIORIRGN

if this text remains here..... then system is stuck in watchdog. (testing bios mirroring)

Yeah I did the same byteswap:
IF THIS TEXT REMAINS HERE...
THEN SYSTEM IS STUCK IN WATCHDOG

What is strange is I have only the first line on screen...

https://wiki.neogeodev.org/index.php?title=Watchdog

It can be physically disabled by putting pin Sig.pngDOGE to ground (J2 jumper).

I've tried that too but it had no effect...

NEO-B1 problems? Bit weird, it doesn't seem like the system is resetting constantly, which I would think you'd expect from watchdog issues.

It definitely doesn't reset constantly and I can't hear the clicking sound of death...

Do you have a logic probe? Might need to bring mine around.

Unfortunately not...

 

[Apocalypse]

OK, I've managed to get further:

The full error message:


TTACHTDOG DELAY...


I've even managed to get in the menu but due to graphics being garbage it's not easy to know what does what. What I'm sure is watchdog is fine and controls work:

CALENDAR TEST


WRAM/BRAM test displays "PASSED"

MISC. INPUT TEST

 

[HeavyMachineGoob]

Do you not have a standard BIOS to compare with? The screen shouldn't be stuck on garbage like that. Diagnosis would be a little bit more helpful if we could see what happens on a stock BIOS too.

At times like this, I quote channelmaniac:

Model: MVS MV-4F
Symptom: Stuck in Watchdog (Click of Death)

Have a 4 slot board in to repair that was giving the click of death - stuck in watchdog - problem.

This can be caused by many different problems, including:

Bad CPU
Bad Work SRAM
Bad Backup SRAM
Missing control signals on Work or Backup SRAM
Shorted 74AS245 ICs buffering program ROMs on top board to CPU on bottom
Short on address bus
Short on data bus
Missing address or data signals to chips

First thing... Check for gouged traces. Found and fixed one. No effect.
Next... Substitute BIOS. Nothing.
Next... Check control signals on the Work RAM - Stuck high on pins 22 and 27.
Next... Check control signals on the Backup RAM - Stuck high on pins 22 and 27.
Next... Check control signals on the BIOS ROM - These were working.

So we had a problem somewhere in the system to where it was trying to read the BIOS but not able to initialize hardware.

Pulled the 74AS245 ICs buffering the data lines to the top board. Had no effect.

Checked for more broken traces. No more were found.

Checked the address and data lines going to the BIOS and SRAM ICs. All were good. Checked the data lines going to the NEO B1 chip. All were good. Checked the data lines. All were good.

Next checked the A22* and A23* alternate lines. Missing one one! Ran a jumper wire from Pin 55 of the NEO-E0 IC to pin 117 of the NEO-B1 IC and the board came up!

OOPS. Z80 Error. Reset line stuck high. On a Z80 since the reset line did not transition from low to high the Z80 never booted. Ran a jumper and the audio section worked. Replaced the pulled 74AS245 ICs and the board wouldnt' boot. Reset was stuck low.

Turns out the reset line for the Z80 CPU and chips on the left side of the board goes through the 74AS245 IC at E11. Removed the jumper for the reset line and board booted normally.

Plugged the top board back in, inserted some test carts and played a couple of games.

 

[fluxcore]

Good stuff Apoc, was that by repairing more damaged traces?

I certainly have a stock bios to try it with, kind of just assumed the diag bios would be more useful in every way at this stage...

 

[HeavyMachineGoob]

In this case, where the garbage tiles persist, it'd be useful to see if that same behavior occurs on a stock BIOS as well. It looks like the MVS isn't clearing the VRAM, or the Diagnostic BIOS did try to clear it and one or both of the VRAM chips are bad, or not working correctly.

 

[Apocalypse]

Do you not have a standard BIOS to compare with? The screen shouldn't be stuck on garbage like that. Diagnosis would be a little bit more helpful if we could see what happens on a stock BIOS too.

At times like this, I quote channelmaniac:

Yes I do have an OG BIOS. I can swap it from my working MV1FZ.
I'm sure it has nothing to do with the watchdog and the message stays there because the screen isn't refreshed properly (VRAM problem?).

Good stuff Apoc, was that by repairing more damaged traces?

I certainly have a stock bios to try it with, kind of just assumed the diag bios would be more useful in every way at this stage...

Just reflowed few pins. To me the board is working fine now (can run the BIOS and make tests, controls work, etc.) except there's something wrong with the video (VRAM?).

In this case, where the garbage tiles persist, it'd be useful to see if that same behavior occurs on a stock BIOS as well. It looks like the MVS isn't clearing the VRAM, or the Diagnostic BIOS did try to clear it and one or both of the VRAM chips are bad, or not working correctly.

It should be the same with any BIOS. I've got the same feeling: something wrong with the VRAM.

 

[fluxcore]

Sounds reasonable to me, let me know what the VRAM chips are and I'll try to order some replacements.

Thanks HMG

 

[mikejmoffitt]

Maybe the /WE signal is no good on one RAM chip, but is fine on the other. The fact that the text is showing up on the FIX layer without issues, but the rest remains garbled makes me wonder if the upper 8 bits are stuck while the lower ones aren't, somehow.

For the curious, the reason the screen seems to fade downwards on the garbage pattern is that the overscan border color is not black, so the clamping circuit in the monitor is clamping to the value (in this case looks like a red or pink color) and subsequently subtracting it from the image once the black level is set.

 

[Xian Xi]

This looks like a missing link between LSPC2 and B1.

Also, do you hear any tones while doing the full test?

 

[Apocalypse]

Sounds reasonable to me, let me know what the VRAM chips are and I'll try to order some replacements.

Thanks HMG

I should have fresh ones from when I repaired my MV1FZ which had a backup RAM error.
I think they are identical.

Maybe the /WE signal is no good on one RAM chip, but is fine on the other. The fact that the text is showing up on the FIX layer without issues, but the rest remains garbled makes me wonder if the upper 8 bits are stuck while the lower ones aren't, somehow.

For the curious, the reason the screen seems to fade downwards on the garbage pattern is that the overscan border color is not black, so the clamping circuit in the monitor is clamping to the value (in this case looks like a red or pink color) and subsequently subtracting it from the image once the black level is set.

I've checked ALL the traces between the 4 VRAMs and the LSPC2 and didn't find any broken one. But that doesn't mean the /WE signal is good (faulty LSPC2?).

This looks like a missing link between LSPC2 and B1.

I think it's something around the LSPC2 too (either VRAMs, bad traces, etc.). I'll check that tonight.

Also, do you hear any tones while doing the full test?

No but as always on MV1F boards when the battery has leaked, the audio part is badly damaged.

 

[Xian Xi]

No but as always on MV1F boards when the battery has leaked, the audio part is badly damaged.

Fix the audio area first. The diagnostic bios can give you errors via audio tones.

 

[fluxcore]

Hopefully it's not the LSPC2. This board is really not worth replacing that for.

 

[Apocalypse]

Fix the audio area first. The diagnostic bios can give you errors via audio tones.

I know but I was being lazy

Hopefully it's not the LSPC2. This board is really not worth replacing that for.

With an OG BIOS I get the same garbage screen but without the text.
I've replaced the 2 chips of the lower VRAM with faster ones : no change...
I've checked ALL the traces going to/from the LSPC2 and they all are fine.
Could it be something wrong with the upper VRAM chips?
Replacing the LSPC2 is a real pain...

 

[fluxcore]

Probably just consider it a parts board at this point. I think you've done more than enough Apoc

 

[Apocalypse]

Probably just consider it a parts board at this point. I think you've done more than enough Apoc

It's a bit disappointing though.

 

[fluxcore]

It's a bit disappointing though.

Yeah, I know what you mean. I can only hope it provided many hours of awesome entertainment during its day _b

 

[Apocalypse]

Yeah, I know what you mean. I can only hope it provided many hours of awesome entertainment during its day _b

RIP MV1F... Your name will stay in our hearts forever...

 

[channelmaniac]

There is a custom chip in the battery backup area with 8 data bits that connects to the SRAM for the audio section. Those like to get eaten.

Next there are traces from the SRAM to the Z80 CPU and the ROM in the sound section. Next there are traces from the SRAM to the custom chipset chip right next to it and a trace from that custom to the 2610. You need to check all of those.

It would be very odd to see any of the traces in that area affecting the board... EXCEPT for the backup RAM.

I'm thinking your watchdog mode is caused by damaged traces in the battery backup section that control the access to the backup RAM. Pins 20, 22, and 27 are the ones to look at with a cautious eye. If they are stuck in the wrong state then the backup RAM will put spurious data on the address bus and crash the CPU.

Oh, and HMG, I'm glad you find my posts valuable enough to quote!

 

[Apocalypse]

There is a custom chip in the battery backup area with 8 data bits that connects to the SRAM for the audio section. Those like to get eaten.

Next there are traces from the SRAM to the Z80 CPU and the ROM in the sound section. Next there are traces from the SRAM to the custom chipset chip right next to it and a trace from that custom to the 2610. You need to check all of those.

It would be very odd to see any of the traces in that area affecting the board... EXCEPT for the backup RAM.

I'm thinking your watchdog mode is caused by damaged traces in the battery backup section that control the access to the backup RAM. Pins 20, 22, and 27 are the ones to look at with a cautious eye. If they are stuck in the wrong state then the backup RAM will put spurious data on the address bus and crash the CPU.

Oh, and HMG, I'm glad you find my posts valuable enough to quote!

I've checked all connections between SM1/Z80/6116 are they are all good now (fixed).
I'll check the traces around the backup RAMs when I'll have time.

 

[HeavyMachineGoob]

Oh, and HMG, I'm glad you find my posts valuable enough to quote!

Me? I am honored to have the privilege of quoting your repair logs, resident master of the Neo Geo, Mr. Neo Fix-it. It is repair techs like you who breathe new life into our favorite arcade games, a skill we should all praise and be thankful for.

Seriously though, your repair logs are awesome, thanks so much.

 

[fluxcore]

Yeah, I'm thankful to anyone that posts logs and info about their repairs, every bit of data is helpful when trying to restore these poor boards. Some of us gave up training in electronics fairly early but like to tinker around, so hints on where to focus attention is incredibly valuable.

Doesn't really sound like the CPU is crashing though if menus are navigable etc?

 

[HeavyMachineGoob]

That's the main reason why I think the board should also be tested with a stock BIOS. Unless we have smkdan here to explain, it's hard to say how lenient the Diagnostic BIOS is with severe system errors like this. If the Stock BIOS gets the click of death or gets random RAM errors, then that is very telling, for the very reasons channelmaniac mentioned. .

 

[Apocalypse]

That's the main reason why I think the board should also be tested with a stock BIOS. Unless we have smkdan here to explain, it's hard to say how lenient the Diagnostic BIOS is with severe system errors like this. If the Stock BIOS gets the click of death or gets random RAM errors, then that is very telling, for the very reasons channelmaniac mentioned. .

With an OG BIOS I get the same garbage screen but without the text.

Also, no click of death.