Neo Geo MV1AX Click of Death

[mehguy]

Hello everyone,

After getting my minigun supergun, I had hooked up an atx psu to the minigun's molex connector without realizing that the pins were different (So I effectively ran 12v down the 5v rail and shorted 12v to gnd). That caused the psu to activate the overcurrent protection and just shutdown immediately after I turned it on. It wasn't until closer imspection was when I noticed that the pinouts were different and I rewired it. I now have video but it seems to have a click of death (more closer to 8hz).

I was really hoping that split second of 12v on all the rails didn't kill it but I guess it did. Switching BIOS's doesn't help and it still does this without any bios installed.

I've tried probing around with a logic probe and the m68k's clock is tied low and the Ym2610 gets unusually warm. Is there any recovery from this? I just hope it didn't take my KOF 99 cart with it

 

[maki]

the m68k's clock is tied low

thats no good, is the crystal running?
different MMs will give different values, on mine it shows 2.6V when its running, on others 1.6V

or do you mean 0V?

and the Ym2610 gets unusually warm

50C is for that one is normal, even when is not doing anything

good chance that ICs are fried all over the place, by the sounds of it the 68k might be, but probably others too, the ASICs, RAM, logic ICs etc.

if you do not have a lot of time and lots of spare parts I'd say this is a goner, shouldn't even be used for spares TBH
sorry for not being optimistic, I could be wrong, could just be the 68k but its very unlikely IMO

 

[BIG BEAR]

Yeah.. It reads like you fried that baby.. but your KOF '99 should be fine...I have some spare parts for that though.
I may have a board,same model but would have to double check.
BB

 

[mehguy]

thats no good, is the crystal running?
different MMs will give different values, on mine it shows 2.6V when its running, on others 1.6V

or do you mean 0V?


50C is for that one is normal, even when is not doing anything

good chance that ICs are fried all over the place, by the sounds of it the 68k might be, but probably others too, the ASICs, RAM, logic ICs etc.

Just measured on my MM, it gives 2.5v, but my logic probe says it's low.
Measuring the clocks coming from the Neo-D0 (24M, m68kclck, 12M etc) and they all seem to be tied low. Even measuring Xout seems to be tied low.

if you do not have a lot of time and lots of spare parts I'd say this is a goner, shouldn't even be used for spares TBH
sorry for not being optimistic, I could be wrong, could just be the 68k but its very unlikely IMO

No worries. It's just my negligance and being overly eager to see it running without being cautious. The plastic casing is still good, and isn't marred with "CHINA SALE ONLY" so I could definitely use it if I ever come across one like that.

Yeah.. It reads like you fried that baby.. but your KOF '99 should be fine...I have some spare parts for that though.
I may have a board,same model but would have to double check.
BB

Great thanks. I'd definitely be interested. Let me know what you have.

 

[slugger_dan]

You have video sync, some graphics onscreen and click of death. That means some clocks are working including 24M. You mentioned the clocks are low on the logic probe so something’s wrong there. It should be set to TTL if you see a CMOS/TTL switch. Try 24M again if it was on the wrong setting.

Problem is the PSU overcurrent protection won’t be instant and it would do a better job of saving the PSU than whatever is plugged into it. You could get lucky and find it’s just 1 or 2 parts but I agree with maki, this could be a ton of time and effort to fix.

 

[maki]

Just measured on my MM, it gives 2.5v, but my logic probe says it's low.
Measuring the clocks coming from the Neo-D0 (24M, m68kclck, 12M etc) and they all seem to be tied low. Even measuring Xout seems to be tied low.

A logic probe ain't good for 24Mhz, mine is totally quiet there but sowing low, but these things are usually very cheap, thats why I've mentioned the MM.
A perfect square wave would be as much high as low, and logic probes usually aren't for frequencies this high.

If you can swap the BIOS, it worth a shot IMO.

 

[slugger_dan]

A logic probe ain't good for 24Mhz, mine is totally quiet there but sowing low, but these things are usually very cheap, thats why I've mentioned the MM.
A perfect square wave would be as much high as low, and logic probes usually aren't for frequencies this high.

If you can swap the BIOS, it worth a shot IMO.

That’s a fair point, yeah I don’t know how good the cheaper ones are for anything past 10mhz. I have a 50MHz probe that was good for anything in the Neo but tbh haven’t touched it since buying the scope. It would’ve been 2x or 3x the price of the EISTAR ones that are all over eBay though.

 

[mehguy]

You have video sync, some graphics onscreen and click of death. That means some clocks are working including 24M. You mentioned the clocks are low on the logic probe so something’s wrong there. It should be set to TTL if you see a CMOS/TTL switch. Try 24M again if it was on the wrong setting.

Problem is the PSU overcurrent protection won’t be instant and it would do a better job of saving the PSU than whatever is plugged into it. You could get lucky and find it’s just 1 or 2 parts but I agree with maki, this could be a ton of time and effort to fix.

Yea I just noticed my logic probe won't support more than 20mhz,
My supergun also has a THS7374, what are the odds that it took that with it as well? All my testing was done by just tapping the RGBS signals from the jamma connector and not through the video output of the supergun.

Also I'm assuming non of the passives were killed, right?

A logic probe ain't good for 24Mhz, mine is totally quiet there but sowing low, but these things are usually very cheap, thats why I've mentioned the MM.
A perfect square wave would be as much high as low, and logic probes usually aren't for frequencies this high.

If you can swap the BIOS, it worth a shot IMO.

I've tried a Unibios, SMK Dan's diag bios and no bios at all. No change in behavior.

That’s a fair point, yeah I don’t know how good the cheaper ones are for anything past 10mhz. I have a 50MHz probe that was good for anything in the Neo but tbh haven’t touched it since buying the scope. It would’ve been 2x or 3x the price of the EISTAR ones that are all over eBay though.

A scope would be really useful right now, but unfortunately mine isn't accessible atm

 

[BIG BEAR]

I found an MV1A and it has the vandalized cartridge holder with letter C wriitten on it in various places, I'm guessing for CHINA.I will test it out overnight to see what gives..
BB

 

[slugger_dan]

Yea I just noticed my logic probe won't support more than 20mhz,
My supergun also has a THS7374, what are the odds that it took that with it as well? All my testing was done by just tapping the RGBS signals from the jamma connector and not through the video output of the supergun.

Also I'm assuming non of the passives were killed, right?

Wouldn't worry about the passives as much as the chips. Datasheets usually have an absolute max rating table and the THS7374 part you mentioned is max=5.5v, so that doesn't look good. No way to know until you try. For 5V parts I'd assume absolute max is 5.25v unless the datasheet says different. Doesn't mean 5.25v is safe to run boards at though, it could work fine for a while but it's like burning the candle at both ends.

If you want to try trouble shooting then have a look at a 68k pinout and try probing these. Slow probe shouldn't be an issue for these, as long as the 68k is powered and it looks like the clock is working:
--- RESET and HALT: should be mostly high and every time you hear the click of death, both should be low for a split second.
--- AS: should pulse. Bare minimum to expect from the 68k if it isn't dead. If it's not dead but quiet then a custom part connected to the 68k could be dead as well.
--- A1-A3: should pulse. Low address pins should be busy if it's doing anything useful.

Could also check the OE pin on BIOS to see if it's pulsing. This could be hard to spot if the 68k crashes immediately. It definitely shouldn't be stuck low though.

 

[mehguy]

Wouldn't worry about the passives as much as the chips. Datasheets usually have an absolute max rating table and the THS7374 part you mentioned is max=5.5v, so that doesn't look good. No way to know until you try. For 5V parts I'd assume absolute max is 5.25v unless the datasheet says different. Doesn't mean 5.25v is safe to run boards at though, it could work fine for a while but it's like burning the candle at both ends.

If you want to try trouble shooting then have a look at a 68k pinout and try probing these. Slow probe shouldn't be an issue for these, as long as the 68k is powered and it looks like the clock is working:
--- RESET and HALT: should be mostly high and every time you hear the click of death, both should be low for a split second.
--- AS: should pulse. Bare minimum to expect from the 68k if it isn't dead. If it's not dead but quiet then a custom part connected to the 68k could be dead as well.
--- A1-A3: should pulse. Low address pins should be busy if it's doing anything useful.

Could also check the OE pin on BIOS to see if it's pulsing. This could be hard to spot if the 68k crashes immediately. It definitely shouldn't be stuck low though.

Ill get some spare THS7374’s regardless. They’re like 88 cents on digikey so might as well have some extras kicking around too. I still need to get a proper 8 pin din connector to wire it up to the RGB output. Although whenever this mentality ensues for any cheap part, I always end up with a hoard I never use :P

I’ll try probing those pins tomorrow when I get home. Im vaguely familiar with how the neo geo works on a surface level, but not enough to diagnose it on my own. The neodev wiki seems to be good but I’m still not exactly sure how everything works together as a whole system

I found an MV1A and it has the vandalized cartridge holder with letter C wriitten on it in various places, I'm guessing for CHINA.I will test it out overnight to see what gives..
BB

Let me know how it goes and hopefully we can cut a deal. Thanks again.

 

[maki]

My supergun also has a THS7374, what are the odds that it took that with it as well?

IMO odds are high good thing the THS7374 is readily available for cheap, thats not to worry about too much IMO.

It would be an interesting project to try to fix this.
At the same time, the MV1AX isn't really rare or that expensive (compared to a board with a 1st gen chip set), if you want something working quick, look for replacement.

 

[mehguy]

IMO odds are high good thing the THS7374 is readily available for cheap, thats not to worry about too much IMO.

It would be an interesting project to try to fix this.
At the same time, the MV1AX isn't really rare or that expensive (compared to a board with a 1st gen chip set), if you want something working quick, look for replacement.

Yes that is exactly what I’m thinking. I’ll do this as a long term project and just get another 1 slot board in the mean time.

It’ll definitely be a fun and interesting project but I have no idea where to begin with something like this. I found an interesting tid bit on the neodev wiki regarding the PALs.

MV1A PALs - NeoGeo Development Wiki

wiki.neogeodev.org

If any of them fails, the board certainly won't start up.

Now I can’t tell if they mean “start up” as in being in a state able to run and execute code (which mine isn’t even doing) or if it’s just watchdoging?

 

[maki]

you can disable the watchdog (short JP2, Pin 83 of the GRC2 to GND) and see how far it gets

with the older models, RESET could be generated by the 68k, but also from the NEO-B1, lifting the pin fro either would reveal who is pulling RESET high, but I don't see the RESET signal for the GRC2 or the MGA, but there is bigger gaps in their documentation on Neo Geo Dev, you might be able to trace the pin out, should be connected to the 68K RESET pin

 

[mehguy]

you can disable the watchdog (short JP2, Pin 83 of the GRC2 to GND) and see how far it gets

with the older models, RESET could be generated by the 68k, but also from the NEO-B1, lifting the pin fro either would reveal who is pulling RESET high, but I don't see the RESET signal for the GRC2 or the MGA, but there is bigger gaps in their documentation on Neo Geo Dev, you might be able to trace the pin out, should be connected to the 68K RESET pin

I've disabled the watch dog and have a solid red screen now and no click of death.

Wouldn't worry about the passives as much as the chips. Datasheets usually have an absolute max rating table and the THS7374 part you mentioned is max=5.5v, so that doesn't look good. No way to know until you try. For 5V parts I'd assume absolute max is 5.25v unless the datasheet says different. Doesn't mean 5.25v is safe to run boards at though, it could work fine for a while but it's like burning the candle at both ends.

If you want to try trouble shooting then have a look at a 68k pinout and try probing these. Slow probe shouldn't be an issue for these, as long as the 68k is powered and it looks like the clock is working:
--- RESET and HALT: should be mostly high and every time you hear the click of death, both should be low for a split second.
--- AS: should pulse. Bare minimum to expect from the 68k if it isn't dead. If it's not dead but quiet then a custom part connected to the 68k could be dead as well.
--- A1-A3: should pulse. Low address pins should be busy if it's doing anything useful.

Could also check the OE pin on BIOS to see if it's pulsing. This could be hard to spot if the 68k crashes immediately. It definitely shouldn't be stuck low though.

Checked all those pins and they seem to be working as they should. Although for whatever reason A1-A3 pulsing between Low and Floating according to my logic probe.

When I checked those pins again with the watchdog disabled, they're all static. RESET and HALT are stuck High while AS and A1-3 are stuck low. I even checked the clocks from Neo-D0 and they're all stuck low with the watchdog disabled.

 

[maki]

I've disabled the watch dog and have a solid red screen now and no click of death.

what BIOS was that with?
AES will report a work RAM error with a solid red screen
maybe give the Diag BIOS another go with the watchdog disabled

 

[mehguy]

what BIOS was that with?
AES will report a work RAM error with a solid red screen
maybe give the Diag BIOS another go with the watchdog disabled

Change of bios makes no difference. It still gives me a red screen even with none installed.

 

[BIG BEAR]

I powered up my 1A and it has a sync issue and no socket for bios... so I'll be throwing it back in the pile for now. Oh well..I tried.
BB

 

[mehguy]

I powered up my 1A and it has a sync issue and no socket for bios... so I'll be throwing it back in the pile for now. Oh well..I tried.
BB

No worries. Thanks for taking a look. Looks like I’ll have to buy from one of the AliExpress sellers again heh

But with regards to diagnosing it, is it even possible to figure out what chips are bad? If it’s true that so many could’ve been fried, can we even pin point which ones are bad?

 

[maki]

If the Neo Geo doesn't boot at all, the DIAG Bios rarely does help IMO

try to check that all the signals are present that would lead to a "click of death" if missing

Signals - NeoGeo Development Wiki

wiki.neogeodev.org

 

[BIG BEAR]

No worries. Thanks for taking a look. Looks like I’ll have to buy from one of the AliExpress sellers again heh

But with regards to diagnosing it, is it even possible to figure out what chips are bad? If it’s true that so many could’ve been fried, can we even pin point which ones are bad?

Just by the clicking you're receiving ,that right there is a tell that at least one of the SMT's got fried. You could power up the unit and feel the tops which will usually get real hot if they're bad.
BB

 

[mehguy]

Just by the clicking you're receiving ,that right there is a tell that at least one of the SMT's got fried. You could power up the unit and feel the tops which will usually get real hot if they're bad.
BB

None of the chips get really hot. Some get slightly warm, but nothing more than that. The warmest is the ym2610, which I assume is normal.

 

[BIG BEAR]

None of the chips get really hot. Some get slightly warm, but nothing more than that. The warmest is the ym2610, which I assume is normal.

Yeah , that's normal for the YM2610...Try leaving it on for like 15 -20 minutes and see what happens...
BB

 

[mehguy]

Yeah , that's normal for the YM2610...Try leaving it on for like 15 -20 minutes and see what happens...
BB

Ok, after about 20 minutes, quite a few chips get warm. All of them are less warm than the Ym2610 but I'd say the Neo GRC is the warmest next to the ym2610. Nothing hot to the touch though.

 

[mehguy]

Had just checked again and what's interesting is the solid color changes between red, yellow, green, orange etc whenever I cycle the power.