mv2f watchdog

Hairy Otter

Neo Bubble Buster
Joined
Aug 11, 2012
Posts
53
I've got myself a 2 slot mvs board with a watchdog problem.
The connections all seem to be allright. But if I look at the Address lines on the 68000 I find the A4 and A5 (pin 32 & 33) to stay low, other Address lines give a short pulse to high with each kick of the watchdog. What can cause this? Or do I have look somewhere else for the problem?
 

Xian Xi

JammaNationX,
Joined
Dec 1, 2005
Posts
27,702
Could be bad work RAM but most likely a bad trace. If A4 and A5 are bad, you need to check the connection between both work RAMs. Address lines are the same between both. It's possible that the connection for the address lines between the two RAMs is bad.
 

Hairy Otter

Neo Bubble Buster
Joined
Aug 11, 2012
Posts
53
@Xian Xi
I used your pinout found here:
http://www.neo-geo.com/forums/showt...-1st-Gen-AES&p=3301797&viewfull=1#post3301797

Pin 20 is connected to Pin 59 on the NEO-GO chip. This one is stuck high?

all connections are ok
all address lines tested to both work ram's and backup ram (all connected together)
data lines only go to one of both work and backup ram (this is similar on my working 2 slot)

does it help to mention that i get the same screen, with or without the bios. Also only the top 1/4 of the screen is filed with garbage on the rest of the screen are only a few blocks (one is always visible) the rest at random.
 
Last edited:

Xian Xi

JammaNationX,
Joined
Dec 1, 2005
Posts
27,702
Do you have an eprom burner? I would suggest burning an smkdan Neo Diagnostic Bios.
 

Hairy Otter

Neo Bubble Buster
Joined
Aug 11, 2012
Posts
53
I have no ROM burner, and also do not know if I'm going to buy one in the near future. My budget is limited and I do not know what I'd best buy. Any tips?
However "smkdan Neo Diagnostic Bios" seems useful. Is it possible to buy a burned EPROM?
 
Last edited:

Pedrobear

Cheng's Errand Boy
Joined
Mar 1, 2012
Posts
113
You can disable the watchdog (short J2 above NEO-B1), but it seems problem is deeper than that, how does the board operates exactly?
 

Hairy Otter

Neo Bubble Buster
Joined
Aug 11, 2012
Posts
53
You can disable the watchdog (short J2 above NEO-B1), but it seems problem is deeper than that, how does the board operates exactly?

I already tried this, the only thing that happened is that the image remains stationary and the clicking stops. Or do i need to reboot with J2 shorted?
The board doen't operate at all, here is a short video with multiple attempts to start the board. The first time after some time the lower half is black, next time I reboot within a short time it becomes red.
I hope to revive this board as these two slots seem very vulnerable.
 

Hairy Otter

Neo Bubble Buster
Joined
Aug 11, 2012
Posts
53
Got my smkdan Neo Diagnostic Bios. Popped it in there, but nothing changed. The OE* is stuck high so the bios is never read so no mather which bios it is. (the Diagnostic Bios wil come in handy for my other faulty boards, and maybe for this one if I get it to boot)
I followed the signal from the OE to the NEO-GO. Here the signal is split into two, namely SROMOEU and SROMOEL. These are also high, and I can follow them up to the NEO-C1. Then I get lost. Is there any way to test if the NEO-C1 is working ok?
 

smkdan

Galford's Armourer
Joined
Dec 30, 2009
Posts
452
Some traces to check if BIOS /OE is completely dead:

NEO-C1:
-Everything under the 68k section on the wiki including the A22/23 pins from NEO-E0
https://wiki.neogeodev.org/index.php?title=NEO-C1

NEO-E0:
-68k address
-"VEC" (from 74259)
-A22/A23 outputs (to NEO-C1)
https://wiki.neogeodev.org/index.php?title=NEO-E0#MV2B_.40_H7_pinout

74259:
See page 3 of the schematic.

Here's an older thread where you can get some troubleshooting ideas. This one was traced back to a bad 68k but it's just one of many possible causes:
http://www.neo-geo.com/forums/showt...eo-AES-3-6-click-of-death-(Stuck-in-watchdog)

If the BIOS isn't being read then what you see on screen doesn't matter. It's just displaying garbage from the video memories until the 68k actually writes something to them, which it won't without BIOS access.
 

Hairy Otter

Neo Bubble Buster
Joined
Aug 11, 2012
Posts
53
I've checked the continuity of all address lines of the 68k to the NEO-C1 and NEO-E0 and they are all ok. The VEC is at logic low but this compares to my booting mv2f. The A22/23 outputs have continuity to the NEO-C1.
On the NEO-C1 I checked:
R/W,UDS,LDS,AS,DTACK: 68k bus control signals
68KCLK: (SCLK?) 12MHz 68k clock

ROMWAIT,PWAIT0,PWAIT1,PDTACK I didn't check as they come from the cartridge, right?

the 74259 I didn't check as I cant make out the text in the schematic. But it is ok I think, as the output to the VEC is ok, or am I wrong here?
If I boot the board with the watchdog disabled (JP2) the A17 keeps pulsing and the rest go to a high impedance state, if I compare this to my booting MV2F none of the address keep pulsing? Doe this mean something?
 
Last edited:

smkdan

Galford's Armourer
Joined
Dec 30, 2009
Posts
452
BR + BERR on 68k can be checked too. They should be high through some resistors near the 68k. Apart from that, 68k CLK should be pulsing as with the C1 CLK.

With the watchdog disabled, what does 68k HALT look like? It should be low only for a short time on power on and completely high after that. The faulty 68k in that other thread was grounding HALT for whatever reason. How does yours look?

the 74259 I didn't check as I cant make out the text in the schematic. But it is ok I think, as the output to the VEC is ok, or am I wrong here?

If the reset input and VEC are OK then it's probably not the 259. The reset is the same reset line shared with the other components.

If I boot the board with the watchdog disabled (JP2) the A17 keeps pulsing and the rest go to a high impedance state

That's strange. You're sure that some address are lines high impedance but others are active *at the same time*?
 

Hairy Otter

Neo Bubble Buster
Joined
Aug 11, 2012
Posts
53
BR + BERR on the 68k are kept high by a 4.7KX4 resistorbridge: OK
68k CLK and C1 CLK both nice signal on my scope.
68k HALT line is always low (only a short peak when turning on, also those peaks in watchdog)

checked the addresslines on the 68k again, didn't find any pulsing this time but:
A6 LOW, A16 A17 HIGH, A18 A19 low, A20 A21 High, the rest is at high impedance.

Reading up on the older thread.
If all the 68k signals are mirrors of the reset then it's either dead or not getting the 12MHz clock. They should be pulsing much faster than the reset. See if it's getting the clock that it should and if it's 12MHz like expected, I think swapping the 68k is a good idea.
I took my scope to the 68k signals but I don't see any pulsing just straight high. But this is maybe because of the old 20Mhz analoge scope I use. Maybe I need a logic analyzer for this.

Status update:

WOHOOOOOO!!! Fixed it! :) :) :)
I had the unibios socketed, so I checked the original JAP Bios... It was fried too. So I'll leave the Unibios there :)
It seems I fried the original Bios from this board as it tested OK before I did more test, now it's not working in my other bord. Is this posible?

If I want to replace the 68k, I would socket it, but what socket do I need for this type?
 

smkdan

Galford's Armourer
Joined
Dec 30, 2009
Posts
452
It has an SDIP64 package. They came in both the full size DIP64 and SDIP64 packages so make sure it's the right one if you decide to try a swap.

A quick summary of what I think since I don't have much time right now: The stuck HALT pin is coming from either the 68k or B1. There's a very specific case where a working 68k on a Neo will output HALT like that but the address measurements are weird. I'm pretty sure (not certain though) that the entire address bus should be at high impedance in both STOP and HALT state, and both states are signalled with low HALT. I don't know of any hardware on the board either that can output onto the bus except the 68k itself, so I'd suspect the 68k if I had to guess based on what I'm seeing.
 

Hairy Otter

Neo Bubble Buster
Joined
Aug 11, 2012
Posts
53
Thanks for your reply, so the 68k should be bad.
I put this board on hold for now. Just until I get my 8 channel logic analyzer so I can measure the behaviour of the 68k in question, for future reference.
I'll update this thread when I make some progress. Either positive or negative.
I still have some other boards to fix. These 2 slots seem like a real bitch to me.
 

Hairy Otter

Neo Bubble Buster
Joined
Aug 11, 2012
Posts
53
I got my logic analyzer and checked de 68k address lines that were odd.
A6 drops to low on the first reset and pulses high?
A11 looks ok
A12 start making junk after e few seconds (between 1 and 5 seconds)
A13 starts low pulses high?
A20 drops to low on the first reset and stays low?





If I disable the watchdog I get this:



But with the watchdog disabled the reset line looks like this:



If I zoom in on the edge with watchdog enabled I see a nice clean signal.
Can somebody explain this?

To mee it looks like the 68k has seen better days so now I will be looking for a replacement, with a socket.
 
Top