MV-1AX Flashing Green Screen and Watchdog Clicking

GadgetUK

Ace Ghost Pilot
Joined
Sep 27, 2013
Posts
1,323
I've got 3 boards here that I think have bad Palette RAM and the neo-buff associated, and it cause the same sort of behaviour you are seeing. Although I've not measured for shorts like you did there - so maybe you have a faulty MGA? With the Palette neo-buf it seems to totally kill the system as the neo-buf interferes with the 68K databus. The other thing I've found on the 8 boards I am working with (all 1A/1AX boards) is that often more than one neo-buf chip fails. I've seen the one you removed (Z80 interface) cause the 68K to watchdog with a diag bios or official bios. These boards are a pain to work on due to all the damn neo-buf chips. 2 or 3 of the boards have issues with nearly every neo-buf chip on them. You could remove the one from GA11 (used by the PROG interface to 68K) and fit it into the Palette spot to see if behaviour changes, but I've found that when that dies in that spot it can take the Palette RAM out (or vice versa). You can also btw remove GA11, GA7, GA8, and GA9 to rule them out as interfering with the 68K databus - the board should boot without them. But maybe the neo-bufs aren't your problem?!?

EDIT: That exact blue or green screen is what I've seen when Palette RAMs have died AND associated neo-buf.

Do you have the M1 test cart to go with diag ROM btw? If your system is booting, and the controller neo-buf is working, hold down D when booting and listen for sound, because that can provide a clue - if you hear the sound test being carried out despite a solid colour on the screen.
 
Last edited:

Dropwire

Kuroko's Training Dummy
Joined
Jul 27, 2018
Posts
75
I've got 3 boards here that I think have bad Palette RAM and the neo-buff associated, and it cause the same sort of behaviour you are seeing. Although I've not measured for shorts like you did there - so maybe you have a faulty MGA? With the Palette neo-buf it seems to totally kill the system as the neo-buf interferes with the 68K databus. The other thing I've found on the 8 boards I am working with (all 1A/1AX boards) is that often more than one neo-buf chip fails. I've seen the one you removed (Z80 interface) cause the 68K to watchdog with a diag bios or official bios. These boards are a pain to work on due to all the damn neo-buf chips. 2 or 3 of the boards have issues with nearly every neo-buf chip on them. You could remove the one from GA11 (used by the PROG interface to 68K) and fit it into the Palette spot to see if behaviour changes, but I've found that when that dies in that spot it can take the Palette RAM out (or vice versa). You can also btw remove GA11, GA7, GA8, and GA9 to rule them out as interfering with the 68K databus - the board should boot without them. But maybe the neo-bufs aren't your problem?!?

EDIT: That exact blue or green screen is what I've seen when Palette RAMs have died AND associated neo-buf.

Do you have the M1 test cart to go with diag ROM btw? If your system is booting, and the controller neo-buf is working, hold down D when booting and listen for sound, because that can provide a clue - if you hear the sound test being carried out despite a solid colour on the screen.

Thanks for the advice Gadget, I do have the M1 Test Cart, just have not been able to get the system to boot yet. Still stuck on the watchdog reset with garbled data on screen.

- No progress thus far on locating the cause of the short, removing the NEO-MGA improved the resistance to ground but there is still something else on the board that is possible shorted from VCC to GND.

- Removed Neo-BUF (GA4) near the Palette RAMs, and re-checked the board resistance between VCC and GND no change. I may try also removing the 2 Palette RAMs and checking the resistance again. I figured if the NEO-BUF (GA4) was bad I would have seen the short to ground disappear when I removed it. I may also try stealing the NEO-BUF (GA11) as suggested by GadgetUK and putting in GA4's position.

Also in case anyone is interested below are the images of the NEO-MGA & NEO-BUF (GA4) positions with the ICs removed. In-case you want to see where the traces go, and the pin pads connect to.
 

Attachments

  • NEO-MGA Traces.jpg
    NEO-MGA Traces.jpg
    935.6 KB · Views: 76
  • Neo-BUF GA4 Traces.jpg
    Neo-BUF GA4 Traces.jpg
    1.4 MB · Views: 76

RyuX

Crossed Swords Squire
Joined
Sep 19, 2018
Posts
179
Really interesting to see the troubleshooting progress. I would really love to open the images - however I am a new member and it seems I can't open files yet ? I Would really love to check if your errors look similar to mine (I have 3 broken MV-1FZ that all have a similar issue and also troubleshooting at the moment). Good luck with your repair!
 

Dropwire

Kuroko's Training Dummy
Joined
Jul 27, 2018
Posts
75
Really interesting to see the troubleshooting progress. I would really love to open the images - however I am a new member and it seems I can't open files yet ? I Would really love to check if your errors look similar to mine (I have 3 broken MV-1FZ that all have a similar issue and also troubleshooting at the moment). Good luck with your repair!

I think there is a short probation period before you can view the images.

What is the trouble with your MV-1FZ? You should open a new post for it and we can discuss it. In the meantime if you want any of the images I have in this post let me know and I can email them to you until you get your access upgraded.
 

Dropwire

Kuroko's Training Dummy
Joined
Jul 27, 2018
Posts
75
I have found the Short to Ground, unfortunately it is the 68HC000 Processor. Looking to try one from a donor board or possible order a replacement with an PLC68 SMT Socket.

Some Observations

- After checking several MVS boards (MV-1AX, MV-1FZ, MV-1FZS, MV-1C, MV-1B) the normal "Board" resistance when taken at the VCC (+5) & GND Test points next to the JAMMA Edge is averaging 300k ~ 600k Ohms depending on the board. The board I am working on currently (MV-1AX) was showing a short of about 50 ohms, and now is back in the normal range after removing the 68HC.

- When using a DMM to get as close to accurate resistance readings, if your meter has a "relative" (REL^) function use it to calibrate your leads prior to making measurement. This will set the meter to take into account the inherent resistance your leads have. Effectively treating your shorted leads as "Zero". While the difference will be minuscule, probably in the order of .10 ohms it makes a difference when trying to hunt for a short on a component as you will be looking for that minute change as you measure from component to component.

- I think this will be a new 1st step method for me when testing a watchdog resetting board, I have 3 other MVS's (MV1B) that are Watchdog reset and decided to check the "Board Resistance" before doing anything else. Lo and behold 1 of the 3 has a board short to ground same as this MV-1AX.

- As I have been removing components off the board, I have been placing sockets where possible for the through-hole ICs, and I plan to place a PLC68 surface mount socket for the 68HC000

- Once I obtain a replacement processor I'll post my findings, hopefully successful repair of this board.

:scratch:
 

RyuX

Crossed Swords Squire
Joined
Sep 19, 2018
Posts
179
My garbage screen / watchdog board also shows around 85 Ohms to Ground while my other Board that "only" has some graphical problems but is booting fine will show 480k - So there is something to it.
However on my bad board I replaced the CPU already and it didn't do any good.. might be a trial and error process then..
 

Dropwire

Kuroko's Training Dummy
Joined
Jul 27, 2018
Posts
75
My garbage screen / watchdog board also shows around 85 Ohms to Ground while my other Board that "only" has some graphical problems but is booting fine will show 480k - So there is something to it.
However on my bad board I replaced the CPU already and it didn't do any good.. might be a trial and error process then..

I am trying to work on a process for finding the shorted component without having to randomly remove ICs. In my case it was the CPU, however it could have been any other component. I am thinking the key here will be to use the current measurement at the VCC for the IC you are testing to see if the amperage draw is higher / lower than the other components. Have not tested this idea yet...:eye:
 

Dropwire

Kuroko's Training Dummy
Joined
Jul 27, 2018
Posts
75
Here are some images of the 68HC000 VCC short to GND test. Also included are images of the pre-test I do with my meter leads before and after using the REL function on the DMM to compensate for the inherent resistance in the meter leads.

The 58 ohm short is between Pin#14 (VCC) & Pins #16 /#17 (GND).

68HC000 Datasheet.jpg68HC000 VCC_GND Short.jpg68HC000 VCC_GND Short2.jpg68HC000 Resistance.jpgLead Test.jpgLead Test REL.jpg
 

Dropwire

Kuroko's Training Dummy
Joined
Jul 27, 2018
Posts
75
- Still working on the board, figured I'd take some pictures during this process, in case anyone wants to see the traces or pads with the components removed. Bear in mind that this is the insanity I went through trying to find the shorted component. Still working on ideas to simplify this process. :loco:

- Also I received the PLCC68 socket and installed it on the board, just waiting for the replacement 68HC000.


CPU empty.jpgCPU_Socket.jpgNEO_D0 Empty.jpgNEO_F0 Empty.jpgNEO_GRC Empty.jpgNEO_IO Empty.jpgNEO_ZMC2 Empty.jpg
 

ack

Ninja Combat Warrior
15 Year Member
Joined
Apr 9, 2009
Posts
538
Thanks for the pics.

I'm curious if there is much interest in making board scans with all components removed. I've been thinking about making them for a while, but its a lot of work.
 

Dropwire

Kuroko's Training Dummy
Joined
Jul 27, 2018
Posts
75
I have been scouring various internet sources and some old documents I have regarding PCB component fault locating. There are a few methods that may be used to find either "Hard Short" or "Low Resistance Path to Ground" condition on boards. Some involve unique test equipment like the "ToneOhm" , "LeakSeeker 89", "ShortSniffer", or using an ESR meter for its resistive measurements. As I have a few DMMs and an ESR meter I have been trying to figure out how to use what I have. While there are many techniques, none are fool proof and ultimately get you close enough to make an "educated guess" as to which component is defective. Trying to get that "educated guess" narrowed down so as not to have to remove too much.

- ESR Meter or DMM with a good resistive resolution, thinking is to measure resistance between VCC and GND on each component in question and see which shows the lowest resistive value. For example the board resistance between VCC and GND is 50 ohms, so measure each component and look for the lowest value. this is where the meter your using needs to have a good resolution so you can see the minute differences. If your measuring the components, since they are all tied to the VCC and GND plane of the board they will all show the 50 ohm short, and as you measure you may find a component that reads 49.99 ohms which could be the defective component.

- DMM Meter checking for voltage drop using a power supply with a controlled voltage and current. Setting a power supply to 5Vdc at 100mA (low current to not further damage components), then start checking the voltage drop across the components. First measuring the voltage at the VCC / GND board main test point to get the reference values. Then moving on to each component (areas of the board). Setting the meter to the DC range and then measuring between the VCC test point and the VCC Pin on the component in question to see the voltage drop. The component with the "highest" voltage reading (which will probably be in milli-volts) should be the bad component. Found a write-up on this " http://areksnotes.blogspot.com/2012/09/how-to-find-short-circuit-on-your-pcb.html "

- DMM Meter checking for current Draw based on Ohm's Law. V=I*R -or- I=V/R. So in a "Dead Short" scenario the current would measure at 0 amps, but in the case of a low resistance short (using the 50 ohm as example) the current at 5V would be 100mA. So looking for the component pulling the most current. This is why using the regulated power supply with a controlled current would be important to limit the current in the case of a low ohm short. In theory as the short resistive value approaches 0 it would draw more current. For example a 1 ohm short would draw 5 amps.

- Also the Temperature test with either a thermometer, or just feeling with your fingertips for the component that is the warmest.

Just some of the thoughts I have, please chime in if you have any ideas or if something is mis-stated. Thanks .... :scratch:
 

Dropwire

Kuroko's Training Dummy
Joined
Jul 27, 2018
Posts
75
Thanks for the pics.

I'm curious if there is much interest in making board scans with all components removed. I've been thinking about making them for a while, but its a lot of work.

I was thinking the benefit would be to see the traces and vias, in the event someone is repairing a board and they are damaged, or missing. As some traces can start under and IC and continue through to the Solder PAD then out to the board, and if the trace is broken outside of the IC you may not realize that it continues under the IC. Just a thought.

If it helps let me know if you want more pics or need me to check something on the board.
 

RyuX

Crossed Swords Squire
Joined
Sep 19, 2018
Posts
179
I was thinking the benefit would be to see the traces and vias, in the event someone is repairing a board and they are damaged, or missing. As some traces can start under and IC and continue through to the Solder PAD then out to the board, and if the trace is broken outside of the IC you may not realize that it continues under the IC. Just a thought.

If it helps let me know if you want more pics or need me to check something on the board.

I have one MV-1FZ for exactly that reason - as a reference - i removed all parts from that one. It helps quite a lot
 

Dropwire

Kuroko's Training Dummy
Joined
Jul 27, 2018
Posts
75
Current update:

- Began reassembling the board with the components I removed that have tested good.

- Still waiting for shipment of replacement 68HC000 FN12 Processors

- Discovered 2 more shorted components, the NEO-MGA & RAM7

- Will begin searching for a NEO-MGA replacement.

- RAM7 is a standard component, may order new replacement or temporarily steal one from another board.

- Wondering if NEO-MGA is used for addressing, until I can obtain a replacement is it possible to still test the board and will it “Watchdog” with the NEO-MGA removed (as it currently is off the board due to the hard short to ground it has)

- Socketed all the through-hole components as well as the Main Processor, if I can get this board running I may use it as a test-bed.

- Pondering the meaning of the universe and why SNK designed the MV-1AX so oddly...
 

RyuX

Crossed Swords Squire
Joined
Sep 19, 2018
Posts
179
Current update:

- Began reassembling the board with the components I removed that have tested good.

- Still waiting for shipment of replacement 68HC000 FN12 Processors

- Discovered 2 more shorted components, the NEO-MGA & RAM7

- Will begin searching for a NEO-MGA replacement.

- RAM7 is a standard component, may order new replacement or temporarily steal one from another board.

- Wondering if NEO-MGA is used for addressing, until I can obtain a replacement is it possible to still test the board and will it “Watchdog” with the NEO-MGA removed (as it currently is off the board due to the hard short to ground it has)

- Socketed all the through-hole components as well as the Main Processor, if I can get this board running I may use it as a test-bed.

- Pondering the meaning of the universe and why SNK designed the MV-1AX so oddly...

Nice work again.. I will also socket my CPU as soon as I get my sockets.
I will also start work on my very bad board - how about the 5V to GND issue ? Any progress on that ?

A test board is always a good idea..
 

Dropwire

Kuroko's Training Dummy
Joined
Jul 27, 2018
Posts
75
So to recap, it seems I have 3 bad "Shorted" components on this board.

- 68HC000 FN12 Processor, NEO-MGA, and RAM7 (BR6265BF-10SL)

Having Difficulty re-installing the NEO-GRC as some of the PADS are either missing or in bad shape, as well as pins 200 ~ 205 seem bent/deformed. It would seem whomever had this board before may have attempted some sort of parts swapping after realizing that the board may have been over-powered or something to cause the multiple components to be damaged.

So as it stands if I can re-install the NEO-GRC and repair the connections to it, replace the 68HC000, and the RAM7. It still leaves me with the issue of finding someway to replace the Neo-MGA.

Trying not to get frustrated, or give up on this one. May just put it down for a few days then re-visit it with a clear mind. :very_ang::mad::blow_top:
 

RyuX

Crossed Swords Squire
Joined
Sep 19, 2018
Posts
179
So to recap, it seems I have 3 bad "Shorted" components on this board.

- 68HC000 FN12 Processor, NEO-MGA, and RAM7 (BR6265BF-10SL)

Having Difficulty re-installing the NEO-GRC as some of the PADS are either missing or in bad shape, as well as pins 200 ~ 205 seem bent/deformed. It would seem whomever had this board before may have attempted some sort of parts swapping after realizing that the board may have been over-powered or something to cause the multiple components to be damaged.

So as it stands if I can re-install the NEO-GRC and repair the connections to it, replace the 68HC000, and the RAM7. It still leaves me with the issue of finding someway to replace the Neo-MGA.

Trying not to get frustrated, or give up on this one. May just put it down for a few days then re-visit it with a clear mind. :very_ang::mad::blow_top:

I feel your pain. My LCPS2-A2 has broken off pins and the traces/pads are also damaged. I wanted to give up on this already but now managed to get a very very thing wire soldered to the broken pin and then glued it in place. slowly starting the repair - trace after trace even knowing that this for sure might not even be the main problem. But I see it as a good training to improve the learning curve on these boards.

I think a overpowered board will have so much damage that a repair is almost not worth it - most defects are hard to spot and some components may only have intermittent behavior at this point.
I am having around 2-3 boards worth of old desoldered components around. Most of them untested - but you never know when you gonna need them.
Resoldering the tiny pitched SMD Components like the LSPC2-A2 is especially a pain in the $$$ when the pins are bent - the pads are broken. Looking at the NEO-GRC QFP208 Package I can only imagine the trouble you are going through.

But wish you the best of luck to get it at least to fire up.
 

BIG BEAR

SHOCKbox Developer,
20 Year Member
Joined
Dec 14, 2001
Posts
8,230
You are doing a lot of good work.
Keep at it after you take a breather...I can supply you with the NEO-MGA.
BB

So to recap, it seems I have 3 bad "Shorted" components on this board.

- 68HC000 FN12 Processor, NEO-MGA, and RAM7 (BR6265BF-10SL)

Having Difficulty re-installing the NEO-GRC as some of the PADS are either missing or in bad shape, as well as pins 200 ~ 205 seem bent/deformed. It would seem whomever had this board before may have attempted some sort of parts swapping after realizing that the board may have been over-powered or something to cause the multiple components to be damaged.

So as it stands if I can re-install the NEO-GRC and repair the connections to it, replace the 68HC000, and the RAM7. It still leaves me with the issue of finding someway to replace the Neo-MGA.

Trying not to get frustrated, or give up on this one. May just put it down for a few days then re-visit it with a clear mind. :very_ang::mad::blow_top:
 
Last edited:

Dropwire

Kuroko's Training Dummy
Joined
Jul 27, 2018
Posts
75
Thanks all for the help and encouragement, working on a game plan now for this board, going to try to get back to it this weekend. I’ll keep you all posted.
 

Dropwire

Kuroko's Training Dummy
Joined
Jul 27, 2018
Posts
75
You are doing a lot of good work.
Keep at it after you take a breather...I can supply you with the NEO-MGA.
BB

Still working on this board, just took a little time off to clear my head on this one.

- Big Bear if your offer still stands, let me know how much you want for the replacement NEO-MGA.

Also if anyone needs my help with anything, just shoot me a PM. I may be on the boards a little intermittently during this upcoming holiday season.

As usual thanks all for your help and support.

:loco:
 

BIG BEAR

SHOCKbox Developer,
20 Year Member
Joined
Dec 14, 2001
Posts
8,230
good to see you back. Still stands.I'll be dropping you a PM this morning. Now where did I put that PCB?!
BB
 
Top