Neo Geo Arcade board problems

[Uriah]

I have about four Neo-Geo six slot arcade game boards with Video RAM
errors. It is not the actual video RAM, that much I have determined.
It has a lot of large surface
mount chips with upwards over over a hundred pins each. It uses a
Motorola 68000 chip. Can someone give me some clues on how to track
this problem down? With this kind of error am I going to find an
address or data line that is stuck high or low? Or could it be another
pin. Nothing on the M68000 seems wrong, The chip checks out fine on a
tester. I have replaced the video RAMS and this M68000 so it must be
in another chip or in a trace. If I probe each and every pin of all of
the components would I for sure find the error or can this type of
error not show up as a hardware difference? Can this be a software
error? Lets say I find the pin that is not working, What would that
pin look like on a scope? Just a high 5 volt or low ground, or could
it be at a logic high or low, that it is just not transitioning from
one state to the next as it supposed to? Could it be more then one
pin? There is no documentation for this board at all. I am not sure
where to began and what to do next.
Thanks very much
Russ

 

[Nobody]

On Tue, 03 Mar 2009 14:25:12 -0800, Uriah wrote:

I have about four Neo-Geo six slot arcade game boards with Video RAM
errors. It is not the actual video RAM, that much I have determined.

Then why do you conclude that the problem is "Video RAM errors"?

What are the symptoms?

If you are getting graphical glitches, it could be the video RAM, or the
circuits which write data to the video RAM, or the circuits which take the
data from the video RAM and generate the video signal.

It has a lot of large surface
mount chips with upwards over over a hundred pins each. It uses a
Motorola 68000 chip. Can someone give me some clues on how to track
this problem down? With this kind of error am I going to find an
address or data line that is stuck high or low? Or could it be another
pin. Nothing on the M68000 seems wrong, The chip checks out fine on a
tester. I have replaced the video RAMS and this M68000 so it must be
in another chip or in a trace. If I probe each and every pin of all of
the components would I for sure find the error or can this type of
error not show up as a hardware difference?

Any error will ultimately manifest itself through the signals which appear
on the tracks. Whether or not you can recognise the error on a scope trace
is a different matter. If a track carries a stream of 0s and 1s from one
chip to another, you probably aren't going to be able to recognise the
right data from the wrong data from a trace (and not necessarily from a
capture unless you understand the functioning of the chips involved).

Can this be a software
error? Lets say I find the pin that is not working, What would that
pin look like on a scope? Just a high 5 volt or low ground, or could
it be at a logic high or low, that it is just not transitioning from
one state to the next as it supposed to? Could it be more then one
pin?

An error could be a signal which is 011100110 instead of 010100110. You
aren't likely to recognise such an error without at least a good
understanding of correct behaviour and a logic analyser.

There is no documentation for this board at all. I am not sure
where to began and what to do next.

Personally, I think it's somewhere between "optimistic" and "futile" to
locate faults on such a complex board if you have neither documentation
nor a general understanding of how such systems function. Even if you
fully understand the board, it may be hard to identify a fault without a
logic analyser.

However, there are some types of fault which might be simple to identify.
If you get a response to shaking or flexing the board, that suggests dry
joints (finding them is a different matter). Also, try scoping the supply
voltage (Vcc) pins on all of the chips; if the decoupling capacitors have
failed, you would expect to see ripple at high (MHz) frequencies.