View Full Version : Capcom A Board Repair

24th July 2008, 10:37 AM
Have had a stack of Capcom A boards from my contact recently, all with the same fault - no sound.

Usually you suspect the amp chip, but pressing on the pins on all the amp chips on all the boards resulted in the angry buzzing noise that means the amp is working.

It turns out that all the boards bar 1 had the exact same fault, dead Zilog Z80 chips. I have no idea how or why the Z80s all ended up dying, usually they are fairly bullet proof, but hitting the data lines with a logic probe showed some pins were floating.

Desoldered the Z80s, fitted sockets and installed new Z80 chips, and the sound was restored. Just for completeness I put the old Z80s into my Galaxians board, and ended up with a screen full of colourful crap - confirmed the Z80s were toast.

Piggybacking a new Z80 on the old one didnt achieve anything either by the way.

The single board that had a different fault did have a dead Z80, but it also had a dead 2018 ram chip that is the sound cpu's work ram. Replacing both the 2018 and the Z80 fixed the board.

I must have fixed about a dozen of these boards now, all with the above fault, at the bottom of the pile I came across an older style Capcom A board, it had a Sharp Z80 chip instead of a Zilog and the sound section is not quite as close to the board edge as it is on the later more compact boards. By this time I had stopped plugging the boards into my cab and had just resorted to hitting the lines with my logic probe on my test bench, which currently has no speaker connected. I went over the Z80 and everything looked fine, all the lines were active, the ram chip was happy. I couldnt see anything wrong, based on what I had been looking for. So I plugged the board into my cab and the sound was fine, there was nothing wrong with it. No idea how the board had ended up being sold in a batch of faulty ones.

Still, it kinda proved my point about spotting what seems to be the common fault, the only time I couldnt find the fault it turns out there wasnt one. :)

24th July 2008, 12:24 PM
Thanks for the post. I cant say that I have ever encountered that problem and I have owned my fair share of cps2 stuff. I wonder if someone went on a spree of plugging them into the harness the wrong way round. :unsure

24th July 2008, 01:16 PM
Its CPSI gear, nah - plugging into the harness the wrong way round would kill an awful lof of the stuff on the board.

I can only assume the Z80s were heat damaged when the board was flow soldered, and they managed to survive but the life of the chip was much reduced.

24th July 2008, 01:20 PM
Another problem too can be poor encapsulation - a whole batch may not be hermetically sealed, and can have a reduced life. That has happened quite often... Ask Fujitsu why they no longer make consumer HDDs!

2nd September 2008, 11:31 AM
You could be right there - another 3 in this month, all with dead Z80s and no other faults, its bizarre its so common a fault, and on decent grade hardware too.

15th January 2009, 12:53 AM
Apparently the CPS1 boards are supposed to use a Z80-A CPU but instead many have been fitted with the cheaper Z-80 cpu which runs hotter and is more prone to failure because its clock is over driven.

21st January 2009, 01:09 PM
That would certainly explain it - will have to see what my collection of toast Z80s consist of.

30th March 2009, 08:47 AM
Had another couple of CPS1 A boards in this weekend with "no sound" stickers attached. Poked my scope at the Z80 on the 1st one and its data lines were pretty ugly looking, very very faint signals, waving all over the place. Desoldered it, dropped in a socket and a new Z80 and the board was 100% again.

I looked at my ever growing heap of Z80s from these boards and they are all Z0840004PSC chips - which according to the Zilog site are 4Mhz chips, MAME says the audio for CPS1 games has the Z80 clocked at 4Mhz so they are all in theory the correct chip, yet I now have nearly a dozen of them from A boards that are toast. I assume they are badged correctly.

The second no-sounder had a dead crystal oscillator, pin 6 on the Z80 was doing nothing, went back to the oscilator and it was getting 5V and its ground connection was good, but its output was stuck low, so no clock signal at all. Replacing it with one from a scrap A board fixed the sound, it still has its original Z80 too.

30th March 2009, 11:50 AM
I've heard/read that the boards are silkscreened to take Z80-As, the original Z80s are rated to 2.5Mhz IIRC. The Z80-A is rated for 4Mhz.
I have also seen Z80s that are labelled Z80, with a string of numbers like "Z0840004DSE" and yours, to be rated for 4Mhz operation... no idea why they'd make our life so difficult, I now to stick to stuff that's labelled Z80-A for 4Mhz operation, since the Z80 labelled ones seem to be dropping like flies.

30th March 2009, 12:16 PM
Yes, the board is silkscreened for Z80A underneath the CPU, and the dead CPUs on these boards are Z80As in that they are rated as 4Mhz chips. No idea why the seem such a common point of failure.

30th March 2009, 04:10 PM
very instructive, thank you !