CPS1 Repair Logs

From Arcade Otaku Wiki
Jump to navigation Jump to search

Repairer: channelmaniac

The CPS1 games take different types of EPROMs depending on the B board type and the jumper settings.

For 32pin 1Mb EPROMS:

If the silkscreen says 27C301 then it takes NON-JEDEC type 1Mb EPROMs.

If the silkscreen says 1M ROM then it could take either and if it says 27C1001 then it takes the JEDEC type 1Mb EPROMs (27C1001, 27C010, etc...)

For the 40 pin EPROMS/Mask ROMs:

Look at the silk screen... if it says 27C4096 then it's the JECEC style pinout. 27C4002, 27C4095, etc. will work.

If it says HN62404 then it is a totally different pinout. 27C400 EPROMs are what you need to replace these. (NOT 27C040 which are only 32pin 8 bit EPROMs)

Let's look at board jumpers on board 89624B.

Now you might have noticed some solder jumpers down towards the 7F and 8F board area. These jumpers (JP0 - JP1 and DJ0 - DJ1) allow you to change the EPROMs in sockets 28 - 31 and 33 - 36 from the JEDEC to non-JEDEC pinouts.

If the jumpers are set like this, then it is set up for a JEDEC style pinout EPROM.

1  4


1  4

And if they are set like this, then it is set up for non JEDEC style EPROMs

1  4


1  4

DJ0 controls 28/33
DJ1 controls 29/34
PJ0 controls 30/35
PJ1 controls 31/36

VJ0 and VJ1 do the same thing for the Voice ROMs.

VJ0 controls 18
VJ1 controls 19

I haven't examined the jumpers at LJ3, LJ4, RJ3, or RJ4. They appear to control the A16 and OE* signals for the ROMs at 1-8, 10-17, and 20-27 but I haven't needed to pin them out yet.

Troubleshooting CPS 1 audio problems:

Turns out Capcom did the same stupid thing as SNK... The board is silkscreened for a Z80A and they put a cheaper, slower rated Z80 on it. That makes this component very suspect if you have audio problems.

However, that's not always the problem on the board.

Tools needed: Multimeter to check for bad traces and a logic probe (preferably one with audio) - bonus if you have an oscilloscope.

Symptom: No FM audio or no audio at all: (And no suicide battery on a QSound audio board)

First check the YM3012 DAC IC. If there is audio coming off of it check the analog part of the audio circuit. The audio part of the logic probe will help here. Look for bad op amps or capacitors. Check for a broken volume pot. If there's no audio coming off of the YM3012 DAC then check the digital portion of the audio circuitry.

To troubleshoot the digital section start by checking the EPROM in socket 9 - the Audio EPROM. Are there any signals on the data lines? No? Check the reset line on pin 26 as you power up the board. If it starts out low then flips to high (after a small pause) then the reset line is OK.

If the reset line is bad, trace it back and fix it.

Next check the clock line on pin 6 of the Z80. If it is bad, replace the 3.579 oscillator. If it is good, replace the CPU.

If the data lines and address lines show signals then check pins 2, 5, and 6 on the YM2151. These are the IRQ*, RD*, and WR* signals. If they are stuck high or low then replace the CPU.

If replacing the CPU at this point doesn't solve the issue try replacing the SRAM IC. (2k x 8 - a 6116 SRAM IC can be used to replace it)

If there is activity on those lines of the YM2151 then check pin 21 of it. It will have a steady pulse if no audio is coming out (Serial Data out). When audio is coming out there will be a series of scratchy pulses on it - this is more evident if your logic probe has audio. If there is no audio pulses coming out on pin 21 then replace the YM2151.

Your FM audio should now be working.

The non-FM sound effects (voice, punching, and other sounds) are simply controlled by the OKI M6295 IC. Check EPROMs 18 and 19 as the digital code for the sound effects is stored there. Repair any bad traces or replace any bad EPROMs. Double check to make sure the jumpers for the EPROMs are set correctly for the EPROM type used. (There's another post in this thread on how to set the jumpers)

I've pinned out one of the connectors far enough to use in troubleshooting audio.

CN1 (Sound)

A1 - VCC
B1 - GND
A2 - VCC
B2 - GND
A3 - VCC
B3 - GND
A4 - EPROM 9 CE*
B4 - A16           -------
A5 - A15             ^
B5 - A14             |
A6 - A13             |
B6 - A12
A7 - A11
B7 - A10
A8 - A9
B8 - A8
A9 - A7
B9 - A6
A10 - A5          Z80 to EPROM 9
B10 - A4
A11 - A3
B11 - A2
A12 - A1
B12 - A0
A13 - D7
B13 - D6
A14 - D5
B14 - D4
A15 - D3             |
B15 - D2             |
A16 - D1             |
B16 - D0          -------
A17 - EPROM 18 CE*	Pin 35 (OKI M6295)
B17 - EPROM 19 CE*	8E pin 6 (74LS04)
A18 - A16	Pin 34        -------
B18 - A15	Pin 33          ^
A19 - A14	Pin 32          |
B19 - A13	Pin 31          |
A20 - A12	Pin 30
B20 - A11	Pin 29
A21 - A10	Pin 28
B21 - A9	Pin 27
A22 - A8	Pin 26
B22 - A7	Pin 25
A23 - A6	Pin 24
B23 - A5	Pin 23
A24 - A4	Pin 22        OKI M6295 to EPROM 18/19
B24 - A3	Pin 21
A25 - A2	Pin 20
B25 - A1	Pin 19
A26 - A0	Pin 18
B26 - D7	Pin 16
A27 - D6	Pin 15
B27 - D5	Pin 14
A28 - D4	Pin 13
B28 - D3	Pin 12          |
A29 - D2	Pin 11          |
B29 - D1	Pin 10          |
A30 - D0	Pin 9         -------
B30 - NC
A31 - NC
B31 - NC
A32 - NC
B32 - NC

Here's another CPS1 A to B board connector pinned out.


CN4 (Char ROM Data/Select?)

G1 - GND
H1 - GND
G2 - GND
H2 - GND
G3 - GND
H3 - 1A PIN 2      -------      PIN 98   -------
G4 - 1A PIN 3         ^         PIN 97      ^
H4 - 1A PIN 4         |         PIN 96      |
G5 - 1A PIN 5         |         PIN 95      |
H5 - 1A PIN 6      PAL16L8      PIN 94
G6 - 1A PIN 7      IC 1A        PIN 93
H6 - 1A PIN 8      B Board      PIN 92
G7 - 1A PIN 9         |         PIN 91
H7 - 1A PIN 11         |        PIN 90
G8 - 1A PIN 13         |        PIN 89
H8 - 1A PIN 15      -------     PIN 88
G9 - 1B PIN 5      -------      PIN 87
H9 - 1B PIN 4      74LS245      PIN 86   CPS-A-01
G10 - 1B PIN 3       B Board    PIN 85   DL-0311-1001
H10 - 1B PIN 2       -------    PIN 84   IC 2H
G11 - 1C PIN 2       -------    PIN 83   A Board
H11 - 1C PIN 3          ^       PIN 82
G12 - 1C PIN 4          |       PIN 80
H12 - 1C PIN 5       74LS245    PIN 79
G13 - 1C PIN 6       B Board    PIN 78
H13 - 1C PIN 7          |       PIN 77
G14 - 1C PIN 8          |       PIN 76
H14 - 1C PIN 9       -------    PIN 75
G15 - NC                        PIN 74      |
H15 - NC                        PIN 73      |
G16 - NC                        PIN 72      |
H16 - NC                        PIN 71   -------
G17 - PIN 119       C Board      PIN 5 IC 8F A Board
H17 - 8E PIN 8        74LS08      JAMMA PIN P (Video Sync)
G18 - GND 
H18 - GND 
G19 - GND 
H19 - PIN 104       -------      PIN 11 IC 4C A Board
G20 - PIN 102          ^         R/W*   -------
H20 - PIN 101          |         OE*      ^
G21 - PIN 100          |         A11      |
H21 - PIN 99                     A10      |
G22 - PIN 98       Custom        A9   MB81C78A
H22 - PIN 97       Capcom        A8   IC 1C & 3C
G23 - PIN 96       IC            A7   A Board
H23 - PIN 95       C Board       A6   (Color SRAM?)
G24 - PIN 94                     A5
H24 - PIN 93                     A4
G25 - PIN 92          |          A3      |
H25 - PIN 91          |          A2      |
G26 - PIN 90          |          A1      |
H26 - PIN 89       -------       A0   -------
G27 - VCC 
H27 - VCC
G28 - VCC
H28 - VCC
G29 - VCC
H29 - VCC
G30 - Pin 108      C Board      PIN 6 IC 1F A Board (Video Clk?)
H30 - VCC
G31 - VCC
H31 - VCC
G32 - VCC
H32 - VCC

Tip: Troubleshooting Sync Problems

If the CPS1 board works, but has no sync then this will apply. This will not apply if the game is not otherwise working OK.

The sync pulses are generated by the custom chip on the C board. Pins 118 and 116 on the C board carry the signal to pins 13 and 12 of the 74LS08 on the B board, directly under the C board. This chip ANDs the signals and the output pin, pin 11, connects to resistor R28 (on the smaller SMT style A board), then on to the sync pin of the JAMMA connector.

If the board is missing sync, first check for damage on R28 (220 ohm) and continuity to pin 11 of the 74LS08 on the B board. If those are good then check for pulses on pins 116 and 118 of the C board. If the pulses are there on BOTH pins and missing from pin 11 of the 74LS08, then replace the 74LS08. If one or more pulses are missing and the pins stuck low on 116/118 of the C board then reflow the custom chip on the C board. If that fails to fix the sync problem then replace the faulty C board.

Tip: ROM speeds

When reprogramming EPROMs to run the phoenix version of code on the CPS-1 boards, the EPROMs on the board may need replacing. The AMD EPROMs labeled -155 (155ns) and -175 (175ns) are just fast enough to run the regular code. When running the phoenix code on these EPROMs the game may crash randomly.

When using AMD EPROMs, use 120ns or faster for EPROMs running the phoenix code. Intel EPROMs rated at -15 (150ns) will run fine. Any AMD EPROMs that do not need reprogramming are fine to keep using.

Tip: Modifying C boards

Captain Commando uses the 90631C-5 labeled C board while some versions of Street Fighter II uses the 92631C-6 and they both have the CPS-B-21 chip on them. This version of SFII is easy to identify as it has the blank spot for the suicide battery on it.

To modify the SFII board cut the ground trace going to pin 42 and 44 of the CPS-B-21 chip. Pins 45 and 46 are already tied high so they do not need to be modified as on the Capt Commando C board. Install 1K SMT resistors at locations R1 and R2. These can be pulled from the bad C board and transferred over.

Ensure that there are no shorts between +5v and ground then install the board and test the game!

Repairer: Womble
Forum Thread: CPS1 A PCB Repair

Jatz posted me his CPS1 A board after finding his Street Fighter producing horrible noises, or silence, or both in quick succession. Being CPS1 I had assumed it was going to be the old bad Z80 fault which could have explained the fault, and I was wrong. What arrived was the older style CPS1 board, its longer than the more common one, has far less SMD TTL chippery and has a different brand Z80 installed.

Pcb repair cps1 a board 1.jpg

The couple of dozen or so Z80s I have removed from CPS1 boards over the past couple of years are all Zilog branded chips, this board had a Goldstar one and it was fine.

I was fairly sure I spotted the fault as soon as I unpacked it, for some reason this board had never had a heatsink installed on the main amplifier chip, and it was standing bolt upright. Even most bootlegs do it better than this and this was an original branded Capcom board. Even without a heatsink the amp chip should lie flat on the board and its legs should have a 90 degree bend in them to go through the board. On this board it was standing bolt upright...

Pcb repair cps1 a board 2.jpg

... how this survived twenty years like this I will never know.

The board was designed for a heatsink, I have one of this type of CPS1 board and it has a proper heatsink, bolted to the board and to the chip, but on this board the bolt holes were still filled with the original solder from when the board went through the flow soldering machine all those years ago. I think this one fell off the production line in the early 1990s and was overlooked by the heatsink guy.

An amp without a heatsink is going to get hot, too hot perhaps but in this case what killed it was the lack of structural support. Two decades of wobbling had fractured virtually every leg on the chip...

Pcb repair cps1 a board 3.jpg

...the ones that were intact had major metal fatigue resulting in an ever more wobbly chip. Fairly certain this was the fault I powered it up and tried to press down on the chip to see if I could prove the amp chip or not. I got silence, so I waggled the chip a bit more to finally get it off the board.

A quick rummage through my scrap pile produced a long forgotten CPS1 A board, the same layout but this board was a wreck so much rust and track damage it is never going to be repaired. So the amp and heatsink were going spare...

Pcb repair cps1 a board 4.jpg

... a quick scoot round with the desoldering iron on Jatz's board...

Pcb repair cps1 a board 5.jpg

.... and the amp and heatsink were re-homed...

Pcb repair cps1 a board 6.jpg

... the trick is to fit the amp to the heatsink (in this case it already was), insert the chip legs into the board and then bolt the heatsink on. Only then do you solder the chip legs in, this ensures that the legs are under zero stress. If you solder it in and then bolt the heatsink on it is likely you will have to bend something and you risk ripping legs out or crushing them into the board as you do up the heatsink bolts.

I also did a quick ESR test on all the caps, all but three 10uF 35V ones were fine, that three were 4x the ESR they should be so were probably on the way out, so they have been replaced.

When powered up under my Magic Sword B board all sounds were back, loud and clear.

Repair Logs converted to wiki format by Brad from Aussie Arcade.