Another TAITO PCB on the bench.This time we have a Rainbow Islands one:
The first thing I noticed was (dust apart..) the lack of the TC0070RGB module.For the uninitiated this custom in SIL package is a DAC so it converts digital RGB output of the PCB into analog signals.I borrowed this custom from my Bonze Adventure PCB and installed the needed sockets as always.Once powered on the board, I was greeted by a solid black screen so I started investigating.Probing the main 68000 CPU I found there was no clock at all.Main clock is generated by a 16MHz oscillator @X1 and then divided by a near 74LS161 counter.I visually inspected this part of circuit and found this:
Replaced the oscillator gave me a fully working board but then it was time to find a proper replacement for the borrowed TC0070RGB.I knew this custom was reverse-engineered by Macro (thanks to him) some time ago:
Another board donated from my arcade operators friends.It’s time for Crime City.
For the uninitiated, Crime City is a beat ’em up developed by TAITO that can be played by one or two players simultaneous.Here are the technical specs of the hardware:
Hardware – Taito System B Hardware
Board Number – K1100470A
Main CPU – 68000 (@ 12 Mhz), Z80 (@ 4 Mhz)
Sound – YM2610 (@ 8 Mhz)
Screen Orientation – Horizontal
Resolution – 320 x 224
Refresh Rate – 60 Hz
Palette Colors- 4096
Players – 2
Buttons – 2
Joystick – 8 way
The board booted fine but I immediately noticed that there was no sound at all except for some scratching noises.Besides, some colours were wrong and some characters tended to pink compared to the correct ones from MAME (see wheels of car in the picture below, MAME screenshot is on the right) :
As usual I started my troubleshooting with a visual inspection of the PCB and found a broken track on the solderside which I promptly patched with some AWG30 wire:
This fix restored correct colours (the broken track was actually connected to the TAITO ASIC TC0260DAR which is a custom pallette generator and this explained the reason of the fault)
As I said board was silent, no sound at all except for some noise sign that the main MB3735 amplifier was working fine.Digital audio section is made by a Z80 CPU which commands a sound synthesis YM2610 chip.With my logic probe I could verify that all was working fine there so fault was elsewhere.So I started to suspect the YM3016F sound DAC:
Connecting its analog output to an external amplifier produced no sound (while digital input was present).This convinced me to replace it and so sound was fully restored .
First of all I want to thank JROK for providing me all the technical info and allowing its public release, without his help I would not have ever made.Let’s start with the story.
Some time ago a friend of mine sent me his Splatterhouse PCB bought as working saying it got stuck upon boot on ‘ROM TEST START!! PLEASE WAIT..’ message displayed upside down on a screen filled with red zero.So I started to investigate on the net and found that usually this kind of issues are caused by a faulty custom chip marked ’64A1′ @M4 on ROM board.This was confirmed by swapping this IC in another CPU board of Galaga88 with this missing component , in this case I got an ‘EEPROM ERROR’ message which is a further proof of the fault.
Thanks always to JROK we know that this ’64A1′ custom IC is nothing else than a HD63701 MCU with the exception of two custom opcodes not present in original part.You can go deeper into technical reading here:
So, programming a blank HD63701 MCU with the available dump from MAME won’t work due the presence of these two custom opcodes.So, patching a ROM is needed.
In the specific you have to change three bytes in the ROM “VOICE0” at a specific offset.For example : in Splatterhouse the offset is $FA8B and the new HEX values are “20 01 56”.
The string to search for in HEX is “B7 C0 00 6E 00” and the first 3 bytes must be changed to “20 01 56”
The patch seems to be needed at offset $FA8B for most games, but not all. Here’s a reference table :
Tank Force – $FA29
Blazer – $FA8B
Dragon Spirit – $FA8B
Galaga 88 – $FA8B
Marchen Maze – $FA8B
Splatterhouse -$ FA8B
Rompers – $FA8B
Blast Off – $FA8B
World Court – $FA8B
World Stadium (and WS ’89, WS ’90) – $FA8B
Bakutotsu Kijuutei (Baraduke 2) – $FA8B
Beraboh Man – $FA4E
Dangerous Seed – $FA8B
Face Off – $FA29
Pisto Daimyo no Bouken – $FA8B
Puzzle Club – no voice ROM !
Quester – $FA1B
Souko Ban Deluxe / Boxy Boy – $FA8B
Yokai Douchuuki (Shadowland) – $FA8B
So, it was only matter to find some real blank HD63701.My first attempt failed since I bought as new some ICs that they turned out to be already programmed and, since they were the plastic HD637B01VOP one-time-progfammable version, they were unusable for my purpose.But then I was luck and found some HD63701VOC UV erasable ones:
So, it was only matter to program one of these MCU with MAME dump using my Hi-Lo Systems ALL-11C EPROM programmer and, at the same time, replace the ROM ‘VOICE0’ with the properly patched one.Once done, I had confirmation of what the good JROK said about. Operation ‘NAMCO custom 64A1 replacement’ accomplished.
First of all I wanna thank the people who donated me this (and other) PCB : “mille grazie” to Mr. Giuseppe (a long-standing arcade operator of my town) and his nice daughters Elena and Antonella!
Now, let’s begin with the log.
When I got this PCB in my hands, I immediately tought it was a Rainbow Islands one:
But I was wrong, it actually was the Extra version (different ROMs and C-Chip).Before start the troubleshooting, I resoldered a wonky TC0070RGB module:
Once powered on, I got a solid white screen with a clicking sound indicating that the watchdog circuit was active:
There was clearly some trouble in main code execution so I started to test with my logic probe the two WORK RAMs and I found some data lines stuck HIGH on the 6264 SRAM @IC25.I desoldered the chip and out-of-circuit test confirmed it as bad:
With a new RAM the board succesfully booted but graphics (tiles and sprites) were all blocky:
This board (and other Taito ones on similar hardware ) uses two custom chips in PGA package which generates sprites (the one marked ‘PC0900J’) and tiles (the ‘PC080SN’).They respectively address GFX ROMs through a couple of 74LS373 latches.Probing the 74LS373 @IC29 I found that all the outputs were absent.So I piggybacked a good one and tiles were restored but sprites were always blocky:
So, I went through the 74LS373 @IC18 and found it had output PIN19 stucked LOW while its input was toggling.Piggybacked this and also sprites were restored.Desoldered the two 74LS373 and they were indeed bad:
Replacing them both back this game to its former glory!
Got this original Konami G.I. Joe PCB from Ebay as not working.Looking at the auction picture I could see the ‘051550’ SIL custom broken in half, this was confirmed once I received the board:
For the uninitiated, this custom is of vital importance since it generates the master RESET for the whole board, infact 68000 CPU RESET line was stucked LOW and I got only a static screen:
I had many Konami faulty boards for spare so I took this custom from one of them and, after replaced it, the board resetted properly but with an error on VRAM @16K:
Desoldered and test the 6264 RAM out-of-circuit confirmed it as bad.With a new RAM game was fully playable (also with sound which is quite rare as this board uses the ‘054986A’ hybrid custom module) but something was wrong :
like some colours were missing from certain backgrounds and this was confirmed also by the color check I ran in TEST MODE comparing the output with the one from a working board:
Color RAM was OK as reported by initial test so I start to suspect the ‘053251’ ASIC since I had a similar fault on a Bell & Whitles PCB.According to MAME this ASIC is a palette/priority chip :
So this convinced me to replace it.And I was right since colours came back to normality:
Gameplay was fine now but some of the sound FXs were muffled and distorted.It could not be due the ‘054986A’ hybrid module since the issue was limited only to some FXs so I started to check the audio circuit (commanded by the usual Z80 CPU).When I piggybacked the 6116 SRAM @6C all the sound FXs were restored so I desoldered the chip in order to test it out-of-circuit:
RAM was really bad though this fault was not reported by initial RAM/ROM test.Not satisfied, finally, to top it all I recapped (with tantalum capacitors) and socketed the ‘054986A’ sound hybrid module’:
