Image Fight repair log

Received from Japan an Image Fight PCB for repair (game is a vertical shooter  developed and published by Irem on M72 hardware)

On power up the board seemed to properly boot, something moved on screen but then got stuck on garbage :

As many of  you may know, this is a three stack set hence, having a good R-Type boardset, I started to swap boards.I was able to narrow the fault in the bottom PCB :

 

First of all I probed these four 8k x 8-bit static RAMs @IC34-IC35-IC36-IC37 :

I found that their WRITE ENABLE input (pin 27) was stuck high, no activity/pulsing like observed on my good R-Type :

Tracing the signals back I could not find anything abnormal until I came across a 74LS32 @IC51 :

 The logic analyzer showed no LOW pulsing on output pin 11 when inputs 12 and 13 got asserted :

I piggybacked the device and board booted into game although background graphics were corrupted :

 

I removed the TTL device and it failed the out-of-circuit testing :

Time to install some machined round socket and a good IC :

The next power up gave me a fully working game with no other issue.Repair accomplished.

1943 Kai - Midway Kaisen repair log

Received from Austria for a 1943 Kai - Midway Kaisen PCB for repair (the game is an update of 1943 with tweaked weapons, fewer levels, and a new soundtrack released in Japan only).Set was in good shape in both CPU and VIDEO board :

 

Game was playable but with severe sprites issue :

 

Objects are generated on bottom video board by a custom IC (in PLCC84 package) marked '86S105'

This is another well known unreliable and prone to failure part (which really would need to be reproduced someday) which I had to replace many times hence I was pretty sure it was bad also on this board.I didn't have any spare but the owner of the board came to help and sent me a working one taken from a Pang PCB.The part arrived some days later :

Time to remove the old one and install the spare :

This lead to a great improvement, sprites looked much better but yet not perfect as they were still slightly glitched with vertical lines through them:

Schematics show that on VIDEO board there are two identical sprite line buffer circuits where two 2K x 8-bit static RAMs lie :

On my board they were replaced :

During removal of the RAMs some traces on solder side were broken and then repaired with some wire .

I probed the two RAMs and found that pin 16 (data line D6) of the one @7D was stuck HIGH :

According to schematics this pin should be tied to pin 17 of the 74LS273 @6E and pin 9 of the 74LS257 @5D :

I checked on the board and found that connection to pin 9 of the 74LS257 @7D was missing.I promptly run a jumper wire between the two points:

This made the trick and fixed board completely.Another successful repair.

Pac-Land repair log

Got for repair this original Pac-Land PCB (manufactured by Namco and licensed to SIPEM) :

The board was in very good shape but it booted all the time to a static screen :

Probing the program ROMs revealed the /CE line of two devices were stuck HIGH or silent :

The signals come from a 74LS139 @7D :

Being a Fujitsu TTL I was aware of the way they usually fail so I didn't spend much time on it.I removed the device, it miserably failed the out-of-circuit testing :

Replacing it allowed the board to boot, game was fully playable but backgound graphics were all blocky :

The issue was obviously located in the tilemap generation and, judging from kind of fault, more precisely in the addressing circuit.Schematics show there are a couple of 74LS86 @9J and 10P that generate some addresses to the backgrounds ROMs :

Devices were again from Fujitsu, they gave troubles when tested in circuit with my HP10529A logic comparator :

They both failed the out-of-circuit testing :

I replaced the two TTLs and powered the board up again sure enough it was fixed and fully working but playing a game I noticed sprites were glitched :

The sprites circuit consists in a couple of custom ICs ('CUS11' and 'CUS12'), four EPROM devices and some logics where I spotted another Fujitsu TTL, a 74LS377 @7K :

The scope showed output pin 9 was floating despite an active input :

Device failed the out-of-circuit testing in that exact gate (pin 9 was, indeed, in 'Z' or high-impedance state)

The replacement finally cured the board.Repair accomplished.

Konami '504' reproduction

This is my own reproduction of the Konami ‘504’ custom IC with the use of simple logic gates and no CPLD or other moderm programmable logics (call it 'poor man's' if you want...).Original part is a 28 pin 600mil IC found on some Konami  PCBs of 80’s , you can refer to this useful spreadsheet (credits to ‘mattosborn’ on KLOV forums) :

 Konami custom ICs


How often this manufacturer has accustomed us, the IC comes with scratched-off part name:

This time I have been facilitated in my work because Konami itself sometimes replaced their custom ICs with daughter-boards which are faithful TTL implementation of the functions.I was able to get the one made for the '504' : 

Then it was only matter to take schematics of this daughter-board and route them to a PCB layout.The final result is a little board with more or less same dimensions of original part (note the simultaneous use of SOIC and TSSOP devices)

 Reproduction was succefully tested on a Track & Field PCB :

Darwin 4078 repair log

Received from Austria this rare original Darwin 4078 PCB (there are only bootlegs around)

Owner told me that he bought the board as working but then it failed after few minutes of gameplay.When I applied power to the board I could not get almost anything but a dimmed rolling picture that sometimes turned in the title screen :

 

All the video timing signals are generated on bottom board :

They are derived from a 12MHz crystal :

The master clock is divided and then inverted several times by a couple of 74LS04.Testing the one @12H (from Fujitsu...) with a logic comparator revealed bad outputs : 

The scope confirmed the unhealthy output signals (on right on the below picture, inputs on left) :

The IC failed the out-of-circuit testing :

Replacing the IC allowed the board to boot up but there were still some loss of video:

 Sometimes it showed only a fuzzy screen :

 I decided to test in circuit with a logic comparator all other Fujitsu TTL.I found a bad 74LS08 @10H :

 And a 74LS04 @10A with a floating output :

It obviously failed when tested out-of-circuit:

After replaced the ICs I tested again the board,it was fully working.Repair accomplished.

Seibu 'SEI0010BU' reproduction

Another rerproduction of a custom IC used on Seibu/Tad Corporation hardware like the past 'SEI0060BU'.This time I took care of the 'SEI0010BU' which comes in a 600mil 42 pin SDIP package like the other Seibu gate arrays used for graphics  :

As always I studied how the functions of this custom IC were reversed on bootleg boards.In this way I figured out that the 'SEI0010BU' is basically a big shift register hence it could be reproduced with almost same design of other custom ICs with similar functions I already reproduced (like the Capcom '86S100' for example).In this case the design was not really complex hence the resulting board layout was of same dimensions of original part :

I could successfully test the replacement in a Cabal and Toki PCBs :

Another custom IC preserved, long live to arcade!

Violent Storm repair log

Got recently for repair this Violent Storm PCB :

Board was in good shape but completely dead, it gave nothing on screen when powered up.This Konami hardware uses a couple of custom ASICs (marked '055673' and '055555') with very fine pitch whose pins most of time lose contact from pads preventing the board from booting.Sure enough I inspected the '055555' for first:

  I found on it amost a whole row of lifted pins :

Promtly reflowed :

This made the trick!

Board risen from its grave and fully working again.Repair accomplished.