Jan 182019

Received from Germany this Gundhara PCB, a top-down shooting game released in 1995 by Banpresto:

According to the owner the PCB suddenly developed a graphical issue.When I powered it up I could actually see some colors problem, the BLUE one was ‘bleeding’ in some parts of graphics:

After a quick check on PCB I could figure out the palette circuit made of two 6264 (8K x 8-bit) SRAMs, a custom (S-DIP 42 pin) marked ‘X1-007’ (which probably contains some latches or flip-flops) and three SIL resistor arrays configured as R-2R ladder to perform the final conversion from digital to analog (DAC)

Color palette circuit

When I probed the two 6264 SRAMs I found weak signals on some data lines of the one @U39, here is a capture from scope comparing the signal with an heathy one :

Good signal on left, bad on right

Sure enough I desoldered the chip, it indeed failed the out-of-circuit testing :

Installing a good chip fixed the issue, board 100% working again.Job done.

 Posted by at 9:03 pm
Jan 142019

Received today from Germany this Guardians PCB, also known as Denjinmakai II, a horizontally-scrolling beat-em-up released by Banpresto in 1995)

According to the owner the game played totally ‘bind’.I could confirm it as soon as I powered the board up :

Board playing ‘blind’

But, as you can see from above video, I was able to restore some graphics by pressing down this surface mounted custom ASIC (QFP 208 pins)

Judging from the heatsink on its top this custom ASIC acts as a GPU generating most of the graphics and video signals too.At a closer inspection I found many lifted pins on all four sides.An example :

I promptly reflowed the IC, this gave me a steady picture but all the graphics were scrambled and, judging from the blueish screen, the RED color was missing too :

Looking at solder side of the PCB I noticed a long and quite deep scratch passing across some traces:

Scratch on solder side

Inspecting the area with a microscope and doing a continuity check with my multimeter confirmed that some traces were severed and other caused unstable contact:

I patched and reinforced the traces with some AWG30 wire :

Traces repair

This lead to few improvements, graphics were still scrambled so I moved on to troubleshoot the lack of RED color.This was pretty easy to fix as the board had a quite clear damage that broke the trace which carries the signal to JAMMA edge connector :

Damage on PCB

As for bad graphics, I decided to double-check the previously reflowed custom ASIC and found a solder bridge shorting two pins:

Solder bridge on reflowed ASIC

Removing it fixed the issue and board completely.Repair accomplished.

Board finally fixed
 Posted by at 10:34 pm

64th Street – A Detective Story repair log #2

 PCB Repair Logs  Comments Off on 64th Street – A Detective Story repair log #2
Dec 212018

Got from Portugal this mint 64th Street – A Detective Story PCB, a beat ’em up released by Jaleco in 1991 :

The board simply booted to a static garbage screen:

At a deeper analysis both 68000 CPUs were not resetting properly on power up :

On the main one there was no transition from LOW to HIGH state, /RESET line went staight HIGH :

On the sound CPU both /HALT and /RESET lines were stuck LOW:

I traced back the lines to a typical power-on reset circuit based on the TL7705 voltage monitor IC which I replaced with no luck:

The circuit is made of few other components, specifically there are two 0.1uF mylar capacitors connected to pin 1 (REF) and pin 7 (SENSE)

I got few Ohms when I measured resistance across their terminals (after detached one leg from circuit), they were almost shorted :

I replaced them :

Board successfully booted into game with no further issue :

Repair accomplished.

 Posted by at 11:08 pm
Dec 192018

Got from Portugal this Out Zone PCB, actually a low-budget korean version :

At visual inspection the board lacked of two 6116 SRAMs (but sockets were installed)

After fitting them I powered the board up and I got the usual black/white wavy striped screen but no boot into game, 68000 main CPU got stalled right after :

Multiple could be the reasons, in particular this Toplan hardware requires the sound circuit to be running otherwise the whole system can’t be initialized.A Z80A rules the sound system :

Z80A audio CPU

When I did piggybacking on it I got a ‘SUB CPU ERROR’ message on screen :

Once removed the IC I had confirm that it was really bad.With a good Z80 the board successfully booted into game but sprites were all blocky (it seems this is a common issue of this hardware…)

Sprites data are stored in four 1Mbit 28 pin MASK ROMs :

Sprite MASK ROMs

Devices were soldered in on my board and not socketed like usual so I was forced to remove all of them for dumping.Two of them gave different CRCs at each reading :

Specifically the ones marked ‘ROM1’ @1B and ‘ROM4’ @1L :

Faulty 1Mbit 28 pin MASK ROMs

I replaced them with two 32 pin 1Mbit non-JEDEC EPROM devices.All seemed fine but while playing the game I noticed some sprites (like explosions) had lines through :

Looking at hardware I figured out that the sprite line buffer is made of four 6116 (2k x 8-bit) static RAMs:

 Probing them with a scope revealed some dead outputs of the one @12L :

On the right no activity on some data lines of 6116 SRAM @12L Good signals on the left for comparison

The chip failed the out-of-circuit testing:

Installing a good RAM chip finished the job.Repair accomplished.

 Posted by at 9:17 pm
Dec 022018

I have two Toki pcbs in storage, both with sprite errors. I saw a video from lukemorse1 (who actually inspired me to start repairing arcade pcbs) on youtube where he worked on a Toki pcb, also with sprite errors.

Here’s how the error on board #1 looked like:

So there are blocks surrounding the sprites.

From Luke’s video (https://youtu.be/Czdt_yTNRTs) I identified the area responsible for sprite handling and I quickly found a 74LS273 with a stuck output at pin 19:

Checking with the multimeter, I found that this pin was almost shorted to ground:

I quickly desoldered the IC and tested it with my VP-398

Replacing it with a working 74LS273 made the sprites appear as they should:


So a big thanks to lukemorse1 and also as always to caius for always helping out 🙂