May 212017

Another board from the recent operator raid, an original Michael Jackson’s MoonWalker (on Sega System 18 hardware)

Despite its age and the use of the FD1094 CPU module with battery backed-up RAM the board was still working except for the sound samples (drums, speeches, etc..), they were missing or replaced by random ones:

The hardware uses a rebranded Ricoh RF5C68 as PCM sound chip :

Not able to find any datasheet I went to “listen” its pins with an audio probe.When I hit the analog outputs (the IC has an on-board DAC) I realized that sound came already wrong out of it.So the chip was the only responsible and it needed to be replaced:

Done and…success!Board completely fixed.

 Posted by at 9:50 pm
May 212017

Another Konami PCB on the bench!

Picked-up this Asterix PCB in a recent operatot raid:

Board was physically damaged, one part of PCB was litterally missing and the nearby 16Mbit MASK ROM was cracked in half:

The affected part was not a vital section of circuit (but only a spare room for not used additional MASK ROMs) so board still booted but sprites were missing.If I pressed the ‘053245’ ASIC they came back but garbled and blocky:

For first I reflowed some bent pins of the ‘053245’ sprites generator:

Then I replaced the broken 18Mbit MAS ROM @3K with a compatible 27C160 :

This improved things but not still perfect :

The MASK ROMs check complained about the other device containing sprites data (the one @7K)

Like said in my previous Parodius DA! repar log, this doesn’t necessarily imply that IC is faulty but also that it cannot be reached  by the device which wants to read/write it.Address bus is shared among the two MASK ROMs @3K and 7K, doing a check with my muktimeter all was fine except for pin 42 of both, I got no continuity between them :

Pin 42 is the higher address line:

I simply ran a jumper wire, this restored sprites and fixed completely the board:

 Posted by at 7:05 pm
May 212017

Received this faulty Parodius DA! PCB for a repair:

The fault concerned the graphics, there were jailbars all over the screen :

Clearly the tilemap generation circuitry was involved, this is made of two 4Mbit MASK ROMs and two ASICs that work in parallel : the ‘052019’ generates the address lines and the ‘051962’ receives the data from the MASK ROMs and process them :

The MASK ROMs check reported the two devices @K13 and K19 (containin tiles data) as bad:

The MASK ROMs @K19 was already removed and socketed.I dumped the device and my programmer warned about a poor contact of pin 13 (data line D0)

Replacing it with a 27C400 produced very little improvements, the test still reported the two bad MASK ROMs.This meant the main CPU could not reach the two devices (thru the ‘052019’ ASIC) so I went to do a continuity check.All the address lines should be daisy-chained between the two 4Mbit MASK ROM and the ASIC but I found some broken/bad contacts from the one @K19, whoever replaced the device managed to break some connections.Here the solderside overview:

Here are some close-up  (taken with a microscope) of the invoved pins (pin 1-5-13)

I restored the connections but still got jailbars:

Chcking the data bus of the tiles MASK ROMs I found a bad via which should have tied pin 29 (data line D19) 0f the device @K19 to the near ‘051962’ ASIC (which processes the tilemap data)

Patching it restored completely the graphics.Job done.

 Posted by at 11:22 am
May 122017

In the past days I dumped the only PAL from a Bells & Whistles PCB (export version of Detana! Twin Bee by Konami).Original device was a secured PAL20L10 which I reversed and tested successfully  into a GAL22V1o.

 Posted by at 6:25 pm
May 032017

Today I received this Wild West C.O.W. Boys of Moo Mesa PCB for repair:

PCB was faulty, initial RAM/ROM check  failed showing a ‘SOUND SYSTEM BAD’ error and then resetted  :

Sound circuitry was already reworked: two RAMs, the Z80, the YM2151 and the ‘054539’ ASIC were replaced, the ‘054986A’ module was socketed too.But this was not enough to locate the fault :

The sound system reported as bad is ruled by a Z80 CPU, probing it revealed its pin 16 (/INT) was asserted, this meant the CPU was undergone an interrupt request stopping execution of code.It was time to use my Fluke 9010A.When I performed a BUS test, it reported address bit 15 (pin 5 of Z80) tied LOW:

This doesn’t necessarily implies the address line is shorted to GROUND but also that something is forcing it low.But in my case it was really shorted (only few Ohms of resistance to GROUND measured)

Doing some continuity test on a good board I could figured out that pin 5 of the Z80 is connected only to pin 5 of the PAL ‘054744’ @E7 and pin 1 of the 74ALS157 @C7 so in my faulty board one of these IC was bad :

For first I intentionally cut the trace going to pin 1 of the 74ALS157 @C7:

This actually cleared the short to GROUND of pin 5 of the Z80, hence the TTL was internally shorted.Tested out-of-circuit it obviously failed:

Now board could pass the RAM/ROM check and enter in game:

But sound was very faint, barely audible.I was forced to set volume to maximum to hear something:

The ‘054986A’ module was already recapped with thru-hole electrolytic capacitors:

A faint sound meant the main amplifier was working but something was wrong in pre-amplification performed by the 4558 OP-AMP on the underneath of the module.So I went straight and replaced it with a compatible LM358:

Sound was back loud and clear!Job done.

 Posted by at 8:52 pm