PCB Repair LogsComments Off on Demon’s World/Horror Story repair log
Jan282017
Got this Demon’s World PCB (also known as Horror Story in Japan but game is multi-regional since it can be configured for different regions via the DIP switches) for a repair :
Board was not working, upon boot it showed the typical black/white stripes (common to all Toaplan boards) sign that main 68000 CPU has properly been initialized but then sat on a black static screen :
As said the CPU was properly running on power up but then halted before entering in game, data/address busses were both silent, usally this happens when CPU can’t find addressable devices.So I went to piggyback RAMs on board.When I did it on four 62256 @18F, 19F, 21F, 23F :
I got this error:
I removed the four chips, most of times they resulted good when tested out-of-circuit but sometimes they randomly failed due probably weak cells.Infact when I installed new RAMs I was greeted by this screen:
Board perfectly working, no further issue found.Another repair log archived.
Probing it revealed static address/data bus, clock was present instead.Since it was not included in the above pinout I spent some time to figure out that pin 59 was the /RESET one.Probing it I found that on power up there was no transition from LOW to HIGH but pin went directly to HIGH state hence CPU didn’t get initialized.RESET circuit is a typical one made of the PST518 voltage monitor IC, a 47uF 50V electrolytic capacitor and a couple of resistors :
Checking in-circuit the ESR of this 47uF 50V capacitor @IC56 gave no reading, value was out of the 100Ohm range of my meter:
I removed it and out-of-circuit testing confirmed this capacitor was bad:
Replacing it restored a proper RESET signal so board successfully booted:
But sound was quiet despite adjusting the potentiometer and some interferences were present on screen too:
I went to measure the impedance of the electrolytic capacitors in analog sound circuit and I found many of them with out-of-specs value:
I replaced them all :
This restored a loud and a clear sound.As said in the beginning, it was only matter of capacitors!
Received this Wild West C.O.W. Boys of Moo Mesa PCB for repair:
PCB was very dirty with flux residual and rust so I first washed it.After dried I powered it up and was greeted by this screen:
The initial self-test reported a bad ROM @T5 and a bad RAM @J4, obviously the board didn’t boot up but kept resetting.Usually Konami testing routines are very reliable so I went to dump the four program ROMs and actually the one @T5 showed at offset 0x3FFB4 a difference of only one byte (well, one nibble..) compared to the MAME dump:
I reprogrammed the EPROM (a 27C020) and error was cleared but obviously the one regarding the RAM @J4 was still present:
RAM @J4 was one of the three RGB 2K x 8-bit palette RAMs :
Piggybacking it had not effect so, before removing the chip in order to test it out-of-circuit, I checked its connections with a multimeter.All was fine until I could measure a short between pin 21 (/WE) of the RAM and VCC :
This explained why the RAM was reported as bad as it was never be written (write enable signal is active low).The /WE signals of the other two RAMs are generated by the near ‘054338’ custom :
so I went to inspect it with a digital microscope and found two bridged pins:
I removed the bridge and board successfully booted but sound was totally missing:
This is a common issue on these Konami board that use the ‘infamous’ hybrid audio module.This board has the ‘054986A’ one :
Usually you can fix this issue by replacing the capacitors but I opted for a drastic solution removing the module and installing two rounded machine-tooled strips with 1.778mm of pitch:
PCB Repair LogsComments Off on Namco System 1 CPU board repair log #2
Jan072017
Another Namco System 1 CPU board on the bench after the past one :
Board (coupled with a good ROM board) gave only a black screen with some green lines, watchdog was active :
I checked the three colors outputs and only the green was active, the other two were dead:
We saw in the past repair that there are three 6264 static RAMs, each one for a color, they are addressed by the custom ‘C120’ and data are processed by the custom ‘C116’ which generates the different shades (then formed into a single color and converted to analog signals from some resistor arrays)
Probing the RAMs revealed stuck signals on data pins of the ones @B2 (RED related) and E2 (BLUE).I removed them :
But during desoldering I decided to remove also the GREEN one @D2.They all failed when tested out-of-circuit:
But also with good RAMs I got no improvement, watchdog was no longer active but all three colors still dead.Probing the other RAMs on board, I found no data coming out the two 6264 @D6 and E6 despite address lines were active:
When tested both out-of-circuit actually only the one @E6 failed:
Finally the board booted but with missing sprites and music (ony FXs were present).Here’s a video using a World Stadium ’89 ROM board:
I noticed that if I piggybacked the 6264 RAM @L3 ( used by the custom ’30’ which controls the sound), the rustling sound I had instead of the music disappeared:
So I removed the RAM and it failed the test:
This restored full sound.Now the sprites issue.I noticed that, when using this CPU board with a Galaga88 ROM board, the sprites (that were competely missing when using the World Stadium ’89 one) were visible but garbled and blocky:
The two customs related to objects (the sprites address generator ’48’ and the sprites generator ’39’ ) were good but probing the two 74LS377 @A3 and A4 revealed stuck outputs:
I promptly removed them and they both failed when tested out-of-circuit:
Replacing them restored the sprites and fixed completely the board.Five RAMs and two TTL replaced to achieve this, a nice booty…
Received this faulty Konami Aliens PCB for repair :
Board was in poor condition, dirty, custom ASICs were full of flux residual, four GFX 4Mbit MASK ROMs were removed and socketed most likey using improperly an heat gun judging from bubbles on solderside of the PCB:
For first I washed the board.Once dried, I powered it up and I was greeted by this repeating screen :
This should be the self-test report but graphics was all messed up sign that there was some problem in the tiles generation (the red characthers most likey referred to some RAMs reported as bad).Therefore I went to inspected the reworked area on solderside and found that a couple of rivets were ripped off during the desoldering operation, these should have tied to VCC since they were pin 31 (/BYTE) of two 4Mbit MASK ROMs
I checked all the connections on the reworked area against schematics :
I found some poor contact between the two tile generators ‘052109’ – ‘051962’ and MASK ROMs/ RAMs
This allowed the board to succesfully boot but some tiles had bad colors:
MASK ROMs test reported three bad devices:
Since they were socketed I dumped them and compared against MAME.They resulted good but I realized they were put in wrong sockets!I swapped them and MASK ROM test succesfully passed :
No other issue was found so also this board was declared as 100% working.
