X Multiply repair log #4

 PCB Repair Logs, Repair Logs  Comments Off on X Multiply repair log #4
Aug 162015
 

Got this Irem X-Multiply PCB (actually a conversion from an Hammerin’ Harry M81 board) from my friend Josef for a repair:

X-Multiply_PCB

It showed a graphical issue where some sprites (like enemy shots) were blocky during gameplay and backgrounds missing during attract mode:

issues

Sprites and tiles data are stored respectively in eight 1Mbit EPROMs on video board for a total of sixteen devices.

sprites_tiles_EPROMs

Original devices were 1Mbit 28 pin MASK ROM:

https://pdf.datasheetarchive.com/datasheetsmain/Datasheets-X2/DSA62700099.pdf

while who made this conversion used ST M27C1000 :

https://pdf.datasheetarchive.com/indexerfiles/Scans-065/DSA2IH00172244.pdf

Video board has some jumpers which must be closed according to the type of adopted devices :

jumper

In this case the jumper configuration was set to  ‘EP’ so not suitable for 27C1000 but required the use of JEDEC devices like 27C1001 or 27C010 (while the ‘MASK’ setting can work straight only with the original 28 pin 1Mbit MASK ROMs and not with 27C1000 which, although the similar pinout, have some extra pins to deal with, so some modifications are needed)

Reprogramming all the data for sprites and tiles into 27C010 1Mbit JEDEC EPROMs fixed graphics completely.

fixed_

 

 Posted by at 1:16 pm

Undercover Cops repair log #1

 PCB Repair Logs, Repair Logs  Comments Off on Undercover Cops repair log #1
Aug 152015
 

Got two Undercover Cops PCBs (a native and a Alpha Renewal version converted one):

Undercover Cops_PCBs

Both had sound issues.

The first one had all music but sound FXs were missing.In this kind of hardware (IREM M92) music is generated by YM2151 sound synthesis chip while audio samples from a “NANAO GA20” ASIC custom chip (QFP 80)

IREM_GA20

Probing the pins of the latter revealed no activity on its DATA/ADDRESS bus (it addresses and read data from the MASK ROM ‘W42’ on ROM board)

I dumped this device and it matched the MAME ROM set so problem could be elsewhere.Tracing back the connection from GA20  to the MASK ROM  I found that some address lines were not in continuity like the other.I could pinpoint the break in interconnect ‘CN3’ which carryes signals from/to the two device:

CN3_interconnect

A simple reflow was enough to restore the full sound.

The second board was totally silent, no audio at all.All the analog audio circuit was fine.I started to test ICs with my HP10529A logic comparator and when I put the clip on a 74LS74 @IC24 on CPU board, suddenly I got sound.So I inspected the solder side of this IC and found some evident dry joints , in particular pin 16 was not soldered on its pad, (probably I restored connection when I put the clip on it):

74LS74_dry_joints

Also here a quick reflow was enough to fix the board.

 

 Posted by at 7:37 pm
Aug 102015
 

Another board from my friend Alexander.It’s time for an original Gun. Smoke PCB:

Gun. Smoke_PCBJPG

Board booted fine but had a graphical fault where backgrounds were totally wrong (MAME screenshot on the right) :

comparison

In absence of schematics, MAME source is always a great help.Looking at it I was able to identify the ROM cointaining the backgrounds data:

ROM_REGION( 0x8000, “gfx4”, 0 ) /* background tilemaps */
ROM_LOAD( “gs14.11c”, 0x00000, 0x8000, CRC(0af4f7eb) SHA1(24a98fdeedeeaf1035b4af52d5a8dd5e47a5e62d) )

I dumped it and it matched the one in the MAME ROM set.So, problem had to lie in the external circuitry.Tracing back connection from this ROM I figured out the its address lines were connected to outputs of some 74LS283 whose inputs were tied to some 74LS273.So, I started to test them in-circuit using my HP10529A and BK Precision 560A.All was fine until my equipment reported a problem on the 74LS273 @8D:

74LS273@8D_in_circuit_

Piggybacking this cleared the issue:

Gunsmoke_fixed

IC failed when tested out of circuit reporting trouble on PIN2:

74LS273@8D_failed

Fitted a new IC fixed the board completely.

 Posted by at 10:27 am

Rolling Thunder repair log #1

 PCB Repair Logs, Repair Logs  Comments Off on Rolling Thunder repair log #1
Aug 092015
 

Got this original Namco Rolling Thunder from my friend Alexander for a repair :

Rolling Thunder_PCB

Board didn’t boot up but got stuck on this screen which repeated itself in a endless loop sign that the watchdog circuit was active :

fault

My friend Alexander gave me an hint to troubleshoot this fault pointing me to a bad custom ‘130’

custom_130_

As I swapped in a know good one from a board of mine, game successfully booted but showed a strange graphical fault where the main character didn’t disappear when entering in a door:

sprite_issue

As usually I started my troubleshooting with a visual inspection and found a deep scratch on the solder side and after checking with a multimeter I could notice that involved trace was severed:

broken_trace_PCB

After consulting the schematics I was sure that the broken trace was indeed in the sprites circuitry :

broken_trace_schematics

Restored the trace cleared the sprite issue.Job done.

 

 Posted by at 7:39 pm
Jul 282015
 

My friend Josef sent me some boards for a repair and among them there was this Vulcan Venture (export release of Gradius II – GOFER no Yabou) :

Vulcan_Venture

This Konami hardware ( called Twin 16) is as beautiful as complex : two 68000 CPUs, one Z80 as sound CPU (Yamaha 2151 + YM3012 DAC, UPD7759C and a 007232 Konami Custom Chip as sound chips).Specifically video board use a lot of custom chips whose functions are not really documented.With this in mind I began my troubleshooting and fired the PCB up:

first_issue

Initial RAM/ROM self test reported a bad RAM @R8, this is located on video board in form of a MB81464 chip (a dynamic 64K x 4 bit RAM, there are eight of them in this section)

MB81464_DRAMs

I was very confident that this would have been an easy repair but I was tremendously wrong because, after replacing this RAM, the board died.I couln’t get anything on screen but only some flashing lines and sometimes I could barely recognize the “RAM ROM CHECK” message.Sadly schematics of this board are not available so you have to seatlle with the Dark Adventure ones which runs on same hardware but has different ICs location.The fault was located on video board since, as I said, board properly booted but it played ‘blind’.In particular looking at Dark Adventure schematics I noticed that the custom marked ‘007782’ generates some video timing signals:

007782_custom_

So, given the typology of  my issue, I started to check this part of circuit.All signal from the custom were fine but when I piggybacked the near 74LS244 @A4 the video output was restored.I fired up my logic analyzer which gave me this result on some gates:

74LS244@A4_logic_analyzing

As you can see output signals were absent (like in gate A1/Y1) or altered.Tested out of circuit, this 74LS244 failed miserably.Replaced it, board booted and displayed properly but was stuck on “RAM ROM CHECK” message.Analysing the CPU board revealed that 68000 sub CPU has /HALT line active so something was wrong in its bus preventing the proper execution of code.Also here the old but effective piggybacking technique came in help since since when applied it to a 74LS244 @10L (which, indeed, sits in the middle of the address bus of the 68000 sub CPU and its work ram) I could enter in game:

Gameplay was fine but background graphics were constantly rolling (besides, some of the audio PCM samples were garbled).I was pretty sure this fault was in the same location of the first one (around the custom ‘007782’ area).Looking at this part of circuit I noticed a small 220 pF ceramic capacitor soldered across ground and PIN15 (input) of a 74LS244 @A10:

capacitor_74LS244@A10

Usually little ceramic capacitors are used to clean signals filtering them.In this case the signal to this input of the 7LS244 came from the custom ‘007782’ and was called ‘NRAS’ on schematics while the corrispective output was connected to the RAS line (row address strobe) of the eight MB81464 DRAMs previously mentioned (they are used for tilemaps).Also here piggybacking this TTL , after removed the capacitor, cleared the issue ( actually this 74LS244 was tested as good in all my programmers but does not mean its internal junctions were fine).I was going to analyse the sound issue but board developed a new graphics issue :

new_issue_2

Text/characters were all blocky but I could quickly pinpoint this fault in a bad 74LS86 @B11 on video board using my HP10529A logic comparator:

So, now graphics were completely restored, it remained to be fixed only the sound.Sound hardware is made of a main Z80  as CPU , a Yamaha 2151 + YM3012 DAC, UPD7759C and a 007232 Konami Custom Chip as sound chips.Music was fine but some of the PCM samples (speeches) were garbled.For example the initial speech “DESTROY THEM ALL!” was played so :

 

while the correct should be:

Studying a bit the hardware with the help of The Main Event schematics and doing some tracing on the PCB, I could figure out that the MASK ROM containing speech samples is addressed by the NEC D7759C chip not directly but through a 74LS273 @7B (while the ‘007232’ PCM custom chip addresses the other samples MASK ROM). Probing this ROM I found that most of its address line were stuck LOW or HIGH.I dumped it and it matched the one in the MAME ROMset.So, at this point fault had to lie in 74LS273 @7B that was in the middle of address bus.Without thinking twice I desoldered and tested it out of circuit where it failed:

74LS273@7A_failed

Fitted a new chip restored all speech samples so board 100% fixed.The ‘Evil’ Konami has been defeated again!Well, at least until the next battle….

Just a quick note : all the  TTLs and RAM replaced were from Fujitsu.

 

 Posted by at 8:35 pm