Space Invaders DX Repair Log

Azure Oz2

Well-known member

Donor 6 years: 2020-2025
Joined
Oct 22, 2020
Messages
1,444
Reaction score
965
Location
Chatswood, NSW, Australia
Here is some info on a Space Invaders DX repair. Hoping it may help others.

There is not a lot posted about this pcb and problems with it. The Space Invaders DX normally has a part label. This one was faded away. After cleaning the original pcb number under it on the silk screen is K1100787A J1100332A. I think the label that goes over this is labelled "K11J0813A Space Invaders".

This particular pcb had corrosion in multiple places on the pcb. I did not take a before picture but will try and start doing that in the future.

Here is the pcb in question now working after a good soak and clean with simple green and blowing under all components with compressed air to dry and hopefully remove anything left behind.
SI DX pcb.jpg

Here are the IC sockets and resistors removed and replaced that showed visible corrosion damage. One electrolytic cap had only legs left, so I replaced all 5 that were close to where the main corrosion was.
Replaced sockets.jpgReplaced passives.jpg

All EPROM and PAL pins were cleaned on the inside and outside before being placed back on the pcb. This was to make sure they made good contact with the dual wipe sockets.
 
Last edited:
After cleaning and first checks it appeared the clock was running and there was activity in the video circuitry.

There was no video output or signs of life from the game.

Further scoping showed most signals around the CPU were tristate or low, the /RESET signal on the 68000 was stuck low. Tracing the /RESET back through IC29, the inputs were also stuck low. This was driven by a single pin on the Taito TC0640FIO IC53 Pin 94 that was also stuck low. At this stage I went down a rabbit hole searching for info on this chip - I came up blank.

After some more investigation I identified IC54 an MB3771 Voltage Monitoring IC which generates a RESET output on Pin 8 that goes into IC53 TC0640 FIO Pin 115. This output was stuck low and all input seemed ok. I changed the 1uF tantalum that determines the reset delay just in case. Given this is a Fujitsu IC and for TTL it is common for them to have a dead outputs, so I tried pulling the reset output high manually.

Reset to the 68000 now worked. I did not have an MB3771 on hand. It turns out the TC0640 FIO also starts generating the watchdog reset once /RESET in Pin 115 goes high. So using this I placed a 10K pullup on the MB3771 output between pin 5 Vcc and pin 8 /RESET.

Here is the pullup resistor added under IC54.
Reset fix.jpg
68000 now resetting and activity on address, data and control pins. Game still not working and still no video out.
 
After some more scoping around, signals on game EPROMS looked a little strange. Before fitting the new game EPROM sockets I did check that the tracks under them were not corroded.

I decided to remove the 68000 as I could not get continuity between some address lines going to the EPROMs.

Under the microscope the tracks looked ok and Pads looked fine. Lesson here is don't let looks fool you. After probing every pad to trace I noticed there were six Pads that were open and making no connection with the tracks. Here are the Pads in question with the resist scraped back on ones not making connection with the pad.
Open Pad connections.jpg

I don't like messy bodge wires when repairing damaged pad/tracks. I soldered some fine Kynar wire to each track and put the wires through the pads. Here are the wires soldered in place.
Repaired Pads.jpg
I checked the socket still fitted and went on to add some resist.
Solder Resist added.jpg
This allowed me to fit a socket and trim the wires flush with the socket pins to provide some almost hidden track repairs.
Pad repairs.jpg

After powering on again, I was greeted with some video.
DX Title.jpg

Sadly, no sound. Still more to do.
 
This game has an electronic volume control. I looked up pinouts for all the audio ICs used. Nothing seemed out of the ordinary, but there was no sound.

I started checking address and data lines from the Z80. Before fitting the Z80 EPROM program socket I checked the tracks underneath were ok. After what I found on the 68000 I decided to remove the soldered in Z80 and check track/pad continuity a far as I could easily trace. Turned out to be worthwhile, this time there were 3 Vias that were not making connection between the component and solder side tracks. These were repaired the same as the 68000 Pads.
Via repairs.jpg

Fitted a socket for the Z80 and powering up again this time I got sound. Game now works. But the text to choose a game is in Japanese.
 
Last edited:
I had to replace IC5 as the one from the pcb was badly corroded and one leg was completely gone flush with the edge of the ceramic. I did a romident for all the other IC's and they matched the mame version. I programmed a replacement D89-05 from the set that all the others matched.
U5 JP EPROM.jpg
After doing a little more searching it turns out that IC5 is the only one that changes between the Japan and US version. I checked with the owner which version they wanted and after getting confirmation they wanted it in english programmed another eprom with d89-xx.ic2.
U5 US EPROM.jpg

It is interesting that only 1 eprom needs to be changed which provides the low data byte and the one with the high data byte stays the same for both versions.

Game is now completely working all inputs tested, eeprom working, sounds/volume and game play all working.
 
TAITO TC0640 FIO Pinout

This is a draft on the pinouts for this IC based on how it is wired up on the Space Invaders DX pcb. I did not remove the IC so was unable to check the few pins I could not figure out and pin vs pcb wiring of Gnd/Vcc, etc.

It seems to serve a few purposes:
1. Take Reset in from MB3771 voltage monitoring ic and generate Reset out for 68000
2. Generate watchdog reset
3. Interface to EEPROM
4. Interface to user controls and outputs

It connects to the upper half of the data bus, uses 5 address lines and 2 CS lines from the PALs/GALs.

Color coding in symbol is Black = Ground, Red = Vcc, Dark Blue = Output, Light Blue = Input, Green = Bidirectional.

I have traced out the wiring to the EEPROM, Opto isolators and output drivers including all pinouts and pullups for a schematic. Not sure if I will get to draw that up.

Here is the draft pinout of the TC0640 FIO
Taito TC0640 FIO Draft.jpg

I you notice any errors or have more info please post or pm and I will update it.
 
Very nice repair log, thank you for posting this. These Taito boards have the tiniest pads/traces making them very difficult to work on without lifting them. I am impressed with the track repair!

p
 
Back
Top Bottom