About me

My photo
Vanløse, Copenhagen, Denmark
Mathematician. Working programmer/system developer. Nerd. Married. Father of 3.


Taito Double Dragon Bootleg Repair Log

This game is a true 80's classic (well at least around Roskilde, Denmark where I grew up). Back then almost all of my arcade gaming was done at burger joints with names like: Torvegrillen, Burger Place, Burger Boss, and Wopsi Burger. The latter was the best as it had 10+ standard cabinets with games changing about once a month...

Well enough nostalgia for this post; I proudly present (a bootleg of): Double Dragon!
I bought this just a few days ago, and by the symptoms it looked like an easy fix, so I started right away. A self test looking like this is always a good sign.

...but then trouble started. An intro screen looking like this:

Also the sprites where pretty messed up with stripes through them and some parts misplaced:

And the game had no sound at all!
Well easy to see that these two cabs in the sound section had seen better days

and not all that surprisingly, changing them brougth the sound back };-)

The board had some other obvious physical damages that needed fixing.
Here a knocked-off smoothening cap

1 transistor with 2 pins ripped off and the other one missing all together

A physically broken 157; the 04 just beside it had also taken a big knock at some point and was replaced for good measure as well. Both spares was taken from scrap boards.

Now I did expect to see some kind of change when replacing the broken 157, but not the kind that I saw when hooking the board up again...

Hmmm, must have done something wrong when fitting them sockets. Pulled the 157 again and had a look. Now I really appreciated myself for always using breakable header sockets instead of normal sockets, as this gave my a clear look down between the two rows. I found one of the tracks between a via and one of the pins suspicious, and the continuity tester confirmed my hunch. 

Patched it up with a little piece of kynar on the solderside.

Now the boot error had gone, but there was still no change in regards to the graphics on screen.
Until now I had only concentrated on the primary PCB, so now I had a look at the secondary one. The component side looked incredibly fine, considering the damages found on the primary one. The solder side however had a fair among of scratches and was also pretty dirty.

I gave the worst place a light scrup with rubbing alcohol and tooth brush, and I found 1 pin touching a nearby via and a track that was scratched in 3 places. Again the continuity tester confirmed...

So straitened the pin (along with about 5 others that I found, for good measure) with a Stanley knife a alot of caution. Then patched the broken track with kynar.

And NOW, the intro screen

and the sprites where alive and kicking again };-)
(please take note of the cute up-skirt view you get in this pic };-P).

Now all was good, and I started test-playing the game. But when I tried a 2-player game, the hit-button for player-2 was unresponsive. A quick glance at the input section on the primary PCB, and my eyes caught this broken resistor array.

By measuring on one it's sisters, I found that this was not the same type as nomally used for pull-ups (with a common point). It was more like 4 parallel resistors just packed into one house. I couldn't find any of those on any of my scraps, so desided on just fitting 4 standard 220Ohms resistors. Now the original bootleggers had put 2 componentpins into each hole, but with the pin-diameter of the standard resistors, I had no chance of doing that. So I desoldered the standard pull-up array

and resoldered it so that it was possible to access the pins from the component side. Then I soldered the 4 standard resistors directly on the pins from the component side.

And once again our Red Hero is able to swing his pork chop against the baddies };-P

And finally this case is closed };-D


SNK Fighting Soccer Bootleg Repair Log

Now this board is a relatively new investment (from the same batch as M.I.A.). A bootleg of SNK's Fighting Soccer. Before even connecting it, I could see that it have seen better days.

Here a broken cappy...

And had someone tried to drill holes in the PCB???

...not only two but THREE PLACES...WTF?!

Without getting my hopes up too high, I replaced the broken cap and hooked the poor thing up in the bench. Well first thing I noticed: No sound! But when I turned the volume pot on the board up, I did hear a hissing sound (It Still Alive };-)). Hmmm, a fast check with the voltmeter and it was obvious that this board didn't get -5V. Soon I found out why?! My JAMMA-extender didn't have -5V wired (note to self: Why do you keep buying premade stuff, that you could have done better yourself? Grrrr). Tried to hook up the board directly in the benches JAMMA-slot...

...and heard the well known SNK-boot-jingle from the speakers };-D
Wired the -5V on the extender and hooked the board up again.
A lot of the sprites had horizontal stripes through them.

By shortening adjacent pins on different RAMs and ROMs I quickly saw, that the graphics was (as with most boards with 2 PCBs) handled by the secondary PCB. So tested all data and address pins on the RAMs and ROMs with the scope, but found no oddyties.
Now, it way getting late and I felt a bit tired. I did however decide to do a little piggy-backing of TTLs before going to bed. The board have a cluster of 3 245's in the middle and I just happen to have a tube of them on the table, so I thougth: Let's give it a try:

 ...and to my big surprise, all sprites looked normal again. Quite a lucky punch (I so i thougth!).
I quickly desoldered the IC and put it in my Top2005+ for testing...hmm, the test came out successful??? The one I've used for piggy-back (the one from tube) also passed the test.
  • Fitted a socket and inserted the new one: Stripes through sprites!
  • Inserted the old one back into the socket: Stripes through sprites!!
  • Piggy-backed the new one on top of the old one: Still stripes through sprites!!!
(BTW: Did you notice that the words "sprites" and "stripes" have the exact same letters...kind'o'funny };-P).
Now I was really puzzled...Then tried piggy-backing the new on top of the 2nd 245 in the batch of 3: No stripes??? WTF!
I wanted to get to the bottom of this, so desoldered the 2 others as well and fitted sockets.

Slammed in 3 new ones from the tube: Stripes!
Put the old ones back: Stripes! (MOAN!)
I discovered that one of the 245's wasn't inserted fully into the socket. With the board still running I gave it a push, and the stripes went away?! Released the pressure and stripes came back. Began pressing on IC's in  the surrounding area: Same thing happend!!!

Picked up the board and bended it slightly with the same result.
Now we are getting somewhere; this smelled an awful lot like a short-circuit of some sort.
When looking closely at the joints on the solderside, there was alot of them looking very nasty:

This is just 2 examples, but this board has extremely many of them.
So started a shotgun-scheme, where I did a reflow on the nasty joints I could see checking the board every time. After about 10 of them with no luck, I decided to change strategy into a more systematic brute-force approach.
Started by cleaning the lower half of the board using rubbing alcohol and a toothbrush.

Then starting at ground-zero (the place where I had to apply the least pressure to provoke the fault), I began the process of reflowing the joints on every IC while working my way to the edges of the board. After having done an IC, I tested the board. And on the 18th IC I hit the jackpot!

It was this PROM who was the little troublemaker };-)

And now all the sprites looked good at all times };-D

Just before I close this log off, I'd like to share with you all this piece of beautiful hairdresser art };-P


Konami Missing In Action (M.I.A.) Rep Log

Now this board, an original Konami Missing in Action (M.I.A.), I snapped up cheap in a batch of 7 non-working boards paying about 100DKK (~$19~£12) each. Standard Konami GX-<something>; only one PCB and quite many custom ICs.

The problem with this baby was, that one (or maby more; a bit hard to tell) of the background layers and all the text had vertical stribes, where the graphics was just simply missing (in action, ha! ha! };-P).

Even in test mode it was same-same!

Now the board have had a rough time at some point; quite alot of scratches on the solderside.

But as the tracks and vias on the GX-boards are quite thin, it wasn't possible to test for continuity without desoldering some ICs in order to follow the tracks around on the component side. As I was a bit too tired for that, I gave the scratched tracks the benefit of the doubt, and started poking all the SRAMs with the scope. But everthing seemed quite healthy on both address and data lines.
I then started shortning adjecent pins on some of the TTLs near the SRAMs involved in the graphics rendering. That did stir the picture up a bit, but nothing that was so obvious, that it was worth investigating further.
Now had a closer look at the 4 PROMS in the top right corner of pic1. By removing them one at the time and trying to boot (and it actually did; that 'self-test' at start-up isn't worth a flying hoot on this boad };-P), it was obvious, that these contained the graphics for the game. Poking each data and address pin on them, I found that pin 18 (Q4) on the PROM labeled "808 E13"

was stuck low.

The corresponding pins on the 3 other members of that "PROM-family" was extremely active...so I decided to try and dump it with the Top2005+. But when trying to do so I got this:

Pulling and reinserting the IC a few times gave the same result.
This was quite art, as this was not pin 18, but something WAS clearly wrong with this mother frakker. Ahaaaa (his cried out loud!).
Now the PROM is an 27C512, so found a nice TexasInstruments TMS27C512 EPROM from a scrap board. Ripped the window sticker off and removed the remaining glue from the window with rubbing alcohol and an old tooth brush. I was quite eager, as I had never had a chance to use the Top2005+ to actually program an EPROM before.
So into the eraser for 3mins it went

Checked it in the Top, and as it was now filled with all nice 1's, 3mins seemed to be enough. Now downloaded the MAME ROM and programmed the right image to the newly erased EPROM.

Verified the IC, slammed into the socket aaaaaaand:

Now wrote a nice little label to put on the window of the EPROM, and the board is as good as new.