As mentioned in a previous post, when I converted my Jumping board to Rainbow Islands I noticed that all the text in game was black. Going back through some pictures I took when I first got the board I see its present there too.

So where do I start?
I know from making my conversion that the palette values are stored in the program ROMs and they are pushed out to the palette RAM at address 200000.
I looked at the RAM with a scope and all seemed fine plus all the other colours seemed fine.
I also know that all the values required for this text colour is when the 2 least significant bits (A0 and A1) are active, so I traced these out down to the video board.

After a long session of following various signals I came across a 74LS374 chip.

This had all its inputs stuck low but according the the scope had a fair bit of noise on them, enough noise in fact to make my logic probe see it as pulsing HIGH/LOW. Following these inputs over to the chip next door I see a 74LS273 with its outputs giving off this crazy signal.
I replaced the 74273 and booted the game

If you look carefully you can see the lower part of the text is missing and a bit scrambled. Taking another look at that 74374 chip and its outputs are still crazy despite having good inputs. After a quick desolder and replacment I got this

All fixed and now my conversion is totally complete.
I have left one bit of graphics the same as the Jumping. If you play the game, in the lower left corner, you number of current lives is displayed as blue stars. On the original these are small rainbows. Just a note in case I ever release my modified files and someone tries to sell it as an original or something.

So I saw this Taito board on eBay. The seller says its broken and unknown but was most likely a Super Chase and supplied a System 16/24 adapter with it, which in its self is very nicely made and they can cost quite a bit.

Only thing is, Super Chase is not a System 16 or 24 pinout.

*pinout supplied by Stefan Lindberg

I made a quick and dirty adapter up with and tested the board, dead.
Now I made a very stupid mistake at about this point. I moved the board down from the test bench and into one of my cabs just to rule out any voltage issues that ive seen before BUT for some unknown reason I left the pinout adapter in the test rig.

After the smoke dispersed I checked the damage.

Power supply is fried. Turns out, without the adapter I tied the +12v line to the +5v line and destroyed the PSU.
At this point I didn’t hold out much hope for the board anymore but tested it anyway.

The board was unsurprisingly still dead.

I check the RESET and HALT lines on the CPU, the watchdog was doing its thing.
Pulled and checked the program ROM’s and 2 didn’t match.
A strange thing here. One has only 1 byte difference right at the end and the other was clearly trashed with random read outs at every read.
I copied the first dump over to MAME and ran it. Turns out this is probably a different region to the one in MAME and that byte difference either displays the country warning screen or not. Mine does not.

I got nothing but a bit of garbage on screen

I fired up MAME and checked what the game should be doing and it does perform RAM checks. I also saw in the code that it has RAM fail messages on screen should it encounter a problem.
Probing around the board in this state didn’t help at all as the watchdog reset everything before I could really see what was going on.
After a quick look at the CPU code I see I can bypass the shared and screen RAM checks very by changing only 1 byte. I did this and replaced the relevant EPROM and watched the game boot

So the screen RAM is at fault. The screen RAMs are the 2 Sony 58257 chips in the center of the board.
I pulled both of these and both failed. Replaced them and the game boots once more.

A narrow escape I think. This will hopefully teach me to slow down a bit, but I doubt it