Sega System 18 (Moonwalker / Shadow Dancer) PCB problem

aclbandit

Member
Joined
Nov 5, 2011
Messages
283
Reaction score
0
Location
Marysville, OH, Ohio
Sega System 18 (Moonwalker / Shadow Dancer) PCB problem

Been out of the hobby for half a year or more with work; finally trying to get caught up on some projects.

I've got a board repair that has me totally stumped.

Origin: a friend asked me to look at his Moonwalker because it had some graphics corruption issues. Basically a few of the sprites (for example, the children's "Help Me Michael!" text balloons) had this corruption pattern that I can best describe as "basket weaving."

So, the troubleshooting steps so far:
1) Eventually, I discovered that if I "pet" the EPR-13228 chip, the corruption wiggles (but doesn't go away).
2) Replaced all the big filter caps around the edge
3) Checked all ROMs, replaced one that only had correct data for some reads (new chip is consistent).
4) Replaced two encrypted ROMs with decrypted ROMs, replaced CPU with 'regular' CPU instead of battery-backed encrption CPU
5) Bought a Shadow Dancer PCB in desperation. Swapped stuff around between the two boards, and discovered that the issue follows the daughterboard, and has nothing whatsoever to do with the mainboard or the ROMs (both games work perfectly if ROMs are in the daughterboard from Shadow Dancer, regardless of motherboard on which the daughterboard is placed; both games exhibit the issue if their ROMs are placed in the daughterboard from Moonwalker, regardless of the motherboard on which the daughterboard is placed). The corruption issue affects Shadow Dancer in that the dog in the attract screen has the 'basketweave' pattern, and so does the control instruction sprites.
6) Reflowed solder
7) Replaced all the ROM sockets
8) Replaced the 74HC139 chip

The graphics corruption issue persists (and no additional issues encountered, so at least I haven't made it worse).


Anyone have schematics for these? Suggestions? Anything at all?

DhIZgUbl.jpg
 
Last edited:
There aren't schematics available as far as I know, I had a nasty moonwalker a while back it had two bad traces under a connector on the main board.

Might have to give up on this one. Building and documenting traces based on the working board is very likely more time than it's worth, unfortunately...
 
If you're sure the problem is on the daughterboard I think you can take some quick steps to narrow the problem.

Remove EPR-13228 and boot the board - are the corrupt sprites gone (replaced with solid block) or different in some way? (I think the sprite roms on this board are paired so I actually expect to see something remain).

Assuming EPR-13221 is the paired rom - remove that and boot - solid block or completely missing now? Problem is now narrowed down somewhat.

Without having a board in front of me, the data bus is likely shared across each of the 4 pairs of sprite ROMs. So put the ROMs back and check continuity across each data pin - so does pin D0 on EPR-13228 have zero ohms to D0 at 13229, 13330 and 13331? And does D0 on 13221 have zero ohms to D0 on 13222, 13223, 13224, etc? If you have a bad data trace this should tell you pretty quickly (because you know 13229, 13330, etc are good because they don't show any corruption, so any break has to be quite local to the ROM socket for 13228. You _don't_ have to trace to the main board connector for this reason).

If all data is ok - do the same for address pins - these won't all be common, because the higher address pins essentially select what pair to use - but I think the lower pins will all be shared.

And post pics :)
 
Thanks, tendril.


I've added an image of the issue to the first post.


Using this diagram for reference in pin naming: http://www.arcadiabay.de/images/elektronik/eprom/eprom_27C020_pinout.jpg.

This is it with the 13228 ROM removed:
iIjLrtFl.jpg


Additionally, when removing 13229, the corrupted portions become missing (which leaves a "jailbar"-like sprite pattern on the ninja dog). I did not photograph this, unfortunately.


I began by checking continuity between the first ROM (13228) and all the rest(13229, 13230, and 13231) at 'same pin' location for each (i.e., checking continuity between D0 on 13228 and D0 on 13229, then the same between 13228 and 13230...)

It looks like the 13228 and 13229 ROMs have continuity between each other on all pins except "/CE" (pin 22), which is expected since the line runs up to 13221 from that pin on 13228, and down to the mainboard plastic connector from that pin on 13229.

It's when checking continuity between 13228 and 13230 that I found the trouble. Pin 7 on each (A7 in the linked diagram) didn't have continuity. I checked against the known-good board, and it *did* have continuity between the two pins. I rechecked continuity of pin 7 between 13228 and 13229 (still good), continuity between pin 7 of 13230 and 13231 (good), and then finally between 13229 and 13230 -- which was no-good.

Double-checking against the known-good board showed that yes, indeed, that was supposed to have continuity between those pins.

I ran a small wire between the two pins on each side of the apparently-broken trace and fired it up again.

lw3CcfDl.jpg



Full ninja dog action! The board is fixed.

kmSUX5sl.jpg


Thanks again so much, tendril: this was some of the most incredibly specific and useful advice I've ever received.
 
For posterity, turns out there was another trace missing between the same two chips: pin 16 (ground). For whatever reason, there were no problems with that trace missing with the Shadow Dancer ROMs, but putting the Moonwalker ROMs back in would result in some substantial video noise around the edges of the stage name sprites when new levels load.

Ran another little orange wire between the pin 16s of those two chips, and everything is 100%.
 
Back
Top Bottom