Ghosts N Goblins Flashes on screen?

[tyranix95]

@K.D.:

Thanks for the props K.D.; however, if I can fix this board, then the credit will go to all who helped out.

Right now, I'm just a little guy trying to stop GHOSTS from GLITCHN.

@CM:

I replaced all 15 LS245 chips on the GHOSTS board.

I replaced the 9 245 chips, on the top board:
----the three 245 chips, on the CPU page (A-1-1/8), @ 5N, 6N, 11H;
----the three 245 chips, on the I/O PORT page (A-2-2/8), @ 1B, 2B, and 3B;
----the one 245 chip, on the CHARACTER VIDEO RAM page (A-2-6/8), @ 10H,
----the two 245 chips, on the COLOR MIXER page (A-2-8/8), @ 5C, 4C,

And I replaced the 6 245 chips, on the bottom board:
----the five 245 chips, on the OBJECTS DATA RAM page (B-2-1/9), @ 14E, 11F, 12F, 13F, and 14F
----the one 245 chip, on the SCROLL VIDEO ROM page (A-2-8/9), @ 10A;

Replacing all the 245 chips was the first shot at the GLITCHES.

 

[tyranix95]

One thing that I've noticed so far is that the the INTRO to the CAVES, with the bats does NOT seem to GLITCH any more. The game used to GLITCH like crazy with the bats fell from the ceiling. So, progress has been made.

BUT, the other areas, like the ICE TOWERS, HOSE OF BIG MEN, HAUNTED HOUSE, and CAVERS (DRAGON) still GLITCH. Just a hunch: Been thinking that these are background and timing related events rather than character related events.

Still working under the presumption that both PROM chip m-01 and Prom chip m-02 are good and the RAM chip at 2C is good because the screen does not white out.

 

[channelmanic]

If the screen fully whites out then the main CPU is locked and the game isn't running.

The video output necks down to a video PROM on the main board and if that PROM isn't doing it's thing then you are stuck on a white screen.

This is easy to tell if you can readily get the game to lock up by letting it lock up with that PROM out of the socket. The screen is there, but you are missing elements.

 

[tyranix95]

OK. Replaced a few more chips today and then play tested the game once with the m-01 chip and once without the m-01 Prom chip.


On the CPU page (A-1-1/8), I replaced one 08 chip at 7K; two 138 chips @ 7L an d 8K; one 21 chip @ 8L; and one 32 chip @ 9L. And, On the MEMORY MAP OUT PUT page (A-1-2-2/8), I replaced one 174 chip @ 3C. Tested game with m-01 PROM chip on the board. No, change: Game glitches.

Removed m-01 PROM chip from board. Tested game. Noticed background and foreground graphics are missing (blacked out); however, header, footer, and title graphics display perfect. Also noticed that game plays blind fine. After a few minutes of messing around, was able to get the TORCH weapon, shoot at a bunch of stuff, HEARD a BIG bonfire, and SAW a color flash on the screen. No, change: Game glitches.

On the LINE BUFFER I page (B-2-5/9), I replaced the two 257 chips @ 14M and 14 N. Tested game with m-01 PROM chip on the board. No, change: Game glitches.

On the OBJECTS DATA CONTROL page (B-2-3/9), I replaced the six 157 chips @ 13H, 13J, 14H, 14J, 9H, and 9J; and, I replaced the three 273 chips at 5J, 6Jand 8J.. No, change: Tested game with m-01 PROM chip on the board. No, change: Game glitches.

 

[T.2193]

Hate to necropost, but I was wondering if tyranix95 ever got this resolved. I have a similar issue on my board (Makaimura rev G), and had just assumed it was a programming fault, hardware limitation, or use of inferior components or something, probably since I have other games with what are apparently built-in bugs.

 

[tyranix95]

Hey T3,

Sorry bud: I haven't fixed my GHOSTS board yet. Been looking over the schematics off and on--planning (and saving up for) my next move.

So far, the flashes appear to be a hardware fault of some sort. Lots of GHOSTS board owners (with different GHOSTS board revisions) also have "The Flash." And, the flashes seem be to located at specific points in the game and occur at different intensities. My board just seem to have it worse than others.

channelmanic suspects that the flash is caused by data not being sent from a ROM to a chip on the COLOR MIXER (Video Card) page. And that missing data is what is causing the screen to "flash."

Back in the day, GHOSTS never had a reputation for being "Flashy," "Glitchy," or "Buggy" like another CAPCOM game--Street Fighter II. But, even that Glitchy game was not "Flashy." GHOSTS just had a rep. for being Good 'N Hard.

 

[Muel]

One thing that I've noticed so far is that the the INTRO to the CAVES, with the bats does NOT seem to GLITCH any more. The game used to GLITCH like crazy with the bats fell from the ceiling. So, progress has been made.

BUT, the other areas, like the ICE TOWERS, HOSE OF BIG MEN, HAUNTED HOUSE, and CAVERS (DRAGON) still GLITCH. Just a hunch: Been thinking that these are background and timing related events rather than character related events.

Still working under the presumption that both PROM chip m-01 and Prom chip m-02 are good and the RAM chip at 2C is good because the screen does not white out.

This is what and where my GnG board starts to flash as well. It will also sometimes flash if you are doing the "extra points manuver" in the cave if you are killing multiple goblins at once.
My GnG board is still virgin and I have not tried replacing any ICs on it, not that I don't want to, I have plenty of other arcade projects goin on right now and this is one that I will eventually get around to fixing.

Just a thought, could this be a bad data cabling problem between the two boards? I know that for example on one of my Williams games, just putting pressure on the ROM board cable so that it would make proper contact was enough to get the game to boot properly - I replaced the ROM board cable and the problem of boot errors and intermittent booting went away. Are you still using the original cable pairs? When these games were new 25+ years ago, so were the cables. These cables age and get as stiff as a board and the copper inside becomes fragile. I haven't even looked, but I'd imagine that if you can't find replacement ribbon ends, the current cable setup could be rebuilt??

 

[orion3311]

Mine flashes too...one spot I know is right at the first level...about halfway past where the hill is.

 

[tyranix95]

...Are you still using the original cable pairs? When these games were new 25+ years ago, so were the cables. These cables age and get as stiff as a board and the copper inside becomes fragile. I haven't even looked, but I'd imagine that if you can't find replacement ribbon ends, the current cable setup could be rebuilt??

Yup...my board still has the original ribbon cables and board connectors.

Early on, I noticed that some of my other working CAPCOM games have the same 50 pin ribbon cables as GHOSTS; so, early on, I tested my GHOSTS board with a couple of extra sets of working 50 pin ribbon cables.

And, I found out that GHOST still GLITCHED with my other working ribbon cables. And my other working CAPCOM boards still work fine with my GHOSTS board ribbon cables. So, I'm thinking I the GHOSTS board ribbon cables are good.

But, what I haven't done yet is replaced the any of the 50 pin board connectors. I gave these a good cleaning; however, I've been focusing more on the chips that control the flow of data (communicate) between the top board and bottom board, from a rom, to a color (video) pin.

Mine flashes too...one spot I know is right at the first level...about halfway past where the hill is.

Yup, sometimes mime flashes there too, when the

shoot at

.

 

[T.2193]

Not sure it will help with anything, but I guess I will outline my experience. It flashes at similar spots to you guys (the blue towers at the start of level 2, the lifts at the end of level 2, the beginning cave at level 3, the dragons, etc), but I've noticed that it is kind of sporadic, especially when it comes to the dragons. I recall that one time it flashed quite a bit at the level three dragon, but I've also had it appear with hardly any flashing at all.

I'm pretty ignorant when it comes to this stuff, but is there any explanation as for why this issue is apparently so prevalent?

 

[tyranix95]

Yup, my board flashes in all those spots too. And I found out that my board FLASHES much worse inside the CASTLE levels.

And I'm not sure why all these GHOSTS boards got the GLITCHES. But, I'm sure I'll make sense when we figure it out.

So far, I've only been able to get the start of the CAVES level (where the bats fall from the ceiling) to stop GLITCHING. Doesn't seem like much; however, it's progress.

I need to mess with my board some more, next weekend.

 

[orion3311]

I initially thought it might be bit rot on the roms...has anyone yanked and verified the roms?

Has anyone tried to see if the same thing happened in mame?

 

[tyranix95]

I have not checked MAME.

But, GHOSTS does not Flash in the console ports

Also, mainman already pulled and checked all the GHOSTS roms on my board. He checked the sum of each of my GHOSTS roms against the MAME files and found all roms matched the corresponding dump file.

Last time I touched my GHOSTS board, I had 1/2 a new set of GHOULS roms burned and I was in the process of testing the board with the new rom set. Thus far, I had tested the board with 3 roms of the 11 rom that I burned and there was no change--the game still GLITCHED.

 

[T.2193]

Yup, my board flashes in all those spots too. And I found out that my board FLASHES much worse inside the CASTLE levels.

Yeah, the worst I've seen so far in mine is right around where you face the dragon in the castle. This part also seems pretty bad on the two different PCB capture videos that I see on YouTube:

http://www.youtube.com/watch?v=DfSna5wCyTk
(at about 5:35)

http://www.youtube.com/watch?v=IizCbGmygIQ
(at about 3:05)

 

[tyranix95]

Yeah, I've seen my board goes nuts (Flash like crazy) in the CASTLE part where the two UNICORNS (?) (Level 1 Forest boss) are. It's so bad its almost unplayable.

Heh, the GHOSTS board in your vids. have the the CROSS. My board does not have the CROSS. It has has the SHIELD.

 

[gomados]

Maybe 9A , 9B or 9C are the problem?

i have original board but came with italian bootleg son 3,4 and 5 eproms not original, and i have issues in ice lift but not on bats, reflashed to japan G and more flashes came on bats for example, i reflashed to italian bootleg and no flashes on bats, any clue.

 

[PJDiCesare]

I have a lot of experience with Ghosts 'n Goblins, and I can say this has happened on every cabinet and board I have ever played on, and usually in the same areas. It almost seems like this is how the game handles sprite overloads. Instead of crazy slowdown/lag, you get the thunderstorms.

When I play on the higher loops (5+) where the enemies spawn much faster, this happens extremely frequently. I've always assumed this was just part of the game and never looked into fixing it. Haha.

 

[gomados]

Guys Mistery resolved!

(From Jotego one of the best FPGA developers)

Ghosts'n Goblins Color Flash

A couple of years ago, someone wrote me asking about yellow flashes in Ghosts'n Goblins game PCBs. I thought it was due to some timing problem in the aged hardware, which probably just got worse with temperature as the board got powered for longer time. However, this week I came across a video from an original PCB that showed this yellow flash effect. It was very prominent. I remembered then that the GnG hardware has a design flaw: it is possible to corrupt the palette memory if the CPU tries to write to it outside the vertical blanking period. I had originally prevented this from happening in the FPGA core for robustness and because I had assumed the game would not try to do that.
After watching that video, I decided to remove the protection circuitry and let the color memory get corrupted. And wow, does it so! The yellow flash effect happens a lot, especially when you fight the dragon boss. That makes me think that it was used purposefully by the game designers, at least on some occasions. The game feels different with this effect and I recommend playing it through. It is especially notable in the last castle stage.
The Speed Rumbler had the music running at a lower pitch. It was a bit hard to debug this issue but it eventually led to a problem in the Yamaha 2203 timer implementation. Those timers are quite tricky because small differences in the implementation can lead to rare problems.