Ghosts & Goblins Bootleg Repair Log

I had a Ghosts and Goblins PCB on the bench recently, it's a bootleg, which I rarely bother with, but GnG is one of very few that actually are worth repairing, mainly because the official Capcom boards are in the ultra-silly money price-bracket.

 

http://i40.photobucket.com/albums/e230/Womble76/Ghosts_Goblins/IMG_4313.jpg

 

http://i40.photobucket.com/albums/e230/Womble76/Ghosts_Goblins/IMG_4314.jpg

 

It was sent with the fault that the game runs, you get sound, but the screen is a complete mess.

 

http://i40.photobucket.com/albums/e230/Womble76/Ghosts_Goblins/IMG_4297.jpg

 

Yep! It's possible to see the Ghosts and Goblins title screen and the high score table among the swirling mess and those components were stable on the screen - so I could rule out a sync or video ground issue.

 

Initial suspicions were that this was at least two faults in action, one that was causing all the crap in the foreground and another that's causing the instability. Normally when you have a gfx fault the faulty graphics are at least drawn cleanly on the screen.

 

The screen full of rubbish was going to be a logic or RAM failure on the board so I spent a few minutes with the scope looking for anything suspicious in the video output stage, just inland from the ladder DACs on the RGB lines.

 

After a few minutes the attract mode stopped running, and was replaced with a rhythmic ticking noise through the speaker. The board has crashed and the watchdog circuit was trying to reset the PCB.

 

The other sign something new was wrong was that the EPROM 12 was now getting stinking hot.

 

http://i40.photobucket.com/albums/e230/Womble76/Ghosts_Goblins/IMG_4303.jpg

 

Pulling this EPROM revealed that pin 4 (address line A6) was bent underneath the chip...

 

http://i40.photobucket.com/albums/e230/Womble76/Ghosts_Goblins/IMG_4304.jpg

 

... which is a sign someone has been investigating this board before, and has tried to reseat the ROMs. The fact this board was able to run means this pin would have to have been making contact with the socket on that pin as it contains the code for one of the CPUs, without that the game would have watch-dogged from the outset.

 

Putting the EPROM in my reader confirmed that it was completely dead, no surprise really, when they get hot it means they have shorted internally. Even if that pin had been making intermittent connection to the socket it would not have killed the chip, something else was going on. TBH I only tried to read it as I was struggling to see how the board could have run with a heater in place of a program ROM, and wondered if I hadn't noticed it getting hot earlier.

 

So I moved on to dump the other ROMs (it is usually something you should do very early on, to avoid wasting time trying to find a hardware cause for a issue caused by bad software) and found a level of destruction I've rarely seen on a board. Of the 17 ROMs on the board all but 5 were dead. The rest were either not recognised by my burner, and/or were dead (either containing FF or F0), or they gave a different checksum every read attempt, meaning they were just putting out rubbish.

 

Of the 5 EPROMs on CPU board, 4 were dead and of the 12 on GFX board 8 dead, and the dead were a mix of 27c128s and 27c246s from NEC, Fujitsu and Mitsubishi brand.

 

At this stage I wondered whether my bench power supply had decided to die and take the board with it, but with those ROMs off the board the logic sections still looked active and healthy, i.e. the system clock and divider logic was still buzzing away normally.

 

As for what the hell happened, I'm not entirely sure. This is a Capcom Classic Pinout board, that uses the same 56 way edge connector as JAMMA, so it is quite likely at some point that someone plugged it into a JAMMA harness, which would pump 12V into one of colour channel outputs, and -5V into another, putting 17 volts across the pair, but this wouldn't just take out only the EPROMs, and a lot of the destruction was on the back board, a long way from the edge connector and behind a lot of logic that was undamaged.

 

To kill so many EPROMs, and only the EPROMs, across two types, and three manufacturers, you would really need to power them up with reversed polarity, probably by taking them off the board and putting them back in their sockets the wrong way round. Plugging the harness in backwards, or giving it a dose of JAMMA love doesn't reverse the polarity on the power rails, and if it did, all the logic would be fried too.

 

The bent pin proves someone has been here before, so I can only assume a load were re-fitted backwards, with most getting killed, but some just hanging on by a thread, most of the graphics ROMs were probably dead on arrival, with the rest dying a bit later on.

 

Anyway - the only way forward was to burn a new set of program ROMs for the CPU board. With those installed the board would boot again, and as there was no improvement in the gfx behind the mess I am fairly sure the GFX ROMs were dead all along.

 

http://i40.photobucket.com/albums/e230/Womble76/Ghosts_Goblins/IMG_4321.jpg

 

So, back to where I started, progress of sorts.

 

http://i40.photobucket.com/albums/e230/Womble76/Ghosts_Goblins/IMG_4323.jpg

 

Replacing all the GFX ROMs meant I needed all the 27C128s I had in stock, after a decade of fixing boards the bottom of my stock drawer contained a mix of manufacturers ICs, some labelled, some not, and none that were blank and ready to use. So while I was putting them all through the UV eraser, 3 at a time, I focused on the issues affecting the video stability. This board is peppered with Fujitsu TTLs which are becoming a plague on old boards. Usually they die cleanly leaving floating outputs but on systems where the output is connected to other chips outputs and time slice controlled by other logic you don't always find the smoking gun on the first attempt. My main suspicion fell on the set of 3 74LS245s in the final video mixed circuit as they looked pretty nasty on the scope, but nothing was entirely floating, due to the shared nature of the bus.

 

http://i40.photobucket.com/albums/e230/Womble76/Ghosts_Goblins/IMG_4331.jpg

 

Murphy's law meant that I had to remove them all before finding the last one I removed was the real cuplrit.

 

http://i40.photobucket.com/albums/e230/Womble76/Ghosts_Goblins/BadL245_9D.png

 

Still, a Fujitsu off the board is a good Fujitsu, and all were replaced with bullet proof Hitachi 74LS245s.

 

With that done the image was finally clean, just lacking a lot of data with only the few surviving EPROMs present.

 

http://i40.photobucket.com/albums/e230/Womble76/Ghosts_Goblins/IMG_4336.jpg

 

http://i40.photobucket.com/albums/e230/Womble76/Ghosts_Goblins/IMG_4342.jpg

 

With a full set of ROMs for the background data on board things were much improved...

 

http://i40.photobucket.com/albums/e230/Womble76/Ghosts_Goblins/IMG_4347.jpg

 

...and after a final batch of ROMs erased and programmed to complete the sprite section all the game characters were back.

 

http://i40.photobucket.com/albums/e230/Womble76/Ghosts_Goblins/IMG_4357.jpg

 

The only issue remaining was twitchiness in some of the graphics, at times it was perfect...

 

http://i40.photobucket.com/albums/e230/Womble76/Ghosts_Goblins/IMG_4353.jpg

 

... but seconds later...

 

http://i40.photobucket.com/albums/e230/Womble76/Ghosts_Goblins/IMG_4354.jpg

 

The issue only affected all text, life icons and the Ghosts and Goblins title banner, and from poking around it was clear this all came out of ROM11. This is controlled by a 74LS86 at 9D(Fujitsu) which had floating inputs on pin 2 of all 4 of the gates, the upstream chip responsible for those inputs being another 74LS86.

 

http://i40.photobucket.com/albums/e230/Womble76/Ghosts_Goblins/BadLS86_9D.png

 

With that replaced the lines lit up correctly and the issue was cured.

 

http://i40.photobucket.com/albums/e230/Womble76/Ghosts_Goblins/IMG_4359.jpg

 

Board fixed, or so I thought.

 

 

The lower one was pumping out the analog audio for the baseline, but the upper one's output was only weakly driven, putting the scope probe on the output pin did cause some of the missing audio to leak through to the pre-amp. Handy as that proved the line itself wasn't stuck, that there was audio getting to the DAC and that the pre-amp could actually amplify. So I whipped one off a scrap board and installed it.

 

http://i40.photobucket.com/albums/e230/Womble76/Ghosts_Goblins/IMG_4376.jpg

 

[video=youtube_share;QPQFg--7rAA]

Edited December 7, 2018 by Womble

Awesome write up Womble :)

I always love reading your threads. They're well written, interesting and contain a wealth of detail. Thanks for taking time to to put the post together for others to learn from :)

Another top write up mate :)

Thanks Womble for rescuing this one for me, fantastic write up, working beautifully. I will be sure to invite the buddy i bought it off over for a few beers - he will be wrapped as it had been in his garage unworking for a decade.

Womble to the rescue again!!!! Arcade games are lucky to have this god around. :)

Hi @Womble, just wondering if you have time to confirm the finger board adaptor wiring used on the bootleg GnG PCB for me?.

As I have documented in my thread on my query

https://www.aussiearcade.com/showthread.php/96017-Jamma-harness-question

 

I have a discrepancy between the documented Capcom wiring and the wiring that came with my board.

The Capcom wiring says Pin 22 on the parts side should be counter 2 GND but my connection is to +12v from the Jamma harness?

I just want to check if I should leave it like this or change to follow the Capcom wiring ?

Thanks

I'm afraid the GnG went back to its owner a couple of weeks ago now, and I don't have one in my own collection to check.

 

I'd say your harness is wrong, GnG bootlegs were either "Capcom Classic" pinout and that was standard for many boards at that time, or a 44 pin variant that looks very different, I can't say for sure that there wasn't a 3rd bootleg standard that differed slightly from Capcom Classic, only that it is more likely that whoever made your harness got it wrong. The fact you have no have no sound is what you would expect with your wiring too.

 

Pin 22 and Z should be ground, and 23 and "a" are the 12v outputs to the coin counters. A transistor on the board will switch 12V from the normal 12V supply onto the counter pins to increment one step, with the two counters connected between Counter 1 and ground, and counter 2 and ground, in these cases "ground" is common on the board to all "ground" pins.

 

counter 2 GND Z | 22 counter 1 GND

counter 2 a | 23 counter 1

 

So, unless your board has been modified then....

 

With the wiring you have your board will short out the 12V power rail, so I'm surprised the whole PSU doesn't shut down entirely, but you won't have any sound as the on-board amp is not powered up.

 

If that black connector labelled "Parts" is being connected to the GnG then the wires next to the "3

 

!!!!! Unless the board has been modded for some reason, damage to the PCB edge perhaps meaning they moved the 12V pins to other pins they didn't need to use anymore, and coin counters would be a good option for pins no one would miss.

 

All of this is best guess, without buzzing it through I can't say 100%, but as a shotgun approach it wouldn't be a bad idea to move the 12V feed to where it should be.

 

If I were you I would buzz through from pin 25 to the pins on the amp, it should connect to one of them, probably with quite a fat track.

Edited December 25, 2018 by Womble

Great, thanks for your reply and help on this. You certainly covered where the 12v should be and why it might not be.

 

Not sure which way to go on this.

I don't have a machine that can run this board now so I want to sell it, but I don't want to sell it to someone if its potentially wired wrong.

 

I might try and buzz pin 25 to the amp, as you suggested, but not knowing anything about PCBs, could you identify where on the 2 boards the amp is?

The amp chip is the one lying on its side by the A marking on my 1st photo in this thread, bottom left corner of the board. On some board it will be mounted to a heatsink.

I just checked all of my GnG bootlegs and pins 25/C and 22/Z and bridged on the underside. So the 12V can go to either one. I'm assuming your board will be the same.

I just checked all of my GnG bootlegs and pins 25/C and 22/Z and bridged on the underside. So the 12V can go to either one. I'm assuming your board will be the same.

 

Yep are correct there must be a bridge between 25/C and 22/Z because my MM gave me a buzz.

Thanks for your help

 

The amp chip is the one lying on its side by the A marking on my 1st photo in this thread, bottom left corner of the board. On some board it will be mounted to a heatsink.

 

Thanks for your help