Got a none working dark seal from @mysticmonk.
It was shown to have some damaged tracks and indeed there were two that were damaged (Marked in red) although there were others that looked like they might be damaged but turned out to be fine (marked in blue).
With that sorted I was just getting a static screen of garbage. No schematics for this one.
After checking stuff for a while (there are a few fujitus on this board but they checked out OK) I started to get the demo playing but I had no idea what I had done to cause the board to start running code. But even though the game was running it's attract mode there were some issues. The sound was intermittent or missing and the speech on the character bios screen sounded like the smurfs (got to admit the ninja saying 'I am a ninja' in a helium voice was funny). When the attract mode got to gameplay the PCB would crash to a blank screen.
Since I had been working on it for a while I left it at that but when I can back to it it wasn't running any code again.
After a while I realized that the board would run code only when it had been allowed to warm up for a while. I've known ram to change it's behavior at different temps so I decided it would be a good idea to check the CPU ram.
Unfortunately I couldn't find the CPU. Mame says it's running a 68000 but the only chip that might be the CPU didn't have the correct data and address pins.
Then I found a dark seal fix log on Jammarcade.net (where would be be without that font of knowledge) that told me that even though there's no schematics for this game it's close enough to other data east games and desert assault schematics were a good guide for this game.
So it turns out this PCB uses a custom chip marked with the number 59 (marked in green) which I assume is a data east 68000 but at least I could follow the logic to the two ram chips that the CPU used (marked in red).
Sliced them both and indeed the bottom one came up as bad. Swapped it out and it's now booting all the time instead of having to warm up and the gameplay worked in the attract mode.
Now for the sound issue. This one was a little strange.
Listening to the mame version I noticed it wasn't just the voices that was sped up but the music as well. So I figured it was a timing issue. There are a couple of LS163s after the 32.2Mhz crystal and one was a fujitsu but unfortunately they both checked out fine. As did the other logic,ram and roms in the sound circuit.
There are also two OKI 6295 chips (marked in yellow) and since I've known them to fail (recently replaced one in a vapor trail) I thought maybe the one responsible for speech is messing up the data or address bus so swapped it out (getting a little better at changing these surface mounts now) but that made no difference.
So the only thing left was the sound CPU. Looking at desert assault again it turns out it's a HuC6280A (marked in blue). It's a custom chip made for the CoreGrafx model of the PC engine console and the only way to obtain a replacement is from one of those consoles.
Dammit. I thought I was stuck on this one then I thought why not check the 32.2 Mhz crystal just in case. I mean it can't be that can it?
Scope said it was running at 55 Mhz.
WHAT!!!!
Could it be that simple?
I checked the 24 Mhz crystal just to make sure my scope was reading correctly and that came up as 24Mhz.
So I dug in my scrap PCBs and found a 32 Mhz crystal (close enough) and with that installed the sound is correct,even the missing sounds were back.
So glad I upgraded my oscilloscope from an old 20Mhz to a new 100Mhz one or I would have never known the crystal was out of spec.
Never heard of a crystal failing in such a way so it never occurred to me that the crystal could be at fault. Learned something useful.
Anyway never encountered this game back in the arcade days so looking forwards to giving it a run to see if everything is present and correct.
It was shown to have some damaged tracks and indeed there were two that were damaged (Marked in red) although there were others that looked like they might be damaged but turned out to be fine (marked in blue).
With that sorted I was just getting a static screen of garbage. No schematics for this one.
After checking stuff for a while (there are a few fujitus on this board but they checked out OK) I started to get the demo playing but I had no idea what I had done to cause the board to start running code. But even though the game was running it's attract mode there were some issues. The sound was intermittent or missing and the speech on the character bios screen sounded like the smurfs (got to admit the ninja saying 'I am a ninja' in a helium voice was funny). When the attract mode got to gameplay the PCB would crash to a blank screen.
Since I had been working on it for a while I left it at that but when I can back to it it wasn't running any code again.
After a while I realized that the board would run code only when it had been allowed to warm up for a while. I've known ram to change it's behavior at different temps so I decided it would be a good idea to check the CPU ram.
Unfortunately I couldn't find the CPU. Mame says it's running a 68000 but the only chip that might be the CPU didn't have the correct data and address pins.
Then I found a dark seal fix log on Jammarcade.net (where would be be without that font of knowledge) that told me that even though there's no schematics for this game it's close enough to other data east games and desert assault schematics were a good guide for this game.
So it turns out this PCB uses a custom chip marked with the number 59 (marked in green) which I assume is a data east 68000 but at least I could follow the logic to the two ram chips that the CPU used (marked in red).
Sliced them both and indeed the bottom one came up as bad. Swapped it out and it's now booting all the time instead of having to warm up and the gameplay worked in the attract mode.
Now for the sound issue. This one was a little strange.
Listening to the mame version I noticed it wasn't just the voices that was sped up but the music as well. So I figured it was a timing issue. There are a couple of LS163s after the 32.2Mhz crystal and one was a fujitsu but unfortunately they both checked out fine. As did the other logic,ram and roms in the sound circuit.
There are also two OKI 6295 chips (marked in yellow) and since I've known them to fail (recently replaced one in a vapor trail) I thought maybe the one responsible for speech is messing up the data or address bus so swapped it out (getting a little better at changing these surface mounts now) but that made no difference.
So the only thing left was the sound CPU. Looking at desert assault again it turns out it's a HuC6280A (marked in blue). It's a custom chip made for the CoreGrafx model of the PC engine console and the only way to obtain a replacement is from one of those consoles.
Dammit. I thought I was stuck on this one then I thought why not check the 32.2 Mhz crystal just in case. I mean it can't be that can it?
Scope said it was running at 55 Mhz.
WHAT!!!!
Could it be that simple?
I checked the 24 Mhz crystal just to make sure my scope was reading correctly and that came up as 24Mhz.
So I dug in my scrap PCBs and found a 32 Mhz crystal (close enough) and with that installed the sound is correct,even the missing sounds were back.
So glad I upgraded my oscilloscope from an old 20Mhz to a new 100Mhz one or I would have never known the crystal was out of spec.
Never heard of a crystal failing in such a way so it never occurred to me that the crystal could be at fault. Learned something useful.
Anyway never encountered this game back in the arcade days so looking forwards to giving it a run to see if everything is present and correct.
Last edited: