What's new
By:
Filters

The Definitive Guide to CPS2 Conversions

That is what I thought. By the time software programming got to 64mbits, they pobably found better ways to store data and moved away from EPROM's.

Alrighty, erasing some 27C322's as we speak. Hopefully this does the trick on a rev4 board. Which reminds me, the only two things missing on your guide is:
  1. What we just discussed
    • Doubling down from 8,192KB -> 4,096KB to burn onto 27C322 EPROM's
    • You have the doubling up part covered
  2. F-PAL to G-PAL board revision re-configuration and/or vice versa
    • Probably can link to Mitsurugi-w's video for the CPS2 multi
After this HSF2 is successful, will ask about converting from Phoenixed-to-MAME roms for 19xxJ. i still have the mask roms. But the main programs are all over the place. I suspect there is need to do doubling up.
 
I did mention the jumper wire but I could add a photo or link to Mitsu's video definitely

I haven't actually seen 64mbit ROMs on CPS2, HSF2 is the only game like that, pretty cool. I'll have to make a note of it

All main program ROMs are 4mbit, always 27C4096 type.

Because it's a Japan variant and the MAME parent is most likely Euro or US you may need to use some from the parent set
 
I just had a quick look at the 19xxj MAME set and it definitely has 4mbit main program ROMs:

19xj-03b.6a
19xj-04b.7a
19xj-05b.8a
19xj-06b.9a
19xj-07.6d

The layout of this ROM set is really strange
 
So here are the 19xxJ MAME roms for the Phoenix deconversion:
Code: 
File Slize - File Name
524288 - 19x-51.6a
524288 - 19x-52.7a
524288 - 19x-53.8a
524288 - 19x-54.9a
524288 - 19x-55.10a
524288 - 19x-56.11a
524288 - 19x-57.12a
524288 - 19x-58.13a
524288 - 19x-63.8d
524288 - 19x-64.9d
524288 - 19x-65.10d
524288 - 19x-66.11d
524288 - 19x-73.8j
524288 - 19x-74.9j
524288 - 19x-75.10j
524288 - 19x-76.11j
524288 - 19x-83.8m
524288 - 19x-84.9m
524288 - 19x-85.10m
524288 - 19x-86.11m
524288 - 19x-93.8p
524288 - 19x-94.9p
524288 - 19x-95.10p
524288 - 19x-96.11p
524288 - 19xj-03b.6a
524288 - 19xj-04b.7a
524288 - 19xj-05b.8a
524288 - 19xj-06b.9a
So from what you are saying, all I need to burn are:

524288 - 19xj-03b.6a -> bank 3
524288 - 19xj-04b.7a -> bank 4
524288 - 19xj-05b.8a -> bank 5
524288 - 19xj-06b.9a -> bank 6

Nothing else needs to be burned? If so, what are all the other ROM's?
 
You also need 19xj-07.6d for main program

If you already have the Q Sound and mask ROMs on board because it was running Phoenix code, you're good.
 
Let's get back to 19xxj later. Back to HSF2, which I am about to give up on. But I do think I have one more thing to explore which is my burn of BGSB3.


My experience with rev4 boards has never been good. I've used several rev4 boards in the past and have always had issues. Whether they be for conversions or the multi. Either something wrong with the sound, graphics, or both. In this case, it is both: artifacts in the graphics and the sound is awful but the game does boot. Here is what I've done on this rev4 board to HSF2:
  • Reburned everything with a SuperPro 610p
    • Banks 1-2 used 27C010's
    • Banks 3-10 used 27C4096's
    • Banks 11-20 used 27C322's
      Spiltting MAME roms from 8,192 KB -> 4,096 KB using HxD
  • Set jumpers to: SOOSSOOSOSOSSOSOSOSOSOSOSOSO
    • Checked jumper settings with a meter
  • Burned this BGSB3 to a Lattice GAL16V8B
  • Soldered jumper wire on the back of B-board from pin 9 on BGSB3-> CN5 pin 13
  • Installed Infikey as prescribed and set jumpers accordingly
I know there are successful uses of rev 4 boards for conversion or use with the multi. But darned if that applies to me and my tools (in the past, I've had the Top3000). The only thing I can think of is that the programmers I've used for burning the GAL's are at fault as from what I've read, it seems to be tricky to burn these things.


Before I decide to invest in a TL866II Plus strictly for burning GAL's (which I've read they do this well via USB power), I am going to pull the G-PAL from my known working CPS2 multi, and throw that into this rev4 board, then cross my fingers. If this rev4 board still gives me issues after that, well then I'll pull EPROM's and put them into a rev7 board and try the suspected faulty G-PAL and the known good G-PAL. I'll know for sure if I need to invest in the TL866II Plus.
 
BGSB3 is the Super SF2 PAL, if you've burnt that you won't get a working board.

You need to burn BGSB3G and connect the jumper wire to fix sound. Make sure the pin is removed from the socket or it won't work.
 
xodaraP said:
BGSB3 is the Super SF2 PAL, if you've burnt that you won't get a working board.
Click to expand...
I burned BGSB3G. I was refering to location silk screened in the back of the B-board.
xodaraP said:
You need to burn BGSB3G and connect the jumper wire to fix sound. Make sure the pin is removed from the socket or it won't work.
Click to expand...
Interesting. I've always tried to do the jumper wire from the back so it looks good per Mitsurugi-w's suggestion in his video. In his video, he never mentioned you have to lift pin 9 if jumpering from the back from BGSB3 pin 9 -> CN5 C13 on rev3/4 boards. I always thought that lifting pin 9 is only needed if you are soldering from that pin to CN3 C13 from the top side. As a matter of fact, he states if soldering from the back, "it is the cleanest way to go. There is no bent pins in the sockets. No cut traces............"

So what I did just now was lifted pin 9 out of the socket on the PAL. Soldered a jumper wire from that pin to CN3 C13 on the top of the board. Plugged it in and no more audio issues! I do still have a graphical glitch though which I am not sure how to solve. I even moved the board to another A-board and it carried over. My guess, the B-board is dirty so I'll pull everything and give it the Simple Green treatment. Graphics EPROM's are seated good since I just seated them. A & B boards are seated well too. Report back then.

So then last question for the night, if you have a rev3/4 board and are jumpering from the back. What else must be done if you want it to look clean from the top?
 
Mitsu used a factory conversion for his example of the bodge wire. His PDF instructions for the multi do say to remove the pin from the socket :)

You would need to cut the trace going from pin 9 of the PAL but I do not recommend this because I HATE destructive mods where they are not required. You can route the wire underneath the board so it's clean from the top and still remove the pin from the socket
 
I can see that now. All these years, that video has been instructing me incorrectly. That is why I've always had issues with rev4 boards. I'm definitely not the only one. My first B-board was sold to me after the previous owner said he couldn't get the audio to work for the multi; that is the same board I am using now for HSF2. Also, the PDF instruction points to the video. The video does clearly state that there are no cut traces, lifted pins, if doing the mod at 10:40 into the video. Perhaps @Mitsurugi-w you have made a comment on that wire mode in the video for rev3/4 boards?

Anyhoo, the remaining graphics glitches, I am hoping are the result of a dirty B-board. Will soak the CN connector holes with Simple Green, blow out all the standing water with a compressor, let it dry and report back if that fixes everything.
 
Could also be the mask ROMs are not in the right slots since they're split? Let me know how you go
 
Mask roms should be in the right place. I split them in HxD one by one and set HxD to prompt me for file naming. Then suffixed each file as A & B. As I burned, I installed them one by one. There is the possiblity I made a mistake and I can burn them again. No biggie. Before I do, I think I will throw them into a known working rev7 board with Infnikey installed. Reset the jumpers on the Infinikey and B-board. If the glitch follows the board, then I know it is not a dirty board. I will burn the patched set from ikotsu then place them in the rev4 board. If it plays clean, then I know my burns are bad and it is not a dirty board.

The graphics glitches are minor. Just boot and title screens. All other screens are fine. Perhaps someone knows what mask roms controls those screens.
 
When you say boot and title screens, which ones are you talking about?

The text only boot screen with the region is on the main program ROMs and then the main splash screen is on the mask ROMs - usually towards the end of the address space ROMs 19-20

Don’t burn the ikotsu main program, just mask ROMs. You can also compare your current burns to that set and see if they match.
 
xodaraP said:
When you say boot and title screens, which ones are you talking about?

The text only boot screen with the region is on the main program ROMs and then the main splash screen is on the mask ROMs - usually towards the end of the address space ROMs 19-20

Don’t burn the ikotsu main program, just mask ROMs. You can also compare your current burns to that set and see if they match.
Click to expand...
The text crawl was unimpacted. The anniversary logo screen had short flickering horizontal stripes across it. The splash screens of the different SF2 versions were garbled. But alas, all is well now.

Moved the EPROM's to a different B-board. Rev7, so no wire jumper wire needed. G-Pal is already there. Jumpers were already set. Reset the jumpers on the Infinikey for HSF2. Fired it up and through the attract mode, there was no graphical or audio issues. So let's wrap this one up and call it a success!

However, I did reburn mask roms 11-B, 15-A, and 15-B all from the set that I split myself in HxD because I didn't label the roms and got them mixed up when I pulled them. Also, I noticed that one of those EPROM's had a leg broken off. Don't know when this happened. But this must have happened after I burned them the first time and inserted them into the rev4 board. So that could have contributed to the issues I was seeing. Or it could be that that B-board's CN connectors are dirty.

I suppose now that all is working in the rev7 board, I can move the roms back into the rev4 board to eliminate the possibility that it was a broken EPROM leg that was the cause of the issue. But nah. It's kind of a bitch to remove EPROM's. HSF2 works now and I'll just leave it at that.

I'll give that rev4 B-board the Simple Green treatment this weekend and give it another wack and resurrect it as MSH, which I still have the OG roms for.
 
Was following your progress @acblunden2. I have a TL866II that you could have borrowed too. Good job on not giving up and finding success.
 
yuro said:
Was following your progress @acblunden2. I have a TL866II that you could have borrowed too. Good job on not giving up and finding success.
Click to expand...
Ohhhhhhhh, that may have been another reason why there was a graphics glitch on the rev4 board. Its possible my Xeltek Superpro 610P didn't burn the G-Pal properly. Let me burn another G-Pal and throw that into the HSF2 rev7 and report back. If it is still funky in what is now known to be a good working conversion of HSF2, I'll just invest in the TL866II Plus. I like how that is opened sourced anyway.
 
In HSF2 conversions, I noticed that the eprom 01 needs to be 12.75v, I used some NEC 27c1000 12v, NEC 27c1001 12.5v and others with the voltage table does not work well, in file 02 it can be used more in 01 only worked with AMD 27c010 12.75v.
 
Last edited:
No shortage of 27c010's over here. Salvaged like 20 of them off a dead Golden Tee '97, I think.
 
Wall-Garants said:
In HSF2 conversions, I noticed that the eprom 01 needs to be 12.75v, I used some NEC 27c1000 12v, NEC 27c1001 12.5v and others with the voltage table does not work well, in file 02 it can be used more in 01 only worked with AMD 27c010 12.75v.
Click to expand...
Makes no sense to me. You're talking about Vpp which is only used during programming. Or maybe you mean your programmer can only do 12.75V EPROMs.
A determining factor however is the chip speed (express in ns access time).
 
You must log in or register to reply here.
Back
Top