CV1000D Multi! NOT A BOOT CHAT THREAD - Formerly: A huge project is approaching....

[buffi]

Idk if I’m reading this correctly but the base config shows the RTC is also an eeprom and has a write port and clock. so it may not be something you can “just patch” without losing some function that it serves

https://github.com/mamedev/mame/blo...ea01346061087520f/src/mame/cave/cv1k.cpp#L437

Yes it is an eeprom and rtc in same package. Addressed at range starting at 0x10C00000.

Its also used for storing highscore and board settings for older games.

Related reading:
https://github.com/buffis/cv1k_research/tree/main/U13_Research#address-decoder

Communicating with it over jtag:
https://github.com/buffis/cv1k_research/blob/main/JTAG/RTC9701_JTAG.py

Datasheet can be found at
http://cave.buffis.com/#datasheets

Afaik cv1k-d games have it populated at u10 though?
Might be that the cpld doesnt allow rtc on them. No idea how that behaves on cv1k-d

 

[ekorz]

Ok so probably not hack time. But I could still use an sh2 wizard for something myself… if anyone’s out there!

 

[rewrite]

1. dealing with the lack of programming access to the sound ROMs, which I think kasaski's solution is quite clever

Sound roms go on the daughterboard, like the others. Similar to kasaki's in practice, but on a full PCB above instead of baby PCBs.

2. dealing with the lack of an RTC, it sounds like you have some concepts, some better than others. if this can be located on the sub-board that would probably be ideal.

The other one has an Epson chip per flashed game.

3. dealing with the lack of memory on -B boards for -D games, unless your solution is: only use -D boards as donors?

Perhaps my biggest assumption here (and I do know what assuming does) is that people are not going to be silly enough to use even the cheapest of real boards as donors when the boots are 1k or less. All the boots so far are D boards, so that seems simple enough. It just seems like no one would decide to use a 15-4500 board as a donor when an 8-1000 board is an option. Could be wrong here, but it's a silly approach if they did go for it, I think. Unless you have dozens of these laying around, at which point you probably have a D board anyway.

without losing some function that it serves

High scores on older games, and it sets the clock in the background of Ibara to match the time of day that the board has stored. Not quite sure what its function is in Pink Sweets, but I know the game won't boot without it. Actually thought I was way off base when I started because I was trying to do Pink Sweets Suicide Club and couldn't get it to work to save my life.

Any which way, we're on rev 1 here. Only 4 games unplayable and one's a romhack. From there there are two main goals:

- Get RTC games working
- Move all the game data to more/larger flash roms and a selector switch

 

[rewrite]

Afaik cv1k-d games have it populated at u10 though?
Might be that the cpld doesnt allow rtc on them. No idea how that behaves on cv1k-d

Sure do. Which I guess is a mystery, then?

What I do know for a fact is that the two games that use the clock setting in-game don't work on D boards as they sit,. What I've been told is Pink Sweets doesn't work for the same reason Ibara doesn't, from someone who has worked on a different multi. and the rest was gleaned from the page on Jammarcade that says the RTC chip is removed on D boards, etc etc.

https://jammarcade.net/cave-cv1000/

Any which way, all surmountable obstacles.

 

[buffi]

Jammarcade that says the RTC chip is removed on D boards

No thats not what it says. It says "Battery for the real-time clock removed" which is true.
The chip itself is still present. Just look at U10. I've sold my cv1k-d boards but I can clearly see a IC marked R9701 on my old pics, which is the marking used on RTC-9701.

It is true that the battery is not present though. I don't know if Ibara boots without battery attached to be honest. Maybe it doesnt? Too lazy to check now

 

[rewrite]

No thats not what it says. It says "Battery for the real-time clock removed" which is true.

And is followed up with "CV1000-D removed the RTC and battery, so high scores and game settings are stored in NAND instead of EPROM." a few lines later.

I may be a bit of a dummy at times, but I can read!

The battery question is a good one to figure out, I think!

 

[ShootTheCore]

And is followed up with "CV1000-D removed the RTC and battery, so high scores and game settings are stored in NAND instead of EPROM." a few lines later.

I may be a bit of a dummy at times, but I can read!

The battery question is a good one to figure out, I think!

Looks like the author of that article on JAMMArcade (me) made a mistake. I see that the RTC is indeed present on the CV1000-D hardware. I've updated the article.

If you catch any other mistakes, let me know!

 

[rewrite]

Looks like the author of that article on JAMMArcade (me) made a mistake. I see that the RTC is indeed present on the CV1000-D hardware. I've updated the article.

No worries! I honestly took it as "further information" at first, didn't read it as conflicting info. Like the first bit says that the battery for it was removed, and then further down it gave more info saying both were gone.

Glad it's fixed!

 

[kasaski]

I’m assuming D boards need to be the donors? Which would be quite expensive unless you use a boot (which is also expensive but not as much )

I hope people only use a bootleg for this . But yes, currently the intention is to use a donor D variant, but I want to experiment with 'upgrading' a b board's ram and seeing if games D games work that way. If that works out, i would imagine the modified b board would be the preferred donor as RTC would be fully functional.

Afaik cv1k-d games have it populated at u10 though?
Might be that the cpld doesnt allow rtc on them. No idea how that behaves on cv1k-d

Ive thought about this but havent tried it, but you may be right. 1000d might just simply have a differently configured CPLD that isnt configured for clock access, so a swap with a b boards cpld could solve our issues. But given im using a bootleg i had already assumed they were just using your replacement bitstream which is based on the 1000b's cpld. If they arent then i guess its a testament to the accuracy of those bootlegs .

 

[advans14]

I guess ill let the cat out of the bag, I funnily enough started looking into a solution sacrificing a bootleg just prior to rewrite making this thread and weve been going back and fourth in dms. My idea works as a quick-swappable pair of cartridges. One cart for sound another for the prog/graphics.. This works as POC but assembly is very error prone and quite difficult, not a great product overall. Connectors are expensive and aren't rated for a large amount of connects and disconnects either. awesome for testing but not something im proud to release.

Regarding RTC game issues, NFI. But i hope once the ball gets rolling those more educated with software REing sh3 code can find some workaround to skip whatever part prevents them booting. Ghidra wasnt playing nice for me when dissasembling prog roms and missed large chunks of functions.

That is cool! I could definitely see this connector getting trashed easily though by folks. Probably best to make a daughterboard with enough flash to cover it all.

 

[buffi]

But given im using a bootleg i had already assumed they were just using your replacement bitstream which is based on the 1000b's cpld. If they arent then i guess its a testament to the accuracy of those bootlegs

I suspect that if they have resources to reverse the gerbers, getting the bitstream out of the cpld is probably not too hard for them.

The custom one i made isnt really thoroughly tested so i sortof doubt its what they are using

 

[kasaski]

TSOP48 sockets arrived today so I've been able to continue experimenting with my last few daughter boards.
I can confirm swapping the CPLD from a cv1000b board seems to allow the RTC games to run on a cv1000d, tested with ibara black label. Ive also tested SDOJ and Akai katana after the swap and both appear to still work as normal. the last hurdle seems to be solved

 

[rewrite]

Sounds like your work's better than you thought it was @buffi !

Excellent sleuth-work @kasaski !!

 

[Dloop]

So are these available somehow without sacrificing a b board? Sorry if dumb question or are you reprogramming the cpld somehow?

 

[rewrite]

So are these available somehow without sacrificing a b board?

The CPLD? Absolutely. Just pop the part number in google and away you go.

epm7032aetc44-10n

 

[mR_CaESaR]

How is this project even possible? How are parts being sourced? I always thought a big chunk of stuff are no longer available etc... Did the Gerbers from the bootleggers put it out in the open market which then allowed the development of the multi?

Either way, I look forward to seeing this project develop.

 

[kasaski]

How is this project even possible? How are parts being sourced? I always thought a big chunk of stuff are no longer available etc... Did the Gerbers from the bootleggers put it out in the open market which then allowed the development of the multi?

Either way, I look forward to seeing this project develop.

You can substitute all the flash with other parts, and patch u4 to remove the check on the U2 flash, basically allowing you to use any similar spec flash chip. For the cpld, thanks to buffi's work you can replace them with alternative chips now. Surprisingly simple board with almost no protection

 

[brandon.arnold]

@kasaski @buffi @mR_CaESaR may I ask for more information about who was able to pull the bitstream off the EPM7032? I am looking to do the same with an EMP7064 and would love to pick someone's brain about it.


Update: I see, @buffi has a lot of context on the EPM7032 in U13_Research on that cave repo. Nice!

 

[mR_CaESaR]

I definitely don't know. I wouldn't have a clue. 99.9% of the stuff being mentioned is another language to me

 

[rewrite]

I always thought a big chunk of stuff are no longer available etc.

You can buy everything that's on a CV1000 board still. The SH3, the RTC, the flash roms, the alteras, etc etc etc. @buffi and @alamone 's works have made it so you can easily replace some of the chips, even with ones from other manufacturers.

Which leaves us at a point where there's nothing in the way of converting from one game to another. With the exception of the RTC games on D board hardware, which @kasaki just solved using @buffi 's bitstream.

From there it becomes: How do we get the game data off of the main board and onto a sub-board? (we have succeeded at this, both @kasaski and I in different approaches).

Then now that the data is on the sub-board, how do we select between multiple sets of data? Which shouldn't be terribly hard as it's just selecting different chips and different banks on those chips with a rotary dipswitch.

I'm not sure what the answer is to the RTC business as far as a multi is concerned. Put 16 RTC chips on the multi? There's a TON of room, I suppose you could, but it seems a silly solution. Maybe there's something more elegant?


Now if you want to know why no one's done this before, well, they have. Several parties. But they're all either for profit and thus can't be out and about for risk of exposure and lawsuits, or they're one-offs or two-offs for self and friends.