- Thread starter djsheep
- Start date
A CPS1 core or a neogeo core would be awesome, but despite the rapid release of cores i think these are much further away. NeoGeo has a crap ton of IO apparently thanks to the insane cartridges. All of the other cores on MiSTer already had pretty good FPGA implementations already and were ported over to MiSTer. I'm not aware of any cores that started out on MiSTer that didn't come from somewhere else (but please point out if i am wrong here). I really hope to be proven wrong about CPS and neogeo 
There are still lots of inaccuracies on MiSTer, even though it's an FPGA, most of the cores are not developed using logic analyzers of the original hardware so they are basically low latency emulators of the original system behavior based on documentation on how things should behave and are not copies of the original hardware circuits. The genesis core has audio issues even with the cycle accurate 68k core (probably because audio is the Z80 and yamaha chips a lot of the time and needs to be both synced and accurate). The recent MiSTer SNES core was developed using BSNES documentation and the author stated he never had access to an actual SNES ever.
That said, the cycle accurate 68K CPU core by ijor is VERY significant, he has been asked his methods and he did use a logic analyzer, documentation and hires pictures of a decapped 68k. That's basically everything you need to "copy"/"clone"/"simulate" something to FPGA almost exactly right there.
If we start getting more chips done in that kind of quality MiSTer has the potential to become more accurate on multiple systems than anything else before.
We should hopefully see more and more of that kind of work as time goes on
PS: i really hope to be wrong on neogeo and CPS cores
There are still lots of inaccuracies on MiSTer, even though it's an FPGA, most of the cores are not developed using logic analyzers of the original hardware so they are basically low latency emulators of the original system behavior based on documentation on how things should behave and are not copies of the original hardware circuits. The genesis core has audio issues even with the cycle accurate 68k core (probably because audio is the Z80 and yamaha chips a lot of the time and needs to be both synced and accurate). The recent MiSTer SNES core was developed using BSNES documentation and the author stated he never had access to an actual SNES ever.
That said, the cycle accurate 68K CPU core by ijor is VERY significant, he has been asked his methods and he did use a logic analyzer, documentation and hires pictures of a decapped 68k. That's basically everything you need to "copy"/"clone"/"simulate" something to FPGA almost exactly right there.
If we start getting more chips done in that kind of quality MiSTer has the potential to become more accurate on multiple systems than anything else before.
We should hopefully see more and more of that kind of work as time goes on
PS: i really hope to be wrong on neogeo and CPS cores
I think its too optimistic to say that Mister is going to support CPS1 or CPS2 or other obscure arcade systems in a near future.. we are still not at the point where perfectly documented hardware systems are perfectly supported in a FPGA, so imagine how hard must be to support systems that were engineered with obfuscation as the main design principle.
That Mister or ArcadeFPGA support PCBs like Xevious, is already a lot
Anyway, 68000 is perfectly documented and is fully supported in FPGAs anyway.. look at Mininig, or Amiga or ST support on MiST o even the Amiga on MiSTer is pretty good already... and of course all the Amiga Vampire FPGA solutions.
That Mister or ArcadeFPGA support PCBs like Xevious, is already a lot
Anyway, 68000 is perfectly documented and is fully supported in FPGAs anyway.. look at Mininig, or Amiga or ST support on MiST o even the Amiga on MiSTer is pretty good already... and of course all the Amiga Vampire FPGA solutions.
What obfuscation is yet to be solved? All anti-piracy stuff is solved (decryption keys, PALs, etc.). Once 68000 core is solid, the big thing that is really needed is proper implementation of audio hardware (Z80's, Yamaha ICs, etc.). My guess is that the drive for a full Sega Genesis/MegaDrive implementation will get some of the audio stuff off the ground and then that research will flow back into the arcade stuff.
-ud
What do you mean that the 68000 core is not solid? I use it since a long time (years?)with the Amiga core and never had problems with it. I use nowadays more Amiga FPGA solutions than the real thing. And not only 68000 but also 68020.
Or am I missing something?
Or am I missing something?
Perhaps "solid" wasn't best wording, but situated in a portable fashion, such that it can be readily plugged into other 68000 based system cores (something other than Amiga, Atari-ST, etc.). I don't know the actual status in that regard, but it appears the existing 68000 based stuff are not sharing same core (being written by different authors). It would be nice if these building blocks could be shared in order to speed the process.
-ud
ShootTheCore
Legendary
I think it's the video hardware that's going to be tricky to accurately recreate for these FPGA solutions. An accurate 68000 CPU is definitely important, but that CPU was wide-spread among tons of 16-bit games - Amiga, Atari ST, Sharp X68000, Sega Genesis, Neo Geo, Macintosh, CPS1, CPS2, and countless other arcade boards.
Games don't look the same and run the same across all of those platforms because of drastic differences in the video hardware. It's also been video display processor implementation glitches that have taken months to iron out on the current Sega Genesis core.
Last edited:
SmokeMonster
Champion
ElectonAsh's CPS-1 is just getting started, so that is definitely a long way off. He is reverse-engineering from actual hardware though, so it will be good if it is ever completed. He has audio working more or less already and is relatively far along considering how early it is. He also just started porting Jaguar over.
A cycle-accurate 68000 soft core was released recently that will be making its way into lots of cores eventually, and it just got added to Genesis.
Neo Geo absolutely will happen, and Furrtek is actively working on that. He said that even though it may require more SDRAM (that is not confirmed though, it may be doable without), he will make it happen. I organized a Patreon raid a few weeks ago when he expressed interest in MiSTer and he got a ton of support and said that he's dedicated to making it happen now. So we may need a Neo Geo hardware add-on board, but I have no doubt that is in our future. How far into the future though is unknown. Remember that he's been decapping Neo Geo MVS/AES/CD hardware for a while now for preservation purposes, so when that core hits, it could be the most accurate core ever released from the sound of it.
A new scaler is the big project that Sorge and everyone is wrapped up in currently. Once completed, it should drastically reduce the HDMI latency, give tons of new options, add coefficient support to all cores, and is immensely easier/faster to compile (and it's open-souce). That will plug right into all of the current cores and is a direct replacement for the old scaler without much work. There are some preview builds of it in the NES core already.
Oof, I'm out of breath!
A cycle-accurate 68000 soft core was released recently that will be making its way into lots of cores eventually, and it just got added to Genesis.
Neo Geo absolutely will happen, and Furrtek is actively working on that. He said that even though it may require more SDRAM (that is not confirmed though, it may be doable without), he will make it happen. I organized a Patreon raid a few weeks ago when he expressed interest in MiSTer and he got a ton of support and said that he's dedicated to making it happen now. So we may need a Neo Geo hardware add-on board, but I have no doubt that is in our future. How far into the future though is unknown. Remember that he's been decapping Neo Geo MVS/AES/CD hardware for a while now for preservation purposes, so when that core hits, it could be the most accurate core ever released from the sound of it.
A new scaler is the big project that Sorge and everyone is wrapped up in currently. Once completed, it should drastically reduce the HDMI latency, give tons of new options, add coefficient support to all cores, and is immensely easier/faster to compile (and it's open-souce). That will plug right into all of the current cores and is a direct replacement for the old scaler without much work. There are some preview builds of it in the NES core already.
Oof, I'm out of breath!
pacoarcade
Professional
This YM2151 FPGA implementation pretends to be (or already is) cycle accurate: https://github.com/jotego/jt51
It's great to hear that there are ongoing efforts on porting CPS-1 and Neo Geo.
It's great to hear that there are ongoing efforts on porting CPS-1 and Neo Geo.
Raph_friend
Grand Master
Really thank for your work, i watched your live on My Life Gaming channel, but i can't find it anymore. I'm waiting for more news!
P.S.: (Só fui saber que era BR depois de olhar o twitter).
EDIT: It was off the list:
Last edited:
It's on a role - I love it
it looks like most sellers are out of stock of the IO boards until next year. Does anyone know where I can get one now so I can start developing a case?
oneleaf86
Grand Master
Hey TS,
I'll be building some IO boards soon. I have the PCB's and parts on the way. More than happy to help out and send you an IO board if you just cover the shipping.
Most Excellent.
I'll gladly cover the cost of the board as well just let me know
Thanks!
SNK-NEO-GEO
Enlightened
@oneleaf86 - Can you set me up with one of those USB Hub Boards? My understanding is that it is preferred to have external power source instead of taping into the main board mini..
USB HUB option 1&2
https://github.com/MiSTer-devel/Main_MiSTer/wiki/How-to-start-with-MiSTer
which one is recommended?
USB HUB option 3
tinkerplunk recomendes
https://tinkerplunk.com/usb-hub/
https://www.amazon.com/TENINYU-Adap...=1538038652&sr=8-12&keywords=5.5x2.1+splitter
Newbie MiSter questions for everyone that has one with the I/O, memory and mainboard..
the I/O board has a USB port, what is the used for? I connected a keyboard, a few USB controller and I get nothing from the I/O board USB port on the MiSter menu.. What am I missing? also, what type of controller does the MisTer detects over the USB HUB? any recommended cables for component out of the I/O analogue VGA thing?
Last edited:
oneleaf86
Grand Master
OK! So to answer a few of your questions,
The USB port on the IO Board isn't for connecting peripherals and is instead used as a serial connection.
The best option is to use the dedicated MiSTer USB hub and wire in +5V directly from the PSU. You can either solder a wire from the underside of the MiSTers DC Jack to the Hub or use a 2.1 x 5.5mm double adapter, cut one end and connect the +5V that way.
A powered Hub is recommended due the the on board micro USB being very low current. If your peripherals pull too much current they just won't work.
I'm building 20 USB Hubs so I'll have plenty spare
The picture attached is from the MiSTer wiki and shows where you solder in the external +5V to the Hub and the jumper you need to cut.
The USB port on the IO Board isn't for connecting peripherals and is instead used as a serial connection.
The best option is to use the dedicated MiSTer USB hub and wire in +5V directly from the PSU. You can either solder a wire from the underside of the MiSTers DC Jack to the Hub or use a 2.1 x 5.5mm double adapter, cut one end and connect the +5V that way.
A powered Hub is recommended due the the on board micro USB being very low current. If your peripherals pull too much current they just won't work.
I'm building 20 USB Hubs so I'll have plenty spare
The picture attached is from the MiSTer wiki and shows where you solder in the external +5V to the Hub and the jumper you need to cut.
Attachments
Last edited:
I can confirm this - for example some of the controllers i have (some 3rd party xbox360 ones) don't work without a powered hub. Another issue is if you use a wireless keyboard a lot of them just stop working randomly because of insufficient power to the dongle things.
I ended up using an old dlink powered usb2.0 hub i had
