Sharp X68000

[icy shuttle]

But again, the black screen with osd screams bios problems

When you boot the core, does it try load boot.rom

[sand girder]

It doesn’t seem to try to load boot.rom. I would see characters on the screen, and not a blank screen, correct?

[icy shuttle]

yeah correct. you'd see a load bar like when you load a rom in another core

do you have the bios database enabled on update all?

[sand girder]

I’m not sure if I have it enabled. How would I check? Also I just noticed my boot files seem to be the wrong size: 174 kb! Let me reinstall them.

[icy shuttle]

here instead of doing that

delete boot.rom and boot.3vhd

run update all and go into the settings (up after it loads)

in there there is an extra files or something like that menu option

the bios database is in there

enable that and it will grab the correct bios for the 68k

[sand girder]

It’s working!!!! Thank you so much!

[icy shuttle]

if you haven't already, check out neon68k.com

[sand girder]

I have. It’s a great site. Thanks for it.

I love Amigavision and I’m glad you made a similar thing for the X68000.

[primal sandal]

Ok... I'm just starting to look at this now, and I can't figure out how to do stuff on X68000

Before I try to play a game, I want to be able to interact with the system itself

So when I boot the core, I get:

Human68k for X68000 version 1.00
Copyright 1987 for SHARP/Hudson

I am given the impression that this is a boot ROM, rather than a disk operating system

Pressing F9, I get a prompt to enter a command, but either "dir" or "list" will bring back a Japanese response stating that it can't find the command.

I feel like I must be missing something....

It took me a while to find out that I needed to press F9 to even get to that point

Anyway, I'm going to need to know at least a little bit about the system if I'm going to do any dev for the core

I appreciate the streamlining that is done in Neon68K for newbies, but before I even load a single game, I want to figure out what they've done, and what the "standard" access mode should be (and if I'm missing anything)

In my games/X68000 directory, I currently have only:

boot.rom
boot3.vhd
BLANK_disk_X68000.D88

BTW, the boot3.vhd looks way too small (16KB)

[icy shuttle]

That’s probably correct. The OS is on the floppy disk/hdf. If you grab DiskExplorer, you can open up the D88s and hdfs and see the humanoid files. I wonder if the blank disk has the OS on it

[primal sandal]

Is DiskExplorer a ROM ?

[icy shuttle]

no sorry, looking for the link

[primal sandal]

"Blank Disk" has some stuff on it - at least "Human68k for X680x0 version 3.02
Copyright 1987,88,89,90,91,92,93 SHARP/Hudson"

Looks kind of like a boot disk

But I feel like the bootup brings me to a screen where I have to press a magical key combination in order to get out and do something useful

Hmm... F10 brings me to a prompt, but I probably need a reference manual because this isn't a standard operating system of any sort

[robust ore]

The primary operating system for the X68000 is Human68k, which is generally similar to DOS. Common commands such as DIR, CHDIR (CD), and TYPE are the same. Executable files are denoted by the extensions .x and .r, while batch programs retain the common .bat extension. Luckily, the X68000 uses a flat memory model, so there's no need to fuss with traditional DOS headaches like conventional memory vs. XMS/EMS.

However, there is one key difference to keep in mind: drive letters are assigned according to the boot device. If you boot from a floppy disk, the floppy drives become drives A: and B:, while the hard disk drive becomes C:. Conversely, if you boot from the hard disk, the hard disk becomes A:, while the floppy drives become B: and C:.

Almost all game software will run directly from floppy disks with no need to install. Otherwise, boot up Human68k first, then locate the main program from the command line or a filer utility.

[primal sandal]

yeah, but I need to hit either F9 or F10 first, to get to the point where the keys are displayed on screen, and in both cases, DIR provides an error message

[robust ore]

On my original system, I'm using a floppy with Human68k 2.0.2 and SX-Window (for the GUI).

Could this help you ?

[icy shuttle]

I cannot find the download link for this anymore, but this is what we used to assemble neon68k images. It lets you open up and extract the contents of D88s and HDFs

iirc, there is a HumanOS folder

[primal sandal]

Well, this boots up into something that calls itself a terminal

X68000 Terminal V1.0

F9 is "Command" and F10 is "End"

so it seems that it's not booting just to an OS

Anyway, if I can ever get to the starting point on this core, I wanted to add the XE-1AP to it. It seems like the joystick wires are pretty easy to access, unlike the complex world of Megadrive (etc.)

I'd also like to get a Zuiki X68000Z keyboard and make sure that's got a proper keymapping available (not sure if it's already in there)

[icy shuttle]

they should be with the unstable. it added the keyboard option in the menu

[robust ore]

Does the Human68k User manual (in Japanese) could help ?

[primal sandal]

There are keyboard options, but it's not clear if they cover that keyboard

[primal sandal]

Yes, I'm 90% certain it would

[robust ore]

Give me a moment.

[primal sandal]

In any case, I don't have such a keyboard at the moment, but I am planning to buy one

[icy shuttle]

I'm not excited yet........but I'm getting pre-excited 😄

[robust ore]

Human68k is public domain, so okay I could share some documentations here.

[primal sandal]

Thank you, I will study this

OK, I'm not sure what I did, but I must have done something - the reboot sequence is much much longer, with a lot more drivers being loaded

Oh, and DIR works !

There's some video corruption when it scrolls though

some junk

Ok, anyway, It's doing something better now, and I have some reading to do

Thanks

[primal sandal]

I've noticed that this core doesn't have turbo, but the Zuiki controller for the modern X68000Z does.

Was that present on old controllers too ?

(or original)

[icy shuttle]

I think on third party controllers only?

[primal sandal]

But I guess I mean - games benefitted from it ?

[icy shuttle]

Yeah I think so. Would be good for shooters

[thorny tree]

for controllers X68000, core misses 3 buttons and 6 buttons also

[primal sandal]

I've found some information on those. I think I'll try to add 6-button

[thorny tree]

For Games SF2 uses 6 buttons and Chelnov uses 3 buttons thanks !

[primal sandal]

Well, I found info about 6-button, but I don't know how 3-button is different

https://gamesx.com/wiki/doku.php?id=controls:x686button

And these are the games I saw (listed in a post)

https://nfggames.com/forum2/index.php?topic=5544.0

[thorny tree]

https://gamesx.com/wiki/doku.php?id=x68000:game_notes

6 buttons seems to do everything

[primal sandal]

Hmm... looks like the joystick strobe line doesn't work properly

I thought I would just be connecting up a couple of things

[primal sandal]

Well, it was pretty easy to get turbo working, but the 6-buttons depends on getting the strobe working, and that's not working (probably why it wasn't connected from 8255.vhd through X68K_top.vhd to X68000.sv).
That's proving to be non-trivial (especially since I don't know anything about the system or that device)
Once that's working, though, XE-1AP should also be trivial

[primal sandal]

Hmm... Looks like Garou Densetsu doesn't actually support 6-button so I wasted hours in testing....

Street Fighter is showing some signs of life, but the mappings are going to be a discussion I think

[thorny tree]

It doesn't use the one of Megadrive ?

[primal sandal]

Well, I've set up the CPS format as stated in the article

I seem to recall that Megadrive 6-button had some weird delay

between scans or something like that

And Chelnov isn't in the neon68K set

There seems to be extremely little (and poor) documentation about which games support which controllers

I can only find statements made by people which are unverified

Currently I have:

• turbo working (but only tested on Galaga, which senses both triggers for the same function)
• 6-button mostly working, but the mappings are not quite right yet (I think that circuit has a few things wrong)
• XE-1AP roughed-in and compiling, but not yet functional

I will need help to test whether any of this is working properly, and to identify the behaviour of games and whether they support such controls
-> NOT the intense type of testing I often see, where people are trying millions of random situations to see if they can find edge-case problems

Ideally, the help will come from somebody familiar the games being tested, and who can verify against original hardware

If somebody can verify claims about which games support which types of equiipment, that would be fantastic for a start

[thorny tree]

yes not well documented Chelnov have an adapter to connect MD controller provided with the game (maybe usable for 6 buttons but not sure)

[primal sandal]

I believe 3-button was just multiplexed, but 6-button had a delayed second scan

Hmm... maybe not. It seems that 6-button Megadrive just requires several cycles of read

https://www.raspberryfield.life/2019/03/25/sega-mega-drive-genesis-6-button-xyz-controller/

toggle SEL multiple times

[thorny tree]

Pac Land also comes with his own controller

[primal sandal]

OK, I'll make the controller 3 bits.

regular, turbo-2, Capcom 6, Megadrive 6, CyberStick/XE-1AP, PacLand

Still have room for two more

But for now, I'm only trying to implement regular/turbo-2/Capcom 6, and rough in for CyberStick

Probably need help for the others

Even CyberStick is proving to be difficult to test. I need a test program

[thorny tree]

CyberStick seems huge to implement ...

[primal sandal]

No, I did it for PC Engine - it's already roughed-in (copied the module), just needs testing. But I've got some potential concerns with the 8255 and some hacks I made to it while testing with Garou Densetsu yesterday

As soon as I am comfortable with getting the 6-button Capcom mappings right, I will try to restore the 8255 code I changed yesterday, and retest XE-1AP

I "sort of" got it working, but I need to hold down another unexpected key, and the axes are inverted...

So it's "partial" right now

[thorny tree]

For CyberStick may be after Burner 2 could help to test.
It doesn't start with regular joypad.
For the moment the only way is to use button of the mouse to enter the game.

[primal sandal]

But the XE-1AP has many buttons, and I didn't implement all of the joystick buttons into the protocol - only the ones that were meaningful to the PC Engine games.

so this will depend on the game

[thorny tree]

Input configuration on After Burner II available Configuration -> Inputs

[primal sandal]

I've got After Burner but not After Burner 2

[thorny tree]

Maybe it is the same game one thing I found we need to strike F11 then start on the cyberstick

[primal sandal]

I can send you a test core in a few minutes. It will be rough though

[thorny tree]

When I strike F11 on After Burner (II) on title screen i couldn't able to use mouse and joypad so It may activate cyberstick mode

[primal sandal]

interesting

[primal sandal]

OK, I rolled back the unnecassary changes, and making a Pull Request now

New features will be:

• Turbo 2-button controller (done in the same way PC Engine's 2-button turbo is implemented)
• MegaDrive 3-button controller
• Capcom 6-button controller
• XE-1AP back-end has been roughed-in, but is disabled for now

Games supported:
MegaDrive 3-button:

• Chelnov
• not clear to me what other games -> need list

Capcom 6-button

• Super Street Fighter II - The New Challengers
• not clear to me what other games -> need list

#test-builds message

[thorny tree]

thank you 🙂

[thorny tree]

For Capcom 6-button there Super Street Fighter II - The New Challengers and Street Fighter II' - Champion Edition
I need to test also SFXVI (doujin vs game)

[icy shuttle]

don't use headphones :p

[thorny tree]

yes Super Street Fighter II has sound issue on Internal at least

[icy shuttle]

yeah, I am pretty sure the game is totally playable but we put it in major bugs for neon68k because we didn't want to be liable for anyone's hearing loss

[thorny tree]

SF2' working pretty great with 6 buttons
and also we can now turn around on Chelnov 🙂

[primal sandal]

🙂

Let's see if we can create lists of games which are compatible with each

I understand that the total number of 6-button compatible games is less than 10... and that includes both MegaDrive-6-button and Capcom

@south estuary : thanks for this page, it was very helpful, but the wiring on the Capcom 6-button was wrong in a few places (as you had expected)

https://gamesx.com/wiki/doku.php?id=controls:x686button

I can walk you through how the 74157 diagram should change, if you're interested in updating it

By the way, about 6-button games , there is a list here, but it's not exactly correct:
https://nfggames.com/forum2/index.php?topic=5544.0

At least Garou Densetsu doesn't support Capcom-6-button

(Maybe it supports Genesis 6 though)

Yes, probably it does. You get a couple of extra controls on Megadrive-3 on that game

[thorny tree]

With Megadrive 3 the third button is recognized on Garou

[primal sandal]

And Start as well

[thorny tree]

With Midi the issue is present ?

[thorny tree]

yes confirmed the start button is 4th button

[icy shuttle]

is midi working on street fighter?

I couldn't get it to trigger

[thorny tree]

for SSF2 it is on the option

you change from INTERNAL to MIDI

[icy shuttle]

oh i think that crashed it. let me look right now

[thorny tree]

but I have issue with my IO board so I couldn't test

[thorny tree]

ok

[icy shuttle]

Modem sounds intermixed with voice samples saying random things.

[thorny tree]

so no crash but not good for sound either

great port the game is beautiful

[icy shuttle]

really is

and controller works great

[primal sandal]

I'll probably be adding Megadrive 6-button sometime soon, and I want to get XE-1AP working too.

[thorny tree]

The controller of Garou Densetsu 2
https://www.ebay.com/itm/225533431040

[primal sandal]

Interesting. But too rich for me (and I don't even have a X68000)

But if somebody were to open it up, it could be implemented

[verbal jolt]

does garou densetsu 2 work on the core? i recall garou densetsu special having major issues

[primal sandal]

It is almost certainly Genesis 6-button with a different button layout

[thorny tree]

No it has heavy graphical issues for the moment

[verbal jolt]

ah i see

[thorny tree]

just the 1st one is good to play

and works now with 4 buttons

[primal sandal]

This one's cheaper:

https://www.ebay.com/itm/326699824779?_trkparms=amclksrc%3DITM%26aid%3D1110006%26algo%3DHOMESPLICE.SIM%26ao%3D1%26asc%3D289600%2C288653%2C289624%26meid%3Df526f0e21d0c43b78833eaa6bd244d6e%26pid%3D101224%26rk%3D3%26rkt%3D5%26sd%3D225533431040%26itm%3D326699824779%26pmt%3D0%26noa%3D1%26pg%3D2332490%26algv%3DDefaultOrganicWebV9BertRefreshRankerWithCassiniEmbRecall&_trksid=p2332490.c101224.m-1

[thorny tree]

it seems better condition also

For Garou 2 start button (Megadrive 3) pause the game so different behaviour.

[primal sandal]

Wow, there are some bizarre controlelrs for this machine

there's a special one for Cameltry

2-Dpad for crazy climber - another (maybe the same) for Libble Rabble

[icy shuttle]

Arcade Perfect ™

[robust ore]

The game have some DMA issues.

[verbal jolt]

i think a lot of games do no?

[icy shuttle]

yes

Rebel did a revert to a 2022 change that fixed it in some things

but it caused a bunch of regressions so it seemed like it was a one step forward one step back thing

[verbal jolt]

unfortunate 😔

[primal sandal]

Going to need srg320 on this at some point

[icy shuttle]

Oh no, don't give him another reason to not work on 3DO

[thorny tree]

The special is little better in game (but more ugly on menus)
Megadrive 3 => Same behaviour of Garou 2
So Garou 1 may use different mapping

to fix all issues core will be huge task (memory issues, floppies format, support hds (hdd), graphical and sound issues and some games hangs)

[verbal jolt]

it would be worth it to finally have a good fpga option for x68k games

the real hardware is just prohibitively expensive

[south estuary]

It's just a rotating disc to spin the standard mouse, not a whole controller.

[primal sandal]

ah cool. I don't have to worry about that then

Libble Rabble looks like it might be the same as Crazy Climber

[south estuary]

And yes the two micomsoft pads for crazy climber and libble rabble are the same.

[primal sandal]

cool

[south estuary]

Not sure if I still have mine, pretty sure I don't. May have pics tho.

Internals etc

[primal sandal]

I saw a page with a schematic - not sure whether you traced that or what

https://gamesx.com/wiki/doku.php?id=controls:x68crazyclimber

[south estuary]

Ah yeah that's mine, I thought I'd done that. 😊

[primal sandal]

The 6-button needs some updates though
https://gamesx.com/wiki/doku.php?id=controls:x686button

[south estuary]

A lot of that was not mine, by the look of it.

[primal sandal]

Yeah, not sure - I don't have the tools to update it, but I can leave some notes here if you like

[south estuary]

I still have a couple of the X68000 fighting stick / mega drive adaptors. They're my retirement fund. 😊

Yeah please.

[primal sandal]

OK, top 74157:

• pin 3 ("select") should actually read "run"
• pin 13 ("R") should actually read "L"

bottom 74157:
pin2 ("L") should read "trigger A"
pin 3 ("A") should read "R"
pin 5 ("start") should read "trigger B"
pin 6 ("B") should read "Sel"

It had me going around in circles trying to get SF2 to run properly

Especially with the 3 or 4 layers of indirection in the MiSTer framework that I already had straight (but got confused about)

[south estuary]

ugh, sorry. I'll fix it.

[primal sandal]

It's OK. I learned a bunch

[south estuary]

OK, schematic redrawn with corrections.

[verbal abyss]

Apologies for the offtopic post (I'm out walking and can't find the answer on discords search) does the 0mhz dos pack work with dv1, and play nicely with tink4k?

[thorny tree]

You have this issue https://github.com/MiSTer-devel/X68000_MiSTer/issues/41

[verbal abyss]

Oh I see thanks for the update

[icy shuttle]

https://neon68k.com/docs#x68000-and-external-scalers

[tawdry creek]

Weren't they asking about the ao486 core though?

[tawdry creek]

You'd want the #1047332497492553799 channel for this question

But also (it doesn't support dv1 but does work with the RT4K)
https://consolemods.org/wiki/AV:RetroTINK-4K/System_Specific_Settings#AO486

[icy shuttle]

I think you still want those resolutions

I think?

[primal sandal]

@thorny tree , I see that you ported some Keyboard layout updates from Puu's code back in February. Do you know whether those include the Zuiki X68000Z keyboard ? Or just generic Japanese keyboards (or both) ?
...or is it not clear to you

[thorny tree]

I don't think it uses x68000z keyboard.
https://github.com/MiSTer-devel/X68000_MiSTer?tab=readme-ov-file#keyboard-layouts
They're is 4 layouts 2 JP and 2 US for common keyboards to have access to all keys of x68000 keys.

[primal sandal]

OK, I've ordered a X68000Z keyboard, and I'll see.

[thorny tree]

If it is working for sure I will order one also 👍
How many keys have the x68000z keyboard?

[primal sandal]

A lot...

And it's expensive

https://www.amazon.co.jp/-/en/ZUIKI-X68000-Z-Keyboard-Black/dp/B0C76ZMWJJ/ref=sr_1_1?crid=1NXS41NQ1OYS1&dib=eyJ2IjoiMSJ9.l7-yh5EGG10hj7JBFzp0nZIX3_Xu15jHVW-bJdm51pjKou3kOnWGU0Za8CYgxZFIukWNWHeCUD4mLPxWfim57XGeVI1pJKUSIJeAwUGuJv7mgSLsEtf0l9lmNUPEHsXJ9_xjDdvTv5KFP81rv_0OqvKUCizA5tzbRofynopc5wrriR3qj9v2oOtKFirVEOTGEMj1vlOoZ5Mauyy9t8b-C1J40Fe8jmXPHTRRoCIPfcSctQlwRJG8S9_Cv89dZynU8KgZTDoddMHbEaFff7nOWIn6pHMn8cwR3hd06bMMZHQ.iMW3zKnLRz2Pq-L3c0TcHvw4PNUtLKasNo5K2MwqlKQ&dib_tag=se&keywords=x68000z+キーボード&qid=1753761129&sprefix=X68000Z%2Caps%2C138&sr=8-1

There are also LEDs on a lot of keys, and printed symbols (which look helpful)

[primal sandal]

If it is not 100% working, I will work to make it work

Amazon also has the Zuiki Mouse, which has a trackball inside. I'm a bit curious about how the system reads this data

https://www.amazon.co.jp/-/en/ZUIKI-X68000-Z-Mouse-Black/dp/B0C771BCJ9/ref=pd_bxgy_thbs_d_sccl_1/356-3010651-2026764?pd_rd_w=Vn3zi&content-id=amzn1.sym.9d5616ec-df84-41cc-8b88-11efeb16ec58&pf_rd_p=9d5616ec-df84-41cc-8b88-11efeb16ec58&pf_rd_r=GK0XV5THAB42Y31S89PA&pd_rd_wg=f69K7&pd_rd_r=5c2d291a-031f-4581-aa9f-d79de8aa2825&pd_rd_i=B0C771BCJ9&psc=1

I haven't decided to get one yet... but sometime

Oh, it's optical, with a trackball !

[thorny tree]

We may need to check if the Linux kernel already support them

[primal sandal]

Sure. When we get close to that point

[primal sandal]

Fedex said my X68000Z keyboard would arrive next Tuesday, but it's already in Memphis (originally Japan), so maybe I'll get it before the weekend (long weekend here).

[icy shuttle]

If it’s fedex it will be there on Tuesday, if not later. I have no idea how they stay in business.

[primal sandal]

It's different internationally than in the USA

From the way it looks, it ought to be tomorrow, but it'll probably be Friday

Anyway, when it arrives, I'll check whether it's already accommodated in the core, and if it isn't, I'll see what might be needed to accommodate

[icy shuttle]

Excited. I am excited to see hear how it feels too. I’ve been eyeing importing the accessories…

[primal sandal]

Yeah, mouse is hard to find

[thorny tree]

10K Yen for shipping and fees to deliver for my place (if i ordered I need to see if i took also the mouse or not because the price for shipping and fee is kind the same) 😅

[primal sandal]

Keyboard arrived

That's a large keyboard.

I will have to wait until after work today to actually try it out

[primal sandal]

So, comparing the keyboard layout of the Zuiki keyboard and the key scancodes against the values listed in the repository for the 4 currently-supported keyboard types.... it's very different

I feel like I need to review what values Puu had actually been tracking in order to make sense of it

Because his "North American" keyboards don't seem to match actual North American keyboards

[thorny tree]

I recall someone have issue with keyboard not US and not JP
He was unable to use some keys

[primal sandal]

And by the way, the switch on the back of the keyboard (Win/X68) seems to affect the scan codes for a small groups of keys - most of them are the same in both modes; only about 9 or so move position

Do you know of any X68 software to check which keys are pressed ?

I could probably write my own based on technical documents like in "Inside X68000", but it's better (and faster) if something already exists

[thorny tree]

I don't know for the moment I will check

[primal sandal]

Even a BASIC program would be good.

[lilac meteor]

found something here : https://github.com/kg68k/xperiment68k -- keyscan one. Didn't tried it yet though, will see hopefully over the weekend to get those compiled

[thorny tree]

for cross compiling
https://nfggames.com/forum2/index.php?topic=4850.msg46342#msg46342

[primal sandal]

It's going to be a while before I'm comfortable enough with how to operate the system that I can compile something.

BTw, looks like the SNK "Magic" controller is slightly different again from Capcom and MegaDrive

So... are there really any games compatible with the 6-button MegaDrive controller ?

I know there are games compatible with:

• 3-button Megadrive
• 6 -button Capcom (8-button actually)
• SNK "Magic" controller

I should rephrase that - I know Street Fighter II is supposed to support 6-button MegaDrive controller. But the Capcom 6-button is sufficient to satisfy that game. So, other than the Street Fighter games, are there any game which support MegaDrive 6-button ?

[thorny tree]

Maybe some homebrews ... like https://www.youtube.com/watch?v=Rv9r4YiKCss it doesn't work on the core currently but it is VS game like Mugen for X68000
@lethal scroll Do you know any games that could use 6 button megadrive that not be SSF2 and SF2' ?

[primal sandal]

Basically, I'm trying to figure out whether it's even worthwhile to worry about MegaDrive 6-button

[primal sandal]

Also, I saw a note from neko68K several years ago on NFG's forums that the Fatal Fury 2 "Magical" controller uses the saem multiplexing as the Capcom controller, but forces a couple of buttons to the on state as identification. I am assuming that they are the corresponding bits to "up/down" or the "left/right" pair... but which one ?

[thorny tree]

it could be added later if we find any game using it.

[verbal jolt]

this is pretty neat. is it known what's keeping it from working on the core?

[elfin slate]

it is essential that we all get to enjoy this video game on our MiSTers

[thorny tree]

Most issues issues are error during execution or loading the game and the game crashes. I don't know the root cause.

[thorny tree]

This doujin VS game "Jotei Senki 4 - Heroines Forever" recognised the megadrive 3 button but not Capcom 6 buttons.
Checking the file of one fighter there is those buttons used
[Remap]
x = x
y = y
z = z
a = a
b = b
c = c
s = s
This series have 4 games. So it could be interresting to test with MD 6 buttons

[primal sandal]

I'll see if I can get the obvious controllers working first. I don't have a real X68000, so I can't test functionality to confirm whether (or how) something works on original hardware

[thorny tree]

On Aquales with Turbo 2button activated all 6 buttons on my M30 pad do a thing

[primal sandal]

I don't understand what that means

Basically, 2 button turbo is the exact same thing as 2 button default (if you don't touch the additional buttons)

[thorny tree]

The behaviour is the same for the first two buttons but if we set turbo 2 on this game others buttons 3 to 6 respond for the action of button 1 or 2.
In case you set button 2 buttons 3 to 6 don't do anything.

[primal sandal]

Yes, that's how it was supposed to work

Turbo is auto-repeat

So, button 1 is no repeat, but button 3 is slow repeat of button 1, and button 6 is fast repeat (A->X->R)
same thing for button 2, button 4, and button 5 (B->Y->L)

[thorny tree]

Ok thank you for the explanation

[primal sandal]

I'm making progress on XE-1AP support. It's gathering data now, but the values are wrong somehow.

I found a program in MICOM BASIC magazine, issue 1990-10, pages 81-82, which requires assembler...

I wonder if somebody might be able to help with that ?

I'm OK to type it in and send it in a textfile if that helps, but I don't I have the tools for assembling the output and making it run

[rotund parrot]

here's an assembler for X68000. You might need a linker too which I will also link
https://github.com/kg68k/has060xx/releases/tag/v1.2.0

https://github.com/kg68k/hlkx/releases/tag/v1.1.0

You'll need a way to copy the binaries to the image you're working with

[primal sandal]

Well, that would be useful if I knew how to use the machine in the first place

I seriously don't have one, nor know how to operate it. It's not like DOS

I wouldn't ask for help if I was self-sufficient

[rotund parrot]

No problem, I can assemble it if you send the source

[primal sandal]

OK thanks, I will type it in

[primal sandal]

Hmm... having a bit of trouble entering an underscore key

[primal sandal]

I can get an overscore, but not an underscore

[primal sandal]

OK, I think there is no underscore character available on the keyboard. I typed in the BASIC part of the listing and sved it to disk; then I had to edit the disk image

Do you have a hard disk image with "AS" and "LK" (assembler and linker) installed ?

[rotund parrot]

Here's the assembler, linker, and utf8 to sjis converter all together in a floppy disk image if that helps

should be bootable

[thorny tree]

Confirmed it is bootable

[primal sandal]

Thanks ! I'll look into using this

[primal sandal]

I think I may need to use an emulator to get some of this set up - what is the best X68000 emulator ?

some sources say px68k (various revisions), and some say XM6 Pro-68k

(or XM6 TypeG)

Hard to figure out the differences between the various sub-releases of each

[rotund parrot]

XM6 Type G is quite good, but Windows only. There's also XEiJ which is pretty good as well

[primal sandal]

OK, I understand the scancodes listed in Puu's documentation. The ones which seem very different from the USB scancodes I've been gathering from the Zuiki keyboard

So, X68000 keyboard operate in a similar way - they create scancodes for each key, and these are interppreted by programs. The X68000-type scancodes are associated with key names, and those are the ones on Puu's documentation. I realized this while reading a bit on keyboards from "Inside X68000". I guess Puu felt it was important to document where the keys had moved to.

Now, X68000 scancodes are not the same as USB scancodes - though in some cases they are the same codes (and in others, they are off-by-one or more). This was why I was confused.

I was really lucky that the Inside X68000 book has a keyboard with basically identical layout to the Zuiki keyboard (minus the left/right control keys)

[primal sandal]

OK, I'm also finding the OSD screen to be tedious. The whole first page is covered with "mount FDD / save FDD / eject FDD / mount HDD / Select SRAM / Load SRAM / SAVE SRAM

I would prefer if the FDD0 and FDD1 choices were sub-pages. and I would really like for writes to be auto-committed.
SRAM should also be a sub-page I think - how often do people even select/load/save that ?

[icy shuttle]

Honestly, I have never used that

[primal sandal]

EXACTLY ! So why take up valuable space ?

[thorny tree]

save FDD / eject FDD are used only with the X68000 core version supporting d88 floppy.
Puu-san has another version of his core supporting only xdf disk format. The save and eject are not present any more.
This implementation is better I think because the read of the floppy is more accurate
With the d88 it takes twice as long the normal time of read the floppy ...
The issue with XDF he used plenty of clocks so it needs some adaptation to integrate it on the mister core.

[primal sandal]

Wasn't that version reading/writing directly a SDCard (i.e. second SDCard) ? Or was it actually going via Main ?

For now, I'm just a bit tired of the layout so I will move some operations to sub-pages. The controller and keyboard layout aren't even visible without scrolling

[thorny tree]

the early version of the core (d88 one) using SDCard
The last two versions of puu-san (d88 one and xdf one) don't use second SDCard

[primal sandal]

OK, well for now, I don't have the time and concentration to dive deep into somebody else's code - I'm just trying to get some controllers and keyboard working as a start. The menu is slowing down testing, so I want to streamline it.

[primal sandal]

OK, I've ordered a "Magical" 4-button pad (bundled with some Garou Densetsu games), and as soon as that arrives, I'll take it apart and see how it works. No schematics online.

I'm also looking at Cyber Stick; my XE-1AP code isn't 100% working on the X68000 games (which want Cyber Stick). The protocol is supposed to be the same, so I want to see if there are changes like timing, etc.

[grand mulch]

i very much appreciate your passion for this

[primal sandal]

Thanks. My goal is to get some of the modern I/O pieces to work with MiSTer.

They have made a new keyboard, mouse, Cyber Stick.... they're all very expensive though

[grand mulch]

i was scared off by the cyber stick price

[primal sandal]

Yeah new one (USB) is almost the same price as the old one (DB-9)

[south estuary]

Man, last time I was shopping for either of those peripherals they were about 40 bucks each.

[primal sandal]

Oof. Lots higher today, even for "junk" units

[lethal scroll]

These pads use the MD chelnov adapter logic

I made myself the 6 button adapter for MD pad and it works with chelnov and garous

GamesX has the schematics

[primal sandal]

Chelnov only looks for button 3

I saw the schematics on gamesx. Garou Densetsu recognizes the third button (plus START) on a MD3 gamepad, but no extra buttons on MD6

Anyway, I have some peripherals in transit and should get some interesting things added soon

[lethal scroll]

It worked

[primal sandal]

Augh. The keyboard processing code is really convoluted

It's reading the MiSTer keyboard based on the PS2 serial interface, rather than the "whole key" format.
Then it has to do some sort of transform on that value (which I am having trouble finding), and then pipe that transformed data to a similar serial port on the X68000 side. Not as straightforward as I had hoped

[grand mulch]

interesting, so a shortcut was once taken but has lead to a dead end for accuracy/function

[primal sandal]

Well, I recall when we changed over the CoCo (or maybe TRS-80 ?) core from serial to MiSTer "key at once" and it simplified the logic immensely

So I got the feeling that serial was the only choice at one point in the past (maybe on MiST)

[primal sandal]

Oh, and I think I understand why I was having trouble with the XE-1AP module on X68000

1. The XE-1AJ and XE-1AP are made by the same company
2. The XE-1AJ is just made for computers, and the XE-1AP is primarily made for consoles, but does have a switch to use with computers....

3. The overall protocol - sequence of operations, timing, etc. is the same between both devices

But apparently, the data is not !

https://sankichi.backdrop.jp/cyberstick_ab_1.html

I have an XE-1AP to run tests on, but not an XE-1AJ, so I never knew this.

Also, I don't have a single computer that will read these - no MSX, X68000, FM Towns, PC-98, etc. None of them.

So if I want to run tests, I need to build a circuit and get a logic analyzer output

...so I've ordered an XE-1AJ just to see it for myself

Anyway, looks like I'll need to add a couple of additional options on that module.

[thorny tree]

https://fxtwitter.com/kata68k/status/1955623542323925172

カタ (@kata68k)

パズルゲーム「X68ZKeeper」 Ver.0.0.7
︀︀・マウス対応
︀︀
︀︀※Ver.0.0.7をうｐするの失念してたなんて言えない。
︀︀github.com/kata68k/X68000/blob/master/Game/X68ZKeeper/X68ZKP007.LZH

Quoting カタ (@kata68k)
︀
パズルゲーム「X68ZKeeper」 Ver0.0.4
︀︀・メモリが足りない場合、遊びやすくしました。
︀︀github.com/kata68k/X68000/tree/master/Game/X68ZKeeper
︀︀X68ZKP004.LZH をダウンロードしてください。
︀︀#X68000 #X68000Z

**🔁 6 ❤️ 12 👁️ 681 **

FxTwitter · Aug 13, 2025, 1:32 PM

[primal sandal]

So... I just got a Garou Densetsu 2 with controller so I could figure out the controller encoding... which I basically did.
Then I tried it on a core built to have that controller, on the Garou Densetsu game - and it couldn't use it.
Then I realized that game (from neon68K) was older than the Garou Densetsu 2 with the controller, so maybe it didn't support the controller (or maybe it would need some sort of patch which wasn't applied).
So then I started looking for a bot drive of Garou Densetsu 2, which definitely should be compatible with this controller implementation.... and I can't seem to find one (other than the diskettes in my hand which I can't read)

Also, is there a repository of scanned game instruction manuals ? These games seem to have a lot of odd options or keys that need to be used for various reasons.

[primal sandal]

OK, I found Garou Densetsu 2 and Garou Densetsu Special, and they have massive graphical glitches - BUT they recognize the Magical controller

[thorny tree]

Yes they have graphical issues.
For scanned game instruction manuals, you have some on archive.org like Garou Densetsu 2

[verbal jolt]

what's up with fatal fury special having such severe graphics corruption anyway? is it a dma thing that hasn't been implemented or

[primal sandal]

The other thing I noticed is that the controllers have 6 buttons, but these two games only use the 4 on the front face.

[primal sandal]

I don't know. But that sounds plausible

[thorny tree]

On the manual of Garou Densetsu 2 you have this mention
It supports 2, 4 and 6 buttons (using Megadrive pad with adapter from dempa) 3 buttons is not supported

[primal sandal]

Hmm. Garou Densetsu - the original - allows for MegaDrive 3, but not MegaDrive 6

But in any case, Garou Densetsu 2 only uses 4 buttons

I don't see the point of adding both Magical 6-button and MegaDrvie 6-button... until somebody can show me a reason. So for now, I am just adding Magical 6

[lethal scroll]

this core needs a lot of fixes

It could be a great substitution for the real machines if it could be ported.

the real machines only give headaches.

[verbal jolt]

not only that, but even buying a junk one will kill your wallet

every day i hope for more fixes to the x68000 core. if it were in the same state as, say, the snes or genesis cores, it would easily be in my top 5 mister cores. there's so many amazing games that are largely unknown/unplayable outside of the original hardware

i would happily pay out the ass for someone to finalize it, or at least make regular updates to it

[primal sandal]

OK, in my research about XE-1AP versus CyberStick, I figured out one of the principal differences, and got further with making games work, but still missing some information.
I made a little microcontroller setup that reads the data, so when my CyberStick arrives this week, I will be able to complete that (assuming it works).

[lethal scroll]

I would support too

[grand mulch]

started looking at the controllers myself

what a wonderfully weird era of experimentation

[primal sandal]

That's for Libble Rabble (and crazy Climber 1 & 2), and I found schematics online for that one. It will be in my next update

[south estuary]

Yeah I did the circuit for that 20+ years ago. Go me, etc.

[trail cypress]

Hmm, I need to for twin stick shooters. Does reflext adapt handle these? Would be awesome for The Binding of Isaac on PC

[south estuary]

It's not a good pad for that, the d-pads are very stiff.

[primal sandal]

CyberStick (~35 years old) came in today, and it looks like I was wrong about the data sequence
It looks to use the same protocol as XE-1AP, although the variety of swtiches are hard to correlate.

But it seems that the timing of several things needs to be more precise than I had been making it

[primal sandal]

Crap, I had it mostly working, and then I made an adjustment and everything broke again. I'm going to have to pull out signaltap and look at this. I'm not sure what the game program(s) are doing to read it.

[warped spindle]

Is the core still in development by puu_san?

[verbal jolt]

last i heard he's currently working on an fpga implementation of fm towns. from what i understand he seems to bounce around between projects often, which is a shame as the x68k core desperately needs more work

[primal sandal]

He has done work on multiple cores of Japanese computer. "still in development" has a wide range of meanings, so I'll put it this way:

• he hasn't said that he's stopping, or that it's finished
• he hasn't published anything in a while (I think since early 2024 ?)
• he rotates the projects in and out maybe every year or second year

But he also doesn't use any sort of code control for his public releases, so it's hard to reconcile against existing core.

Definitely NOT designed for collaboration

[primal sandal]

Still, his releases are apparently ahead of the official MiSTer version in some ways, but need to be reconciled

Right now, I'm looking at peripherals and not the internals, but I might get there at some point

[icy shuttle]

Yeah, puu releases an RBF and source code and then that would slowly get put into the official framework. I believe there are still things from 2024 that aren’t in the mister core

[lethal scroll]

If any day there is a good core of this computer I will sell the real manchine... I don't wanna suffer more...

[primal sandal]

FINALLY got CyberStick emulation working

[rotund parrot]

Nice! That’s awesome!

[primal sandal]

Pull request submitted

[thorny tree]

Thank you 👍

[icy shuttle]

https://bsky.app/profile/timeextension.com/post/3ly3nzpikma2u

Time Extension (@timeextension.com)

"Lost" Gradius Sequel Nemesis '90 Kai Is Getting A Reissue For The X68000 Z.

-# "Lost" Gradius Sequel Nemesis '90 Kai Is Getting A Reissue For The X68000 Z
Will be given away with a one-off issue of Oh!X

Bluesky · Sep 5, 2025, 12:30 PM

[verbal jolt]

"lost" mf you can download it right now

[icy shuttle]

https://tenor.com/0rYH.gif

[digital fjord]

This is my favorite gradius in the whole serie.

[south estuary]

Exactly my thought. "Lost? Bitch I used to own two copies."

[hasty meteor]

I lost the game

[primal sandal]

@thorny tree I'm looking at the coommit history of X68000, and currently trying to figure out which of puu's updates are included, and which are not

I see a lot of commits made in May of 2021 by you, but it's not clear whether you put all of puu's work in

You had said that some of puu's fixes haven't come over to the MiSTer repository, so I wanted to take a look at what was missing

In any case, if you can help identify where I should start to look, it would be helpful

Especially, you mentioned that some things which are still broken in the MiSTer version, were fixed by puu

[primal sandal]

For example, I think you said that there is some garbage left on the screen in the MiSTer core in some cases, but that was fixed in puu's core

If this is true, are you able to locate which version of the core this was fixed in ?

[thorny tree]

I will summarize a bit. Puu-san doesn't use git of other kind tools for this code and his repo doesn't follow MiSter style (with rtl directories etc...).
First work of the Core done by Puu-san was ported by Sorg until april 2020.
Puu-san continues his core but Sorg stops to updates this MiSter Core to evol so in may 2021 I ported last updates of Puu-san
After that, Kitrinx and Sorg work on it until december 2021 (change CPU, replace sound by JT modules, improves graphics, clocks, etc ...). Puu-san continues his own core without using those changes. So both cores diverged.
Later Rebel40 ported from Puu-san (dma, midi and sound fixes) and i ported layout keyboard from Puu-san.
So some part are better on the MiSTer core and some part are better in puu-san core.

1st point may introduced by changes on MiSTer cores and working on Puu-san core, it is texts issues on the shell if plenty are shown and some scrolling. Puu-san fixed it on his release 20200831. So this one could be difficult to track.
https://github.com/MiSTer-devel/X68000_MiSTer/issues/40

2nd point Puu-san in this blog post create a second version of X68000 core using XDF for floppies and continues to maintain other core usign D88 for floppies.
http://fpga8801.seesaa.net/article/489733322.html
The XDF one is quite better some games are graphical fixes like R-Type using XDF format maybe introduced by conversion on D88.
Also it fixes loading time, with D88 core it tooks two times to load content.

3rd point, it is analyses done by Rebel40 regarding two tickets. He managed to fixe issues (with new regressions). Those issues are present on both cores.

• Fixes for TNZ https://github.com/MiSTer-devel/X68000_MiSTer/issues/21 but break nemesis'90 and akujama dracula

• Commit on one of my branch (for better visibility) https://github.com/wwark/X68000_MiSTer/commit/64756ffd48927c587323105a6e02b117eb18690c

• Fixes graphicals issues on some games but introduced some timing/speed issues
  https://github.com/MiSTer-devel/X68000_MiSTer/issues/43

• Commit on one of my Branch (for better visibility) https://github.com/wwark/X68000_MiSTer/commit/08755f7281d1345454213230175d19a432436858, this commit also added INTCont changes from Puu-san not ported on the Mister core. I checked on X68000 inside books it seems follow but i couldn't find any improvement for the moment with this change so not propose

4th point, it is Memtest68k.d88 -> it is just black screen with MiSter Core and working with puu-core https://imgur.com/PG8pAti

[slate timber]

someone should just rewrite it

[icy shuttle]

👀

[verbal jolt]

i would pay someone the msrp of an x68k to do that

[icy shuttle]

No no

that's not how you catch them

if you want to rope a dev in

you need to build a test rom

[verbal jolt]

lol

i wish i was qualified to do so

[primal sandal]

Easy to say, but not so easy to do. In order to rewrite it, one must understand how it's supposed to work. The current implementation is an approximation with issues; only by understanding the nature of the issues (or studying the entire machine from the ground up) can one actually create a fix.

I'm not going to build something from scratch for a machine I neither own, nor have substantial knowledge of. But it might be possible to fix a bug

[thorny tree]

I have also a pastebin with a list of issues (hangs, graphical issues, sound issues, memory issues, etc..)

[thorny tree]

Also for test programs, we have some here but never tried them on the core
https://stdkmd.net/xeij/miscfiles.htm#testprogram

[thorny tree]

Testing prwbit2.x on the core

[primal sandal]

I think I can't even trust emulators to do things properly... Need to test against an original machine, and I won't buy one of those anytime soon

Maybe 2026. Maybe

But I do agree that rewriting this would make it easier to work with

It's just that the knowledge has to first come from somewhere.

[thorny tree]

Sorry I don't have have the real hardware on my side ..

[verbal jolt]

not many do

[primal sandal]

Anyway, I'll look at puu's source progression to see if there was a fix he made that never got imported to MiSTer

I already made a temporary private git repo with each of his versions so I could diff the version changes to track updates, and cross-reference against MiSTer's

One thing I already noticed is that the file vcaddr_pkg.vhd introduced some addresses in the 2020/06/30 update, such as:

constant VC_R0s  :std_logic vector(23 downto 0) := x"e82402";

And it remains that way even in the 2024 version of the file, but the MiSTer version makes it:

constant VC_R0s  :std_logic vector(15 downto 0) := x"e824";

(as well as the various comparisons that involve this constant).

It's not really clear why this change was made in MiSTer; it makes the comparison less specific, and this could account for some behavioural difference

....or maybe not... it's just not clear

Also, the updates to MFPint.vhd from 2020/07/17 don't look like they were applied to MiSTer (unless they were later removed)

I'm not a fan of puu's approach to file organization

[thorny tree]

Normally I have appied changes on the file MFPInt.vhd when I have ported and Kitrinx adapt it for clock enable
What are the differences you saw ?

[primal sandal]

Hopefully this is going to be legible...

This is a diff between 2020/07/15 and 2020/07/17

Look especially at the section from old line 240 (deletion) - it's still in MiSTer's version

The large addition section also appears not to be in MiSTer

It may have been reverted subsequently though

[thorny tree]

you are right this part it not present on MiSTer side

[primal sandal]

If I look at the 2024 version from puu, the deleted part appears to be back

and the added part appears to be substantially changed

So I don't know if this is relevant

I'm trying to walk through the changes, one by one, to see if anything appears to be overlooked (instead of just doing a compare of "current version", which would be overwhelming)

It looks like rebel40 is working on MIDI stuff - which is good, because I don't know about MIDI, and that would be something I would avoid to work on

I added you to the temporary private repo, so you can also view the progression of changes in puu's releases

[thorny tree]

For MiDi there is still games having issue also on Puu-san Core and maybe an option to activate or not MiDI because it is activated by default and in some games, MiDi is enforce in this case and FM sound needs to be change after that.

[primal sandal]

Curently, I only put versions up to 20200913 in that repo

[autumn geyser]

I'm obviously in the minority, never even saw a X68000 in the UK, so I'm only going by what I've played on it, but it seems to be okay to me. What do the devs think is so badly wrong with it? Accuracy, features etc.?

[primal sandal]

Lots of unplayable software, and this issue is probably the biggest issue for me:

https://github.com/MiSTer-devel/X68000_MiSTer/issues/40

[autumn geyser]

AH right okay, fair enough!! Thanks! 🙂

[primal sandal]

Is it unusable ? No.

Does it meet my standards ? also No

[thorny tree]

also indicator like testmemory shows problem https://imgur.com/PG8pAti
but most games don't rely to it

[autumn geyser]

I'll have to check that, how have I not noticed that corruption when the system boots before - mind you I am blind as a bat! Thanks both for the explanation.

[rugged geode]

So, more of an academic question, but I have always heard that the x68k is close in design to the cps1 hardware. Is there any truth to that or is it just similar in specs and implementation is different?

[primal sandal]

No idea about specifics, but it's a similar timeframe. Also, it's interesting to note that Sharp was doing design for other companies in roughly this time period, as Hudson worked with Sharp on the video chips for the PC Engine (in 1985-86). But I can't say anything about CPS and X68K

Also, if you look at Japanese computer magazines in that era (1988-1991 at least), you'll see that there was tremendous energy being spent making licensed home computer versions of arcade classics. So italso isn't necessarily down to just the hardware being similar

[rugged geode]

Well, I’ve always taken it as true that capcom developed cps1 games using it

[primal sandal]

Could be. X68000 was used for lots of development for lots of consoles too - including art and music - but that doesn't mean that it was compatible. Still, I don't know

[lethal scroll]

They share the CPU, same FM synth and simmilar adpcm ( just 1 channel on X68k ) but I dont think they share too much more.

[rugged geode]

Ah, I had always assumed they shared similar custom hardware.

[primal sandal]

A lot of systems had similar capabilities at that time (i.e. graphics), but they weren't necessarily implemented in the same way

[hasty meteor]

@foggy pawn @river mango know anyone that might be willing to lend their x68k to @primal sandal for core development? I know it's a huge ask but I thought I might as well.

Might ask @ember musk too since he's been in Japan a bit.

[primal sandal]

I won't be able to do anything with it until 2026 anyway

So don't

[hasty meteor]

I think I pinged the wrong person sorry xD

Ah I thought you meant for financial reasons. Sorry.

[primal sandal]

Nah, I'm booked up and I have too much on my plate right now

But I do want to learn about the machine little by little

[hasty meteor]

Thanks. Gotcha

[primal sandal]

I'll probably get a X68000 at some point in the next year or so

[ember musk]

I've got nothing, sorry 🙁

[hasty meteor]

Oh I did ping the right one haha. It wasn't showing blue on the name when I tried so I wasn't sure xD

Thanks though

[foggy pawn]

No, sorry

[river mango]

Let me ask around

I am fucked busy right now though so remind me later this week. Real estate bidding shit in the next 24-36 hours

[hasty meteor]

Well he said he wouldn't have time until 2026 haha

[primal sandal]

And probably mid-year at that

[river mango]

haha I just mean someone remind me later next week because I am going to forget about this guaranteed

[pine kindle]

Where are you based for shipping if one had an X68000?

[hasty meteor]

@primal sandal tbh I'm just really hoping for more x68k support xD but your life takes precedence haha

[primal sandal]

I'm in Canada (Toronto area), but really, don't bother - I will get one myself eventually.

[foggy pawn]

I’d prefer if he’d focus on this over his own wellbeing tbh

[primal sandal]

I kind of figured that somebody would feel that way. Didn't expect it to be stated here though.... 😉

[foggy pawn]

lol I’m kidding!

[primal sandal]

I hope so !

[river mango]

best I can do is an X69000...that work?

[primal sandal]

uh.... no thanks.

I can wait

[river mango]

🤣

[verbal jolt]

speaking of which, how's the dreamcast core coming along robby? chop chop

[rugged geode]

I thought Robby was busy modeling the electronics of an LB&SCR Class E2 tank locomotive

[foggy pawn]

yes

[autumn geyser]

Nowt wrong with model railways, minues 16 feet by 8 feet, mainly scratch built, set in the North-East of England in the 1950.s. Built a pit, slag heap, tunnels, bridges, houses, a church and graveyard etc etc. 00 guage. Like this, it's a hobby. 🙂

[rugged geode]

not poking fun at model railways. Robby is a notorious train junkie, to the point he has spent more on esoteric train controllers for his game consoles than on rent and groceries this month

[autumn geyser]

Good lad, it's good to have lots of hobbies. Sorry taking this OT. 🙂

[foggy pawn]

If I could marry my train controllers I would

I would dress Sharply with a big X bow tie as I marry all 680000 train controllers,

[lethal scroll]

if this core will work good one day... will be awesome

[verbal jolt]

it's my most-wanted "unfinished" core

the original hardware is so prohibitvely expensive, and i would much rather play the games on a crt tv and my controller of choice, instead of having a real x68k with a multisync crt monitor, keyboard, scsi emulator, etc.

[thorny tree]

https://x.com/romhackdoting/status/1971713281698967678?t=dFoLcnXH2Fju8Fg7eggxiw&s=19

Romhack.ing (@romhackdoting)

Silk Road 2 - English Translation (X68K) Added https://t.co/mlZnWvKXSc

X · Sep 26, 2025, 11:07 PM

[thorny tree]

MacDonald pub on X68000 (it is a Mugen like)
https://fxtwitter.com/pipixvi/status/1994707922660806962?s=20

ぴぴぴ☆@12/31(水) C 1 0 7 南m-22b (@pipixvi)

【 X68000 】 SFXVI #331 トリチ McDonald's
︀︀ネコサダちゃん(@mokonaXVI)制作のトリチ！！
︀︀この多関節感が未来を感じる（ばく。
︀︀フェルナンデスに食われる姿が一番似合ってて好き。
︀︀#SFXVI #X68000
︀︀youtu.be/N-dtVnFOB8Y

**🔁 6 ❤️ 12 👁️ 403 **

FxTwitter · Nov 29, 2025, 10:00 AM

[icy shuttle]

the moment we have all be waiting for

[cunning sleet]

Where’s the latest RBF file for this?

Welp, scratch that, let me check to see if I got it lmao

[thorny tree]

https://www.youtube.com/watch?v=HoHk7qnURis

[rugged geode]

Why does the soundtrack sound like an old IBM PC speaker composition?

[thorny tree]

https://github.com/albconde/X68000_MiSTer/commit/cbb2ab1e1cf6d9185fdcafab77de13bc877375f5

[robust ore]

Compilation in progress.
EDIT : It seems to work on my MiSTer systems, so let's upload it.

[rotund parrot]

Ooh excellent

[uncut mantle]

how fast you find this ... , do not expect miracles, only little fix , and i hope no broke other things ...

[robust ore]

Wark91 is really, really fast for finding things.

[robust ore]

I need to check the second level on Final Fight one more time, otherwise the core seems solid.

Okay, false alarm, I ran too much games on the core without rebooting completely the core.

[uncut mantle]

Final Fight gives an error that seems random; it happened with the previous core and with this one.

[thorny tree]

Does fixes speed with fixes of rebel40 ?

[uncut mantle]

, it ends up crashing

It only apply the 12 cycles dma when is needed , by default use the 9 cycles

if you find any "regression" with the original comment .

[robust ore]

On Garou Densetsu 2 & Special, the Graphic layer is bugged, but the rest is working correctly.
These games are semi-playable.

I don't remember them this 'playable' on the core before.

[uncut mantle]

thank this happens with the previow too.

[robust ore]

So far, I have not seen any regression or major crash.

[uncut mantle]

there are a lot of think that not work , but not work in the previous too

[robust ore]

I greatly thank you for these fixes @uncut mantle.

[uncut mantle]

i try to fix more things , but in the end this core need to be rewrited .

[robust ore]

Yeah...

I'm already thankful from what we could achieve here.

[thorny tree]

Thank you Albconde on my side I could test in some days away from my place

[mental pine]

👍 👍 👍 many tnx @uncut mantle

[mental pine]

very nice, atomic robokid sprite offset also fixed !!!!!

[true flame]

My X68000 sense is tingling!

Always appreciate updates to this core, no matter how minor.

[foggy pawn]

@icy shuttle

[icy shuttle]

if this gets pushed, we may have to get the band back together @eager yacht

[uncut mantle]

Thanks to your analysis, only the last piece was missing to understand the differences in the DMA.

[river mango]

Ooh fun

[buoyant bay]

Wait, new X68000 build? Guess it's time I finally figure out how to convert the Easy/15khz versions of Akumajou Dracula to HDF.

[robust ore]

I've done it yesterday.

You need to open the HDF file with DiskExplorer as a Human68k HDD.
Inside the DRAC folder, change the original DRAC.X file by the new one.
Give it the same name as the original one (replace it).

[buoyant bay]

Ah, that would be the elegant way. Thanks.

[robust ore]

I can't help much more for today.

[buoyant bay]

That was all the help I needed I think. I've got it opened up and am transferring the new files into the drive image. Appreciate it!

[uncut mantle]

No changes in Akumajou Dracula , i try to fix the clear of the text layer , but without success , in the test core there are a "debug layers" menu i put this to see in which layer the graphic error was , in akumajoy you can see the text layer not clear some data

[buoyant bay]

Yep, I noticed the option to disable layers. That's really helpful as a stopgap to a fix. You lose the HUD but I can live with that.

[eager yacht]

oh! so when you get to glitched parts in AD you can just enable that and have it play “normally”?

I feel like it’s mostly a late game issue for AD?

[buoyant bay]

Yeah it was in the clocktower or whatever when I got glitched out.

[dim violet]

Nice to see an update to the X68000 core, and that Atomic Robokid is running better. I haven't touched this core for quite a while,convinced that some games that weren't previously working now seem to load. A big "THANK YOU" to you guys for pushing this through!

[pine kindle]

Same with Akumajō Dracula? If I get that overlay I can just disable layers to continue?

[eager yacht]

not sure, I haven’t tested yet and I was asking the question myself

[buoyant bay]

Yes, remove the text layer and you can progress without the clutter on screen.

You'll lose the HUD but be able to see.

[eager yacht]

thanks for the confirmation! Cool to have a workaround like that

[pine kindle]

Appreciate that thanks, what a strange issue.

[uncut mantle]

Some more small fixes , the most important , fix the text layer clear error in Akumajou Dracula ( the game clear the screen with raster copy )

[icy shuttle]

@eager yacht ^^

[eager yacht]

awesome!! I’ve finally been doing some testing again, so I’ll grab this build and get to it as soon as I can. Looking like we’ll be able to add some new stuff to Neon68K

[robust ore]

Awesome!

[rotund parrot]

Wow! Awesome work 🥳

[eager yacht]

Found several shmups that now work on the newest core that had graphical issues before: Cy-Lips (still some gfx and sound corruption, but totally playable), Desperado, Hellhound, Kyukyoku Tiger, Parsec, Phalanx, Salamander

[verbal jolt]

does chourensha still have crackling audio

[eager yacht]

haven’t checked that one yet

[verbal jolt]

been a while since i last played it

[eager yacht]

I’m working through our old testing list for Neon68K and going back to the ones that we either included and marked as “minor issues” or didn’t include due to major graphical issues

Marble Madness still has corruption on the title screen, but the levels look ok now and it’s playable

[verbal jolt]

i'll have to see if anything that was glitchy for me works now. fatal fury special and Y2 come to mind

[eager yacht]

lemme know if they look better and I’ll check our sheet!

[thorny tree]

This issue is normally fixed (https://github.com/MiSTer-devel/X68000_MiSTer/issues/40) @uncut mantle with your last update.
Are you able to run Mentest68K program, it is black screen with last release of MiSter core but working with the Puu's core.

[uncut mantle]

From Puu's core i know other two thinks that runs better than in our version rigar ( argos ) work , ,black screen in official official version , and some garbage in Deflektor

[uncut mantle]

where i can fin mentest68k ?

[thorny tree]

PM you

[dull kindle]

That's a cool update, thanks ! Can we have midi support with it ? Midi is not working for me at least...

[lilac meteor]

MMDSP (a mdx player) does not work correctly anymore, graphics are all messed up on the latest core:

it worked on this version with text glitches: #1351743248850227360 message

a while ago puu-san had some DMA and other fixes which also I believe fixed this and back then MMDSP worked flawless, without the text glitches:

[uncut mantle]

i do not have a pi midi so i try to fix other thing before ,

[foggy pawn]

Someone said Pi, I’m here

[uncut mantle]

there are some mistakes in 31 khz modes , as Thera34 found , need made other changes

[near dirge]

As someone that has been stuck on the clock tower level in Akumajou Dracula for months, I thank you! This will make the suffering much more bearable!

[thorny tree]

Music player of Akumajou is great now 🙂

[thorny tree]

On Music player, with Bloody Tears title, one or two drums are not correctly played (it is like buzz or something) other drums are OK (FM sound selected)

[rotund parrot]

Here's Bloody Tears played on a real X68000 Compact XVI model

[thorny tree]

Undeadline Intro and attrac mode on Syvalion are quite better

Syvalion have centering problems on some screens

[primal sandal]

For games of this age, you really need to compare to how it looks on original hardware. Sometimes centering problems are actually game programming (how they wrote it)

[rotund parrot]

It does appear to be a bug, these screenshots are taken from original hardware. Also pretty cool game, I've never tried Syvalion before

[uncut mantle]

Yes is a bug the core is a little disaster , i fix the sprite size, but no the position

[primal sandal]

Oh, I thought you meant left/right centering, not up/down centering. Yeah, that would be an issue

[eager yacht]

this is all the stuff I've found so far that we didn't include with Neon68k due to issues (or ones that we included in the "Minor Bugs" category) that seem to be working pretty well now with the latest core

Baruusa No Fukushuu (Gfx still partially corrupt, but much more playable now)
Cameltry (some title screen gfx corruption, other minor corruption (i.e. out of place 'game over' graphic)) , can’t select menu options [needs spinner? Mouse works, but is a bit wonky]
Cy-Lips (minor gfx/sound corruption still, but less than before!)
Desperado
Genocide 2
Hellhound
Kyukyoku Tiger (shifted sprites fixed)
Marble Madness (corrupted text in menus)
New Zealand Story
Nemesis '94
Parsec
Phalanx
Salamander
Star Fighter
Syvalion
Twin Soul

[grand mulch]

Bonanza Bros?

[primal sandal]

That's a pretty good set of fixes ! Nothing that was working got broke ?

[eager yacht]

I haven't gone back and tested that just yet

but I'll dive into the whole list after my first pass at the games that either didn't boot or had graphics corruption

[eager yacht]

ya know, I don't see that one on our spreadsheet 🤔

[grand mulch]

it’s a good port, there was only some funkiness in the bonus stage last i checked

[eager yacht]

huh, I wonder why we don't have it

[grand mulch]

you’re waiting on jotego to make the System 24 core!

[eager yacht]

there was at least one game that we didn't include that actually worked all along and we just missed that it requires a mouse somehow (Star Blade)...whoops

[tawdry creek]

Oooh, Cameltry might work with a spinner?

[eager yacht]

that's my guess

it works with a mouse, but it's not great haha

i'm presuming that marble madness needs a trackball too...it also works with a mouse, but it's not the best

[tawdry creek]

I've got one of those Thundersticks spinners, I'll have to give it a try

[rotund parrot]

Marble Madness also has gamepad support but you have to go into the options and select it as the input device

[eager yacht]

oooooooh

[rotund parrot]

Does work best with a trackball

[eager yacht]

testing now

one issue with Marble Madness is that the title screen text is corrupt still

but on the previous core, the levels were all messed up

so yeah I can't actually see which option it is to select the controller for marble madness

actually, I just changed something in the input menu...went to the game...it still controlled with the mouse but it seemed better>?

oh I wonder if there's a sensitivity setting or something

whatever I selected definitely made it better with the mouse

Marble Madness: select option 3 in main menu > option 5 > option 2 > option 1 > option 1 in main menu. That seems to make the mouse work better and it works decently well with a controller in mouse mode

one of the other options on that third input menu probably changes it to gamepad 😅

ahhh the second to last option is the winner (option 6)

[rotund parrot]

Sorry had to step away for work for a moment, this is what it looks like on the real machine

and you're right about the top 3 options adjusting sensitivity

I recall I had a lot of issues using my trackball with the real thing until I selected the right setting

before, I had to really move the trackball to even move slightly ingame

but with the correct option it's much better

[eager yacht]

I can add instructions when we add it to the pack if need be for anyone that wants to play it…maybe with some luck those menus will get fixed before we do a new release 🤞

[tawdry creek]

Cameltry doesn't seem to work with my Thundersticks spinner. The buttons on it works but that's it. A plugged in mouse seems to "work" but that seems to be it? I can't seem to do anything to navigate its menu

[eager yacht]

oh yeah that’s right…touching the mouse at all zips you right through into the game

so it’s playable with a pretty big caveat for sure

wonder if it’s similar to MM where you need to go to a config menu to enable spinner (which you obviously can’t access )

[south estuary]

The real X68 Cameltry came with a metal mouse bracket with a spinner. It clamped the mouse in place and the spinner rotated a disc under the mouse. So, there probably isn't a spinner option because it only and always uses a mouse, either way.

[eager yacht]

ahhhh that would make sense

[primal sandal]

That's it. I recall as I was adding special controller support that Cameltry had something bizarre that prevented the "special controller" from being implemented

[rancid compass]

That game is the true reason mankind needs that core

[robust ore]

Still crackling audio issues with Chou Ren Sha.

[uncut mantle]

jt51 is not the last version , i will update to the last available , maybe this improve the sound , or maybe no ....

[eager yacht]

yes that game rules!

[thorny tree]

For information the jt51 was modified a bit so not sure it is possible update it straight away

[buoyant bay]

I'm super appreciative of all the work being done. I think I'm seeing a regression in DaiMakaiMura. 31khz/CRT Mode 2 appears to disable the Sprite Layer. Seems to work on the late 24.05.24 core release.

Not sure that it matters at all since Direct Video doesn't work on the core and I'm not sure that the scaler cares or if in game resolution really affects quality with how this is rendered.

But thought it maybe worth a mention.

[uncut mantle]

thanks , i see , i make a mistake with the sprite doubled , is the same case that thera34 found in MMDSP , the sprite is out of the screen , every info is important thanks for testing , i was trying to fix this , but in the end maybe i need to revert this change , i can not found a solution that work in all case for the doubling sprites ( deflektor in game and Syvalion logo )

[rancid compass]

Porting to the nuked OPM core is not in the cards? Version 1.0 was awarded by nukeykt on January 1st

https://github.com/nukeykt/Nuked-OPM/commits/main/

I guess that the nuked core would guarantee that it is a die based 1:1 replica and that the chance is much higher to have Chou Ren Sha fixed

But it's probably bigger and not just a "drop-in" replacement

[uncut mantle]

not sure about this , i was testing the core and i think the most part of the problems are adpcm ( oki 6258) not related to the opm ( yamaha 2151) . The problems is there are no other oki 6258 implementation , only the puu's one. All things i try this week not improve the core

[rancid compass]

Oh, I was just assuming y2151 and thought that switching the module would have a higher probability to fix bugs than just updating jt51

But it's pure guesswork of course and not based on any understanding of the hardware

[uncut mantle]

previously i think the problem is in the jt51 too , but for example the problem in chorensha now think is pcm , i put in a test mute jt51 option and mute adpcm , and mute adpcm seen to solve ( you do not hear the shoots , etc obiously .only music ..) Some Drums in akumajo are adpcm too

[uncut mantle]

test version , only changes in sound , update jt51 and some little changes in the adpcm , please test the sound and tell me if is better , worse , or equal

[robust ore]

Seems better on Chou Ren Sha, but we need other listening testers.

I can't easily detect the sound cutoffs anymore.

[thorny tree]

Bloody Tears still has issues

[verbal jolt]

what kind of issues?

[thorny tree]

#1351743248850227360 message

Phalanx also has the same issue and more presents

[verbal jolt]

ah i see, thank you

i tried chourensha on the sound test core, and it sounds great. nice to finally hear the game without crackling audio 🙂

i still see a lot of graphical glitches in many games though... garou densetsu special has no improvement. music in cyberblock metal orange and Y2 are still too fast as well

[verbal jolt]

also, does anyone know what model of X68000 the core is based off of (if any)? like the super, compact, XVI, etc.

[uncut mantle]

thanks for the testing , so seen the cracking sound is fixed , but still are other problems .

[rancid compass]

Well. Chou Ren Sha is THE X68000 game. This being fixed is the best possible news regarding that core.

[verbal jolt]

illumination laser is really cool too

[dull kindle]

I love code zero, the music rocks !

[cobalt mulch]

Hi, I’ve been testing the updated x68k cores, but the UART over USB Midilink doesn’t seem to be working. Was this feature disabled in the latest version?

[cobalt mulch]

I tested the latest stable version and it works. I also tested the latest unstable version and it doesn’t work. From what I saw, they recently changed something in the MIDI code in the Git repository, so it looks like the modification made there is what broke it

When the Castlevania issue is finally fixed, the MIDI stops working 😅

I'm using an original Roland UM-One together with an SC-55

[uncut mantle]

is broken in the commit in the oficial repo ? or only broken with my changes ?

[cobalt mulch]

I believe this commit broke it. https://github.com/MiSTer-devel/X68000_MiSTer/pull/42

I reverted this change and I'm installing Quartus to verify if it was indeed the cause. It’s very likely that this modification is what caused Midilink to stop working over UART

I’ve been waiting a long time to play Castlevania with the SC‑55, now that the problem has finally been fixed. 😅

[thorny tree]

This change fixes midi for some games but seems change the way UART is used ... (it is part of changes of Puu's work)
I have only tested with mt32-pi not real midi hardware

[cobalt mulch]

Unfortunately, the real module is no longer working

[thorny tree]

So it will required a proper fix for UART and midi.
Could you create a specific issue and someone with real hardware could fix it and test it ?

[cobalt mulch]

I created the core without the change and now it works midi, now I can finally play castlevania 🥳

[cobalt mulch]

@uncut mantle Are you going to fix other graphical issues? I'm asking because I could report them in your fork on Git.

[eager yacht]

yeah there are probably some things we have documented from our Neon68K testing that could be reported if helpful

[cobalt mulch]

Understood. Castlevania appears to have introduced an issue with the background layer in the first stage. Additionally, in this stage the items seem to appear slightly below the ground line—a minor discrepancy. I am not certain whether this is expected behavior, as in other stages the items appear at the correct height.
Aside from that, I would like to express my appreciation to @uncut mantle for addressing several issues

[uncut mantle]

i will try to made more fixes , but i can not promise . but a compatiblity list will be great , like other cores have . My fork is only "temporaly" when we have a stable version i upload to mister repo .
i touch a lot of thing so regresions are expected. @cobalt mulch can you give me more info in the error in castelvania ?

[cobalt mulch]

After the correction appears as in the image graphic error there in the background layer that does not have in the stable version

I'm not at home right now, but as soon as I arrive I'll take a picture of the items on the floor that seem to be out of standard, although it was good to see on the x68k itself

[uncut mantle]

ok see is in the text layer ( text layer can have graph) , thank

[cobalt mulch]

Another game that presents serious graphical issues is Fatal Fury 2 and its Special version. The problem appears to be related to the background layers of the images, both in the stages and on the title screen. When the background layer is disabled in the debug menu, the issue disappears.

I believe this is an issue that already existed, based on my tests using the stable version as well as with the modifications you made.

The layer where the characters are rendered appears normal, but the rest such as the background and other stage elements shows graphical issues

[uncut mantle]

yes the issue exist previously is a little strange , i I think I know what the error is, but it's very strange: the memory is used differently in the core in 65k color mode and in 256. This is wrong but it seems intentional. The game loads the graphics in 65k and then displays them in 256. I was just looking at it right now.

[thorny tree]

memory of the game or internal memory of the x68000 ?

[uncut mantle]

the core load the memory graph in a diferent way in 256 and 65k in 65k use g_base + m_addr[18:1] but in 256c use g_base + {m_addr[18:10], m_addr[19], m_addr[9:2]} this is wrong .
The problem in this game is the graphics are loaded in 65k mode bus show in 256 mode , so the memory read is not correct

to fix this i need to use the same memory write and read in the two cases ( as mame or px68k do )

maybe the 256 mode is a trick to made the access to the ram faster , i not sure about this

[thorny tree]

I wonder if it is not linked that Memtest68K shows on Puu's core
https://imgur.com/PG8pAti

[uncut mantle]

Is posible , in the 256 mode ( from xm6g ) the memory is :
Logical Address:$00C00000 Physical Address:$00C00000 FC:$5(101)
+0 +1 +2 +3 +4 +5 +6 +7 +8 +9 +A +B +C +D +E +F
00C00000: 00 FE 00 FE 00 FE 00 FE 00 FE 00 FE 00 FE 00 FE
The core seems to no load the "00" part not load the first byte ( 0 , 2 ,4 .....)

[verbal jolt]

@uncut mantle don't want to annoy you since i know you're already busy, but on the sound_test core, illumination laser is a little weird. there are some occasional loud, scratchy noises, and starting a new game after getting a game over caused the music to sound off-pitch

[uncut mantle]

ok thanks i will check too

[verbal jolt]

thank you very much! godspeed

[thorny tree]

@cobalt mulch, Yosinda gives me some hints that could make fixes midi over UART. Could you please test if it is working on your side ?
Could someone test Midi with MT32pi ? I couldn't test on my side for the moment ...
This build doesn't contain fixes from Albconde.
If MT32Pi and Midi over UART work, I will propose PR.

[cobalt mulch]

Great, when I get home I will test it. I tried it yesterday and reached stage 12. In stage 6, when you face the boss before the water rises, there is a graphical issue that occurs when breaking the lion on the wall. I couldn’t take a picture because the error disappeared when I paused the game.

I also noticed that the items on the ground appear slightly lower than they should in stage 9, but I still need to test it in an emulator to confirm whether this is a standard behavior or a random occurrence.

As for Akumajō, I believe the errors are related to garbage data in the text layer.

[uncut mantle]

Some info of my checks , not good . As I think the Problem in Garou Sp are motivated by the change from 65k to 256 c mode , the core do the 256 in the graphic layer "interlaved" and insert this interlaved in 4 caches of 128 Words , quit the interlaved and use 4 caches of 256 words fix the problem in Garou SP , break atomic robokit , but is the correct way . But I find other problem x68000 in 256 c there are two scroll layer the core use layer 0 - layer 1 for the first and layer 2- layer 3 for the second , this is correct in 99 % of the cases , but for example code zero logo uses layer 0 - layer 2 for first and layer 1 -layer 3 for the second , to fix this i need to double the cache again to 512 words , and with this the sdram has problems , probably doing the caches in other way can solve , but i think this is over my knowleg at this point . So i think that we need to stay with the 256 word cache and have the code zero scroll problem for now .
I was checking the "test" sound too , Jt51 only change two signals , but as Wark91 say it changes a lot internaly , and i do not know why but the previous version song better than the actual maybe i was doing something wrong and maybe I revert this .
Nuke OPM is only a emulator , I think there are 3 implementations at this point of Yamaha 2151 , puu's one , jt51 and ikaopm , I was thinking in put an option to change from jt51 to okaopm ( you need to reset in the change ) and check

[cobalt mulch]

I am completely inexperienced in this area, but would dual SDRAM help with this issue?

[uncut mantle]

Sure , but i can not enter in this hole ....

[cobalt mulch]

Please feel free to fix whatever you are able to, this core is something that no one has truly dedicated themselves to improving. I sincerely appreciate all the effort and dedication you have put into it.

[cobalt mulch]

Nice, the MIDI is working again now! 🥳

[thorny tree]

for your test and it is working also for MT32pi I will send the PR.

[uncut mantle]

I found the problem with the change to the new version of jt51 , i need to change one signal , now is working great , but in next "test" version you can chose between Jt51 an ikaopm , i think the two sound great now

[rancid compass]

Jotego would certainly be happy about a small PR

Maybe it fixes something in his arcade cores

[uncut mantle]

no , the problem is in my code , the version jt51 was realy old in the core , and one signal is diferent in new jt51 , so the problem is not in jt51

[rancid compass]

Ah, ok. Thanks anyway

[jovial basin]

Is MIDI for Final Fight working now? That would be marvelous! 😍
Now if only someone would update the mt32 pi to emulate the SC55 MIDI. 🙂
Detana Twinbee and Castlevania sound so fantastic with SC55.

[eager yacht]

I don’t believe this fixes that, but I could be wrong

[thorny tree]

It permits only to use midi over UART it is a regression from previous commit that changes this behaviour

[pale star]

Holy shit Chorensha got fixed

[rancid compass]

Within the game itself I didn't hear crackles with the sound test build

But I seem to hear some rare hiccups in the demo?

[eager yacht]

does the latest unstable include the fixes from @uncut mantle 's recent test builds?

[rotund parrot]

No

[uncut mantle]

no , i do not push my changes in oficial repo yet , i still trying to fix things

[eager yacht]

ah ok, I figured that was the case but I wasn’t sure!

[uncut mantle]

New test version of the core, in the coming days I will have very little time, and although there are regressions (atomic robokit and syvalion are broken), it improves some things.
ADPCM sound improvements, fixed some games that previously had no sound effects (dma problem, arianshu, asuka 120% ...). You can select in the menu between JT51 and ikaopm, need reset to apply.
Background in Garou SP and Garou 2 fixed (load graph in 65k colors and show in 256 colors).
Sprites in Baraduke fixed (chain array in 32-bit DMA mode).
Super Hangon load (int4, still background problems).
More regressions expected due to a lot of changes applied. The core still has a lot of problems.

[thorny tree]

Albconde for all those fixes 🙂
if you need more games having issue you can check Keepers (heavy graphical issues) and Die Bahnwelt (Color issues)

[cobalt mulch]

I have just finished Garou Special, and the only graphical issue I encountered was a minor line in the text layer on Terry’s stage. The music played flawlessly using IKAOPM, and while the sound effects for attacks and similar actions behave inconsistently—being absent in some stages such as Jubei’s and Andy’s—the overall improvement of the core is remarkable. The progress has been substantial and truly impressive. Thank you very much for the dedication and care you have put into this core; it clearly shows.

[verbal jolt]

was just testing garou out and experienced the same issues; a slight graphical error in terry's stage and missing sound effects (regardless of sound module), but it's otherwise an incredible improvement over how it was. illumination laser also seems near-perfect and the music in cyberblock metal orange EX finally plays at the correct speed! i'll do some more testing later, incredible work @uncut mantle

[thorny tree]

On Reinforcer (Zain Soft) game, during the demo mode, one scene is not correct on the bottom screen

[cobalt mulch]

The same stage in Garou 2, Terry’s stage, does not present this graphical issue.

Someone had asked whether the MIDI in Final Fight would be working; well, it is working without freezing.

[foggy pawn]

I should probably do fancy pins for this channel, huh

[cobalt mulch]

It would need to be tested with an MT32Pi, since I am using an SC-55 in MT-32 mode.

Previously, simply switching to MIDI mode would cause the game to freeze when it attempted to play a track.

[thorny tree]

On Phalanx (Zoom), if we choose Midi off, the gameplay is very slow and the fm sound is ok before plenty of crack sounds

[true flame]

You've already reached the pinnacle of your powers. Further improvement would be unsustainable.

[cobalt mulch]

Maybe need to use turbo like on ssf2?

[thorny tree]

With Midi on the speed is correct

[uncut mantle]

thanks for the testing . the core still need a lot of work. Please comment all the errors you find.

[cobalt mulch]

Finally, I managed to progress further this time, reaching level 15, and there were only the two graphical bugs that I had previously reported: one in level 2, affecting the background on the text layer, and another in the level where a lion is broken to start filling the area with water

[uncut mantle]

update the code in my repo if someone want to check

[cobalt mulch]

The game Arcus Odyssey is experiencing graphical issues in the graphics layer, and these issues also occur in the stable version

[icy shuttle]

that is new I think

[uncut mantle]

Arcus and Xenon i think the problems is the vhd image , have the same isuues in mame , px68k and xm6g , If you load form disk this errors are gone

[icy shuttle]

yeah, that's right - I knew I saw that look correct. Is the image in neon68k bad? I can't remember if I redid that one or not

[cobalt mulch]

In the stable version, the game Arianshuu had no APCM audio at all; it now correctly plays sounds such as gunfire and other effects

That is a very interesting shoot ’em up; I was not previously familiar with it.

[uncut mantle]

One question , do you think is interesting have a version with xdf support instead d88 , puu have two version , xdf is a lot of faster loading , i try to put the two in the core at same time with a selector , but do not fit in mister ( labs insuficent ) , maybe i do someting wrong .

[icy shuttle]

i would prefer xdf tbh

[cobalt mulch]

The game Catar Magical Block is experiencing graphical errors in both the graphics layer and the text layer on the title screen. During gameplay, there is also an issue on the side of the screen affecting the graphics layer.

It also happens in the stable version

In the game Chase H.Q., a portion of the track remains stuck in the background. I am not sure whether this is normal game behavior, but it also occurs in the stable version.

[cobalt mulch]

The game CosmoGang is losing its background music. This also occurs in the stable version. The graphics, however, are rendering correctly

In the game Cotton: Fantastic Night, changing the resolution causes the core to freeze; this also occurs in the stable version. There are some graphical defects during the lightning special attack. Additionally, the voices seem to be higher‑pitched, although I am not sure if this is the intended behavior in the original game

[thorny tree]

There is plenty of floppy disk format, could be handle on main Mister side like c64 if it tooks lots of logic.
D88 and others are also used on other Japanese computers so it may could be shared between cores

[verbal jolt]

i've been using .hdf for all games so far, personally. haven't tinkered with floppy formats at all

[thorny tree]

D88 implentarion on x68000 took 2 times than normal to read floppy.
Xdf implentarion fixed this part.
Also d88 version of R-Type has some graphical issues not present in xdf version.

[cobalt mulch]

I've never used any D88 games

If using the XDF format is better than D88, then it should be implemented in the core instead of D88

[foggy pawn]

What is the difference between d88 and Xdf?

[thorny tree]

https://nfggames.com/forum2/index.php?topic=7444.0

[cobalt mulch]

In the game Strider Hiryuu, APCM audio is now working correctly. However, when changing the resolution to 512×256 31hz, graphical issues appear in the background and the entire options menu becomes visually corrupted.

HDF bug?

At 15 Hz, it works without any graphical errors, at least in the first stage

[cobalt mulch]

In the game Genocide, the background layer is being overlaid on top of another layer. When Graph G0 is disabled in the debug options, the background renders correctly

[thorny tree]

This issue is present again
https://github.com/MiSTer-devel/X68000_MiSTer/issues/40
Your Previous build it was fixed.

[thorny tree]

MTEST for SYSTEM

A-Jine is broken in game and the start screen is not ok
MiSter screenshot

Mame screenshot

Naious -> Hangs after the start screen
https://imgur.com/xRuQJq5

Dragon Buster -> First screen is not correct resolution

Lagoon - hangs during the boot of the game

Downtown Nekketsu Monogatari - Layer issues

[dull kindle]

Not new (Arcus Odyssey)

[thorny tree]

HDF is not ok. It works with floppies

[thick kayak]

So glad to see some progress on this core. Thank you to all involved! I've definitely noticed improved sound in a few games I've tested.

While we're noting issues...

Nekketsu Koukou Dodgeball Bu has major graphics glitches. I don't think it's ever worked well on any Mister core.

Yamihime has a similar issue to Genocide above where a background layer, the green vine things (on Graph G3) cover up the sprite/text layers. I don't think this is a regression either... the issue was in previous cores.

[thorny tree]

Thunder Blade (Sharp, SPS) -> Glitch line (see screenshot)

After letting some times, the error is printed and the game hangs.

[thorny tree]

Pacland -> The top line is on the bottom of the screen

[thorny tree]

Bomber Man - Issue with outline of the letters of Bomber Man on Title Screen

[cobalt mulch]

I am not sure whether this is normal, but in Street Fighter II the music seems to lose some instruments when, for example, a special move is performed. It feels as though the sounds overlap, and I do not know whether this behavior is normal on the original hardware

[cobalt mulch]

I have just finished playing SF2 and did not notice any graphical issues. The only concern is that the sound effects of the attacks seem to overlap with some of the musical instruments, and I am not sure whether this is normal behavior. Additionally, the sound of the attacks appears to be somewhat high-pitched.

In Super Street Fighter II, this issue of the APCM overlapping the FM instruments does not occur. It may be normal behavior in the Champion Edition

I am not sure whether this is a graphical issue or if it behaves this way in the original X68000 version; it would need to be verified. However, in version 2 on hit Combo, the top portion appears to be cut off.

I have just finished playing SSF2 and did not encounter any issues during the game, except for the number “2” in the hit combo. At the end of Chun-Li, I noticed something unusual

When I disable the sprite layer, this appears

[foggy pawn]

lol gewd faece

[cobalt mulch]

Up to the second stage, Final Fight is working perfectly, including sound, graphics, and music. Only this part of the bonus stage raises some uncertainty, as I am not sure whether this behavior is normal

[icy shuttle]

I wonder if mt32pi works on final fight

[rotund parrot]

Mt32pi does work with Final Fight on a real X68000. Not sure if the core works now

[icy shuttle]

yeah, it used to crash in the core

[cobalt mulch]

I tested it with an SC‑55 and it worked correctly. Previously, even when using it, changing the option from internal to MIDI would cause the core to freeze. Have you tested it now with MT32‑Pi?

[icy shuttle]

i haven't, but it sounds like it's working properly now

[cobalt mulch]

I have finished Final Fight; I have been waiting since 2019 for this core to reach this level. 😅

Super Hang-On has a black line on the BG1 layer

It also froze shortly afterward. 🫠

It appears to be related to the graphics layer.

The game V-Ball has multiple graphical and audio errors.

[verbal jolt]

active shoot doesn't seem to function properly right now. using .hdf, projectiles freeze and the game slows down exponentially. it also exhibits the same issue as garou densetsu special where certain sound effects don't seem to play

also, i'm unsure if this is intended behavior, but i noticed that some games won't let me start playing until the attract mode/demo plays at least once. illumination laser is a good example of this, also using .hdf

[cobalt mulch]

In the game Overtake, a blue layer appears at the bottom of the screen, which I believe should not be there. When the text layer is disabled, it disappears.

The game Overdriver has graphical issues on the title screen.

The game runs well overall; the issue appears to be limited to the title screen.

[cobalt mulch]

The game Momotarou Densetsu has an issue with the text layer on the title screen, which appears entirely pink. When the text layer is disabled, the screen displays correctly.

[icy shuttle]

I just want to say how much I love the activity in here, friends. Thanks for your work @uncut mantle, and thanks for the thorough testing @cobalt mulch (and @thorny tree of course)

[cobalt mulch]

I can only express my gratitude. I have wished for this core to reach a perfect state for a long time, and it is gratifying to finally see someone giving it the care it deserves.

[verbal jolt]

albconde is a saint

[icy shuttle]

you and me both, pata - you and me both

[icy shuttle]

big +1

[verbal jolt]

i haven't been taking any photos of bugs i've found, but i'll start doing that whenever i encounter anything serious. just been going through and testing random games so far

[thorny tree]

We need also to thanks Puu-san, Kitrinx, Sorgelig, Dshadoff, Rebel40 🙂

[thorny tree]

Patch for AJAX
https://x.com/x68trap/status/2013783154310451653?s=20
Using megadrive pad and other fixes

トラップ (@x68trap)

A-JAX for X68030 パッチ
https://t.co/rLiEQyViBq

X · Jan 21, 2026, 1:18 AM

[thorny tree]

Double Eagle (Artdink) has graphical issues on menu (game still running)

Etoile Princesse - During the intro if we push a button this screen appears

[thorny tree]

Full Throttle has graphical issues and also sounds (ADPCM) are not right

[thorny tree]

https://x.com/bcc2528/status/2023426563858165913?s=20

BCC🌐 (@bcc2528)

X68000版Quake
一応デモが動くところまできた。
ポリゴンが欠けたり、ランダムでフリーズするのは現バージョンのXM6のFPU実装の可能性もあるんであとで確認するけど、68060 60MHzでも重すぎる。
効果は無いと思うけど、Quake 3の高速逆平方根アルゴリズムは取り入...

▶ Play video

X · Feb 16, 2026, 3:57 PM

[thorny tree]

Geograph Seal (Exact) -> Black screen during the boot (with Albconde build). It is working with last unstable.

[uncut mantle]

thanks for the testing , as i said previously i have little time this week , only fix for now is the fade efect in Cho Ren Sha background . Some of the problems you found are relates to change color mode , than the previous in Garou for this i need to go in the rute to change the cache from 256 words to 512 words , this is complicated with the sdram . Some games load all data from disk to memory in 56k color mode ande made some transformation in memory to show 256 o 16 colors with 256 words cache you can not solve all cases at this point .

[rotund parrot]

I couldn’t get this patch to apply to any of the public copies of Ajax floating around. I suspect that the version going around has already had the copy protection cracked, which is very common for X68000 dumps unfortunately

I’m actually considering buying a copy of A-Jax to dump myself

I’d like to try the patch

[thorny tree]

Ok ...

[cobalt mulch]

@uncut mantle I have a question: what is the speed of turbo mode?

Is it 16 MHz?

[thorny tree]

It is more than 16 MHz maybe around 20 MHz

[cobalt mulch]

I am not sure whether supporting multiple clock speeds would be possible, but for example: 10 MHz, 12 MHz, 16 MHz, 20 MHz, and 25 MHz.

Some games behave unpredictably when turbo mode is enabled. If additional speed modes were available, it would be possible to better fine‑tune compatibility beyond the standard 10 MHz.

[thorny tree]

Regarding patch of ajax it permits to fix sound with 16MHz so also it could be the game not designed to run above 10MHz

[uncut mantle]

Turbo is 20 Mhz yes , maybe when the core is more stable at this point i need to put the sdram to 120 Mhz to allow the 512 cache words work correctly , we will see if this change create problems with some misters ....

[thorny tree]

Also information that could be interesting

• First models of X68000 used a clone of 68000 - Hitachi HD68HC000 10 MHz
• X68000 16 MHz model used a Motorola 68000 CPU 16 MHz
• Later model Motorola MC68EC030 25 MHz but compatibility with games is not so good

[verbal jolt]

this might be a dumb question, but what's the purpose of the NMI and power buttons in the core menu?

[uncut mantle]

for an user not too much , nmi launch an interruption is util for develop or test .
In some models the original machine have this buttons.
Thanks to the nmi i find that the problem with super hang on was an interruption , in the oficial core if you load the game , when the game do not load anymore , if you select nmi , you will see a mesage inthe x68000 if you push "a" the game load again and you can see the loading screen .

[verbal jolt]

thank you for explaining it to me 🙂

[cobalt mulch]

I have just tested Street Fighter II on the XM6 emulator, and the ADPCM sound causes some musical instruments to disappear when a special move is performed. I do not recall whether this occurs in the arcade version; it is a rather strange behavior.

[rugged geode]

So, is it the drums that disappear?

[cobalt mulch]

Yes

[rugged geode]

My guess is that is how it is implemented. I think they are treating their sample channel as single voice and not mixing pcm sounds together to save cpu.

[cobalt mulch]

https://www.youtube.com/watch?v=iGaD4mQf8Zk

In this video, it is stated that the footage is from the real hardware; I did not notice the drum sound disappearing when a special move is performed.

https://www.youtube.com/watch?v=LmvBAc8aBKY

In this video, the absence of the drum track is already evident, even with the poor audio quality.

I believe that when using a MIDI module, the drum track should not be missing.

[rotund parrot]

Looks like the channels also disappear on my X68000 Compact system

Here's with an SC55

[cobalt mulch]

So it really is normal behavior.

[thorny tree]

Working great on the core 🙂

[south estuary]

Should we maybe split off a thread for bug and compatibility reports? Every time I come here lately it's a giant list of games that don't work right.

Or maybe that's the purpose of this channel? What else would we talk about?

[rugged geode]

So, I knew this guy from irc once that knew a heck of a lot about the x68000. I think he even maintains a website for it. We could use the channel to aggregate questions to email to him 😉

[cobalt mulch]

Much like the Saturn channel, I believe that this would be the intended purpose, wouldn’t it? Especially considering a core that took years before someone actually stepped in to properly address and fix its issues. Every time the core is updated, it seems that testing and reporting will continue to take place here, particularly since he explicitly requested it.

[south estuary]

Don't you fucking dare, haha

Just seems to me that this channel would be better served as a 'I need help' channel that isn't overwhelmed by endless screenshots of games with layer problems. That's helpful, sure, but it's been a lot lately, and if that happens every time there's an update, it'll drown out everything else regularly.

[verbal jolt]

i mean, all threads for in-development cores start off like that. i don't think X68000 is in a spot where people should be asking for general help just yet anyway

though someone could start a spreadsheet for all the bug reports, if they wanted

[true flame]

I love my games with layering problems, and when I need a recharge I visit this thread for them sweet sweet screen shots.

[eager yacht]

I would say that it’s fine to share those bugs here as long as it’s helpful for @uncut mantle or any other brave soul looking into these issues. If another venue like github or a sub-thread would be better, then it should be moved there

[primal sandal]

Or, if there are various new versions coming from time to time, is it possible to have a sub-thread per version ? It's kind of hard to keep track of which versions have which issues if there is any chance of regression

[uncut mantle]

regressions are spected at this time , for example the layer priority and transparencies are not clear , the original code , mame ans px68k are not perfect and with different approach ,

xm6g ( no source code ) is the best but for example transparency in Keeper are wrong too

[cobalt mulch]

XM6 has only made old source code available

Kamada-san’s Java version is open source, if that helps.

https://stdkmd.net/xeij/source/

[thorny tree]

there is some test programs also
https://stdkmd.net/xeij/miscfiles.htm#testprogram

[lethal scroll]

The X68000 has only 1 ADPCM audio channel and the loss of drums is normal in a 10mhz 68000. There is a way to force the game to perform 4 ADPCM channels ( need at least 16mhz CPU ) and these 4 channels are used by default if the computer is an X68030.

I have an x68000 XVI and by modifying the autoexec of games like SFII, SSFII, Mad Stalker or Nemesis 90 is possible to have 4 ADPCM channels without having slowdowns. The x68030 activates them by default.

On 10mhz machines, forcing more ADPCM channels will resullt in massive slowdowns.

[cobalt mulch]

Is additional channel support being achieved through software‑based emulation, since ADPCM itself only supports a single channel?

For this reason, I still believe that the ideal approach would be to offer multiple turbo-mode speeds: 10, 12, 16, 20, 25, and 30 MHz.

[lethal scroll]

They are generated by software

Brute force

[rotund parrot]

Ah nice, I didn’t know that! Will have to try it out. I have a 24mhz capable Compact

[cobalt mulch]

What should be added to the autoexec file in order to enable the additional channels? Since the MiSTer core reaches 20 MHz, the performance might not be too poor.

[rugged geode]

should be something to do with pcm8.x

https://eab.abime.net/showpost.php?p=1782878&postcount=520

[cobalt mulch]

Does anyone know how much RAM the X68k core has?

I found a post on Reddit containing a Dropbox link to a 7z archive with patches for several games.

https://www.reddit.com/r/sharpx68000/comments/1czg7ey/how_to_use_this_famous_pcm8_driver/

[rugged geode]

I have used the switch program to set max ram and it seems to work

Should be 12mb

[cobalt mulch]

Wow, I applied the patch to the start.bat file and placed the file on the HDF, and now the music is playing perfectly

However, it is now necessary to enable turbo mode in order to be able to play

[thorny tree]

other patches with PCM8
http://retropc.net/x68000/software/games/patch/

[thorny tree]

also Dennou club 22 has patch for super hang on to increase size of moto and character

[cobalt mulch]

Just to confirm, I managed to obtain a Raspberry Pi and set up an MT32-Pi, and Final Fight handles MIDI playback without any issues

[rotund parrot]

Nice!

[cobalt mulch]

I can hardly believe how much the core has improved in such a short period of time.

[verbal jolt]

it's really incredible

[cobalt mulch]

@uncut mantle Thank you so much for all the work you've done

It took seven years of waiting for the core to reach this level of maturity and functionality.

[languid mango]

there any documentation i can follow up on for neon68k? can't seem to get anything to load. just get a black screen

[tawdry creek]

https://neon68k.com/docs

[languid mango]

yeah, i've gone through that, but nothing that seems to match my issue

[tawdry creek]

Well that's the docs

Perhaps tell us about your MiSTer setup

[languid mango]

pretty standard setup, with misteraddons' aluminum case, 128mb ram, uhhh not sure what else i can think of, i'm not using direct video

[tawdry creek]

Where are your games stored?

[languid mango]

x68k core was working with my previous set of games, but when i started using neon68k is when things went awry

exactly as the docs have it

3 folders, copied to the root of the sd card

[tawdry creek]

How are you loading the games? Are you loading them from the X68000 menu?

[languid mango]

yep

[tawdry creek]

Hmm, my setup doesn't have it under /media/fat/games/X68000/media like that, its /media/fat/games/X68000/HDF

Hmm, perhaps I changed that though 🤔

[languid mango]

probably because that's the layout when i look at the zip file for a game

[icy shuttle]

what are you outputting to?

[languid mango]

in what relation?

video?

[icy shuttle]

yeah, are you just going to a TV? a computer screen? CRT?

HDMI?

[languid mango]

hmmm, capture card through hdmi

[tawdry creek]

aaaah

So maybe not a pretty standard setup 😁

[languid mango]

hdmi isn't standard?

[icy shuttle]

try doing straight HDMI

[tawdry creek]

Through a capture card isn't

[icy shuttle]

without the capture card

just to see if that’s gumming up the works

[languid mango]

i mean, it works here and other cores and like i said before, my previous set of games with the x68k core worked

[icy shuttle]

which version of neon68k did you unpack? scaler? non scaler?

[languid mango]

MiSTer Upscaler

this is what i get when i load up a neon68k game, i have to open up the OSD just to confirm it didn't crash or anything

[icy shuttle]

that looks like a bios problem. Can you load up the 68k core by itself, choosing it from the _Computer folder?

[languid mango]

sure

hmmmm, black here too now

OSD comes up

[icy shuttle]

show me the root of /games/X68000

[icy shuttle]

hrm

just for poops and grins, delete boot.rom and re-run update_all

[languid mango]

sure

[icy shuttle]

I’m not in front of my mister, and it’s been an age since I set up neon68k, but I don’t think you have to specify a bios like amigavision

[languid mango]

confirmed it got it

[icy shuttle]

still black screening?

[languid mango]

about to try

[languid mango]

so i recently set this up as well, i didn't need to touch bios, it just booted into the frontend

still black on the x68k

[icy shuttle]

does it say loading boot.rom when you boot it?

[languid mango]

doesn't look it

[tawdry creek]

And what happens if you load one of the HDF files?

[icy shuttle]

ehh, it could be happening during the signal drop

yeah, that is a good test. you boot the core and then load an HDF manually

[languid mango]

still nothing, signal drop didn't occur

[icy shuttle]

ok, in your /config folder, there are going to be a shit ton of files for x68000. Move them out of there and then boot the core normally, without one of the neon 68k mgls

[languid mango]

alrighty

hmmm, nothing

let me blow out the rbf and boot roms, start from square one

hmm, and still nothing. all the cfgs would be prefixed with x68k right?

[icy shuttle]

yeah. the thing that is throwing me off is the menu. seeing that means video is working and the core booted. not seeing the 68k boot up means it’s not seeing a bios. everything is on the SD card, yeah? or are you using cifs?

[languid mango]

all local to the sd card

[icy shuttle]

The last thing I can think to try is to use this core: #1351743248850227360 message

Someone on the forums had a similar issue where the core wouldn't load (but they were getting a black screen) and this core fixed it.

I assume there was something else at play there, and this won't fix your problem, but my brain is gravy at this hour, so I don't have a lot of ideas

good ones, at least

[languid mango]

i'll give it ago

[thorny tree]

CPU speed check for x68000
https://modelrail.otenko.com/assets/x68000/spdcheck.zip

[lethal scroll]

It will work perfectly

[languid mango]

ok, might have tracked down the issue, there was an errant media folder where it didn't belong that it somehow ended up in the root folder and i guess that was mucking things up? somehow?. copying over the neon68k stuff now

[tawdry creek]

@languid mango there's also this script which can do all the downloading and putting into place on the mister for you: https://github.com/Neon68K/Automatic-Updater-Script

[languid mango]

well i'm going step by step, i'm trying to figure out how that folder got there

[tawdry creek]

When you say there was an errant media folder in the root folder do you mean a literal media folder in /root or in the root of /media/fat?

[languid mango]

there was a X68000 folder in /media/fat

and i have no idea why

because there was and X68000 folder in games

[tawdry creek]

X68000 or _X68000?

[languid mango]

X68000 no underscore

only thing i can think of is i fat fingered something last night

[tawdry creek]

That's why we have the updater script so you can't do that 😁

[languid mango]

well i didn't know the script existed until 5 minutes ago

[tawdry creek]

https://tenor.com/view/now-you-know-the-more-you-know-no-kidding-no-shit-obviously-gif-1342210079425559607

[languid mango]

no mention of it on the neon68k site which i assume for legal reasons

[tawdry creek]

@icy shuttle hey reckon we should update neon68k.com to mention the n68k updater or was that intentional?

[languid mango]

hate myself and want to die

[tawdry creek]

Don't do that, just enjoy the rad selection of X68000 games instead

Especially if you have a mt32pi, then the music is

[languid mango]

yeah, yeah. i was using a different setup before. neon68k came into my realm of consciousness when i learned about oneload64 and amigavision

[icy shuttle]

I, uh….can’t remember. I’ll ask limi

[thick kayak]

Speaking of neon68k, I’ve been testing and converting more games to over .hdf that could possibly be included. Many of them are obscure / doujin games of varying quality - so I’m sure not all of them are worth including, but I’ve found some gems as well.

Also, if anyone is missing a game or wondering if one works, let me know. I’ve tested a bunch.

@grand mulch Bonanza Bros seems to work although I haven’t played too far into the game.