MiSTer PCXT

[remax]

Yeah, i met similar errors but can't pinpoint to what exactly it relates.

[suww37]

Juko is fast for sure. I have a somewhat solid setup with DOS 5, 4DOS as the command shell loading into emm, FREESP, UMB+LTEMM, MyMenu and I just transferred about 100 games to it. I am slowly working through testing but I am still running into Drive C disk errors that force me to restart the core. simple things like editing config.sys or autoexec.bat can cause a general failure reading drive c.

I've been debugging serdrive and I can see some crc errors but I can't figure out if it is me or the core. I moved all my setup local instead of over the network and that helped a bit.

Are any of you having stability issues with the disk still? "General failure writing drive C. Abort, Retry, Ignore, Fail?"

It seems to only happen when I use an app that needs to swap to memory with LTEMM loaded, but I am not positive on that yet.

I am too. When I run the game and go back to Dos, I get an error and I have to restart the core.

[flynnsbit]

Couple of testing sessions, I don't think it is LTEMM though it seemed less likely to happen without it running.





11 min in on Testing 2 you can see the drive failures. It could simply be the instability of the COM hosting of the drive image that is causing it. It will be really hard to build anything for people with the disks like this. (again, unless someone has figured out how to stabilize it and this is just me).

Super cool progress on the core.

I'll pause for now on my side to let things stabilize a bit or if someone takes a crack at hosting the .vhd to 300h through the HPS side instead of the COM port.

[spark2k06]

I have added the possibility to activate the Tandy graphics from the OSD, until now it didn't work because they were activated by register. A Tandy ROM is required for this to work, as the BIOS needs to support the new video modes:

https://github.com/spark2k06/PCXT_MiSTe ... ec3e6ad131


The BIOS from PCjr doesn't work, you have to download the first revision of Tandy 1000:

http://retrograde.inf.ua/tandyroms.htm

In addition, there are also problems with the keyboard in this BIOS, so you should add in the AUTOEXEC.BAT whatever you want to test, until it is solved.

On the other hand, there is another problem with the Tandy sound, the sound module needs to be checked as it is not working properly. There are also the ghost lines that appear in the Tandy BIOS text mode, more things to fix.

[flynnsbit]

Archive.org "tanfiles".

You're welcome.

=== Games:

Most of the games are preconfigured for enhanced Tandy/PCjr graphics and
audio. Furthermore, most of these games should run fine on any Tandy 1000,
although some may be sluggish on a 4.77MHz processor.

Games that *required* unreasonably fast Tandys to run at all, like a 12MHz
80286, have been removed. This includes some of the more demanding 3D titles,
like Stunts.

There is a directory called GAMES/TECH which has a .zip of games that have
interesting programming or optimized special-effects. They are included as a
curiosity.

[LamerDeluxe]

Paging @kathleen

[thorr]

Very excited to see the Tandy development start! Thanks a bunch.

[flynnsbit]

XT IDE configured to boot directly from the COM1 address and skip auto detection of the other controllers. It speeds up the boot process a tiny bit for now. Of course this will need to be re-done when HDD/FDD is added. You can regenerate your base bios with this using the python script if you want it built in. I changed the color scheme so you would know it is the one that removed the other controllers.

[Newsdee]

I can't wait for this to get floppy or hdd support

Are both implementations related, or is it it easier to just do floppy first? That would allow having simple boot disks to make it easier mount an hdd via serial.

[spark2k06]

I have made a number of improvements to the prebeta 1.3:

https://github.com/spark2k06/PCXT_MiSTe ... e_beta_1_3

Default loading of BIOS + XTIDE in EC000 from boot.rom is now possible, the first 64Kb will be loaded in F000 and up to 16Kb thereafter in EC00.

boot.rom up to 64Kb + 16Kb for XTIDE
Essential for BIOSes that require the whole F000 segment, like Tandy or IBM.

Dummy LPT1
Necessary to pass tests in some BIOS, such as the latest Tandy 1000 revisions.

Update of ROM download scripts
XTIDE support is given to those ROMs that must be loaded on EC000, and Tandy ROMs are added.

XT IDE configured to boot directly from the COM1 address and skip auto detection of the other controllers. It speeds up the boot process a tiny bit for now. Of course this will need to be re-done when HDD/FDD is added. You can regenerate your base bios with this using the python script if you want it built in. I changed the color scheme so you would know it is the one that removed the other controllers.

Interesting! I suggest you make a pull request on the prebeta 1.3 branch, replacing the file with the same name as the folder with the one you share. This way, we have a record of the users who contribute to this project

I can't wait for this to get floppy or hdd support

Are both implementations related, or is it it easier to just do floppy first? That would allow having simple boot disks to make it easier mount an hdd via serial.

The logical way would be floppy first, but the really interesting way would be ide via VHD or physical SD card. However, even if we have floppy, as long as we don't have IDE, the initial mount from the script in the Linux part of MiSTer is essential.

[wark91]

With the bios Tandy 1000 HX, the core starts to show memory check ! Thanks

[spark2k06]

With the bios Tandy 1000 HX, the core starts to show memory check ! Thanks

Correct, but then it won't boot because it fails the keyboard test. The only BIOS that ignores keyboard operation is the first revision of the Tandy 1000, however, then the keyboard cannot be used.

[Newsdee]

as long as we don't have IDE, the initial mount from the script in the Linux part of MiSTer is essential.

Where should I get serdrive for the linux side?
Is it just a binary from the xtide page?

Edit: answering my own question... it's in this post
viewtopic.php?p=54369&hilit=serdrive#p54369

[spark2k06]

Edit: answering my own question... it's in this post
viewtopic.php?p=54369&hilit=serdrive#p54369

This thread is already very big, but I think it deserves your attention from beginning to end, however, in the Mounting the disk image section of the project's README I also discuss it

[breiztiger]

https://www.xtideuniversalbios.org/chan ... ersalbios/

last build for xtide

[breiztiger]

config.sys

with LTEMM you can add /n to pass the test

DEVICEHIGH=C:\UTIL\LTEMM.EXE /p:A000 /x /n

[breiztiger]

Landmark 2.0 work but report as a 2mhz

landmark 6.00 crash

a strange thing ... pcxt core seem to "view" keyboard windows keypresse ?!? when landmark crash, press windows key beep

CCT 1.00 detect the clock rate at 0.0 MHz

[Newsdee]

Finally joined the cool kids club managed to get pre-beta 1.3 running with Juko ST BIOS and a custom 20MB hdd.

A few comments:

• CGA mode ignores forced_scandoubler=1, so I can't use the core with an LCD VGA screen. Can that be fixed?
• MDA mode does work on the VGA screen (probably because it's already 31khz?)
• For CGA composite, maybe you can reuse the logic in the Apple II core? (not perfect but might work well enough)
• Jill of the Jungle hardware detection seems to hang
• I've never realized this, but many games look MUCH better on Amber monochrome mode. Maybe they were optimized for it!

I've only tried a handful of games, but it's so great to see early CGA games running so nicely at the right speed,
without having to fiddle with settings or slowdown TSRs!

prehist.png (10.32 KiB) Viewed 5019 times

goody.png (8.83 KiB) Viewed 5019 times

digger.png (7.2 KiB) Viewed 5019 times

[spark2k06]

Finally joined the cool kids club managed to get pre-beta 1.3 running with Juko ST BIOS and a custom 20MB hdd.

A few comments:

• CGA mode ignores forced_scandoubler=1, so I can't use the core with an LCD VGA screen. Can that be fixed?
• MDA mode does work on the VGA screen (probably because it's already 31khz?)
• For CGA composite, maybe you can reuse the logic in the Apple II core? (not perfect but might work well enough)
• Jill of the Jungle hardware detection seems to hang
• I've never realized this, but many games look MUCH better on Amber monochrome mode. Maybe they were optimized for it!

I've only tried a handful of games, but it's so great to see early CGA games running so nicely at the right speed,
without having to fiddle with settings or slowdown TSRs!

prehist.pnggoody.pngdigger.png

Try setting vga_scaler=1, vsync_adjust=0 and see if it works.

[Newsdee]

Try setting vga_scaler=1, vsync_adjust=0 and see if it works.

Sure that works for now, but I lose the ability to use dual resolutions.
(I mostly keep HDMI at 1080p for my video capture card)

I notice that the standard MiSTer video_mixer is commented out in the code... integrating that at some point would fix it.

[spark2k06]

Try setting vga_scaler=1, vsync_adjust=0 and see if it works.

Sure that works for now, but I lose the ability to use dual resolutions.
(I mostly keep HDMI at 1080p for my video capture card)

I notice that the standard MiSTer video_mixer is commented out in the code... integrating that at some point would fix it.

I think I have already tried to make changes in this respect and it didn't work well, possibly just out of ignorance. I suggest you to test by yourself, and if you find a good solution, make a pull request in the prebeta branch, after review I will gladly accept it

[Newsdee]

Couple of testing sessions

I tried to use MyMenu from the Shareware Pack as a test... but it really didn't like the core's DOS environment.
Are there settings I could tweak or would it require changes to MyMenu code?

(I imagine that I'd have to give up long filenames, for example)

mymenu.png (18.5 KiB) Viewed 4802 times

[flynnsbit]

Couple of testing sessions

I tried to use MyMenu from the Shareware Pack as a test... but it really didn't like the core's DOS environment.
Are there settings I could tweak or would it require changes to MyMenu code?

(I imagine that I'd have to give up long filenames, for example)

mymenu.png

It's LFN=T being set and DOSLFN or LFNDOS not running. If you set LFN=F in the mymenu.ini it will fix it. BBond and I are working through some stuff with MyMenu to see what is optimal for the core.

FYI, disk errors are the biggest issue I see. You can bomb the serial connection and cause the keyboard to basically start only accepting every other key by running ATTRIB on the file system.

[Newsdee]

I suggest you to test by yourself, and if you find a good solution, make a pull request in the prebeta branch, after review I will gladly accept it

I tired... but no luck I get a blank screen (OSD works just not the internal video)

Code: Select all

wire video_MDA = status[4];
wire [2:0] scale = status[16:15];
wire [2:0] sl = scale ? scale - 1'd1 : 3'd0;
wire       scandoubler = (scale || forced_scandoubler);
wire [7:0] R,G,B;
assign R = {video_MDA  ? r : raux, 2'b0};
assign G = {video_MDA  ? g : gaux, 2'b0};
assign B = {video_MDA  ? b : baux, 2'b0};
video_mixer #(.LINE_LENGTH(650), .GAMMA(1)) video_mixer
(
	.*,
	.hq2x(scale==1),
	.freeze_sync()
);

[spark2k06]

I suggest you to test by yourself, and if you find a good solution, make a pull request in the prebeta branch, after review I will gladly accept it

I tired... but no luck I get a blank screen (OSD works just not the internal video)

Code: Select all

wire video_MDA = status[4];
wire [2:0] scale = status[16:15];
wire [2:0] sl = scale ? scale - 1'd1 : 3'd0;
wire       scandoubler = (scale || forced_scandoubler);
wire [7:0] R,G,B;
assign R = {video_MDA  ? r : raux, 2'b0};
assign G = {video_MDA  ? g : gaux, 2'b0};
assign B = {video_MDA  ? b : baux, 2'b0};
video_mixer #(.LINE_LENGTH(650), .GAMMA(1)) video_mixer
(
	.*,
	.hq2x(scale==1),
	.freeze_sync()
);

As I said, I have been trying to improve it, but for the moment we will have to make do with what we have. It will get better in that respect.

[suww37]

I found something strange when I played the Tetris (Spectrum HoloByte) game. That's what breaks the graphics. The first time you upload a core with a hdd.img mount, I didn’t have this symptom. There seems to be something wrong with the core, so please check it out.

[spark2k06]

I found something strange when I played the Tetris (Spectrum HoloByte) game. That's what breaks the graphics. The first time you upload a core with a hdd.img mount, I didn’t have this symptom. There seems to be something wrong with the core, so please check it out.

Please try the versions in the release folder and tell me just which one starts to fail.

[suww37]

I found something strange when I played the Tetris (Spectrum HoloByte) game. That's what breaks the graphics. The first time you upload a core with a hdd.img mount, I didn’t have this symptom. There seems to be something wrong with the core, so please check it out.

Please try the versions in the release folder and tell me just which one starts to fail.

Now that I check it out, it doesn't matter if it's vesion. The level 6 and level 9 backgrounds are especially crashed.

[spark2k06]

beta 1.3

https://github.com/spark2k06/PCXT_MiSTe ... 45d237b7a4

• Unified chipset clock at 100 MHz.
• Changed read signal to uart module.
• Changed cen_opl2 signal.
• Improved access speed to SDRAM.
• Control sdram refresh execution timing.
• Fixed KF8237.
• Wired between Timer 1 output and DMA0 request.
• Fix VRAM CGA and loader for XTIDE.
• IBM5160 BIOS downloader.
• Fix indentations in make_boot_with_ibm5160.
• Tandy graphics selectable from the OSD.
• EMS pages frame update.
• fix a comment on addressable memory.
• boot.rom up to 64Kb + 16Kb for XTIDE.
• Dummy LPT1.
• Update of ROM download scripts.
• Simple improvements to PCXT.sdc.
• Correct use of address_enable_n signal in ports and memory accesses.
• Initial improvements in Tandy sound implementation.
• Improvements to the implementation of Tandy video

In this version, in addition to the Tandy mode, the palette switching feature also works:



The graphics module has some room for improvement:

• Text mode based on 225 lines instead of 200, has one duplicated line per character that shouldn't be.
• Tandy's 320x200x16 video mode has two additional vertical lines that should not be there.

I start a new prebeta 1.4 branch:

https://github.com/spark2k06/PCXT_MiSTe ... rebeta-1.4

@kitune-san, I have tried to improve the implementation of the tandy sound module by following a little bit the methodology you have used in other modules:

https://github.com/spark2k06/PCXT_MiSTe ... 0052703e0f

Despite some improvement, as you can see in the video, the sound is still distorted... you can compare it with another video where you can see how it should sound:

https://www.youtube.com/watch?v=ersh8Yz-CC8&t=35s

Can you think of a way to fix it?

On the other hand, a user told me that although the direct clock of the sdram works well, in many of the MiSTer cores there is a small offset as explained here:

https://github.com/mist-devel/mist-boar ... on-4-sdram

What do you think about it?

[kitune-san]

@kitune-san, I have tried to improve the implementation of the tandy sound module by following a little bit the methodology you have used in other modules:

https://github.com/spark2k06/PCXT_MiSTe ... 0052703e0f

Despite some improvement, as you can see in the video, the sound is still distorted... you can compare it with another video where you can see how it should sound:

https://www.youtube.com/watch?v=ersh8Yz-CC8&t=35s

Can you think of a way to fix it?

I haven't seen this thread in a while.
It is great progress.

One thing I noticed.
If you need a ready signal, you can do the following.

in Chipset.sv.

Code: Select all

    READY u_READY (
        ...
        .io_channel_ready (io_channel_ready & memory_access_ready & tandy_snd_rdy),
        ...

And I don't think the edge signal for the write signal within Peripherals.sv is necessary. Because it is done within sn76489_latch_ctrl.vhd.