clb wrote on 2024-12-06, 10:22:

... it did make me wonder if there were ever any games that would do something with the border ...

Here are a couple more:

Gunship
Keef the Thief
Pirates!
Sword of the Samurai

Crystal Caves used the border to determine whether all gems were collected or not in a level.

Catacomb 3-D used it when the player gets hurt.

Hey all!

New year, new cards: we are getting closer to processing the second shipping batch of CRT Terminators, so for those in the order queue, look out to your email inboxes in the coming week or two.

Additionally, we worked through the Christmas days to prepare for the first firmware update for CRT Terminator. I am currently doing QA on the release candidate, and writing up a new documentation page in the manual for how to flash a new firmware to the adapter. There will be plenty of bug fixes and improvements packed into the firmware based on feedback and experience so far in the field.

Another update: I now got around to releasing a modded version of Foray in the Forest optimized for CRT Terminator.

I wrote a separate post about that, since this announcement thread is getting a few miles long grinning face

Check it out at -> Commander Keen: Foray in the Forest - CRT Terminator Edition for stutter-free 70Hz gaming

The new firmware version is looking really good. I don't yet have a ChangeLog page set up on the site, so here is a sneak peek to the changes (> 200 commits, >2 months of development) that the new firmware will bring over the first shipped firmware:

DV1000-2025.1.5
---------------
- Improved upscaling for Pinball Fantasies and Pinball Dreams in their High Res modes. (a85df770)
- Improved video signal sampling phase detection stability on Hercules Stingray Pro (ARK1000PV) (f4efaf67)
- Added video signal sampling phase indicators to Developer HUD as a sequence of four +/- characters (for HSync, Vsync, Disp.Enable and Data) (378237c5)
- Improved video signal sampling phase detection stability on Diamond SpeedStar 24X Tseng ET4000AX in DOS 320x200 and 640x350 video modes. (a8ba41d9)
- Improved video signal sampling phase detection stability on video modes that exceed 48.5 MHz pixel clock. (a8ba41d9)
- Dropped maximum generated output video refresh rate from 75.25Hz down to 75.00Hz for improved sync compatibility on displays with max 75Hz refresh rate. (3d0744a3)
- Improved video signal sampling phase detection stability for the Display Enable signal, by accounting for long time spent low during vblank. (63277314)
- Rewrote the video sync code to give precedence to Display Enable signal, and not get confused by timing fluctuations in poor stability of hsync or vsync lines (8e12e187)
- Fixed support for synchronizing to Hercules Stingray Pro (ARK1000PV) SVGA Mode 43h 1280x1024i. (fa2b2b99)
- Fixed support for synchronizing to Hercules Stingray Pro (ARK1000PV) SVGA Mode 32h 1024x768i. (25e6a0ec)
- Fixed video synchronization to not get confused by alternating line counts on even and odd fields of interlaced video modes. (387334ed)
- Improved stability for synchronizing to Tseng ET4000AX SVGA Mode 3Dh 1280x1024i. (bdb1c722)
- Added support for synchronizing to Hercules Stingray Pro (ARK1000PV) VESA Mode 1024x768i 15/16bpp. (95eb1411, a61359e2)
- Improved # of colors indication in Developr HUD to show suffixes 'b' (bpp in hi/truecolor), 'v' (vga), 'e' (ega), 'c' (cga), 'm' (mono) for color count. (7468a88c)
- Fixed an issue reporting color count data from CRT Terminator to SNOOP.EXE. (7468a88c)
- Improved support for synchronizing to interlaced video modes on WD90C31-LR. (1188060e)
- Improved support for synchronizing to WD90C31-LR VESA 320x240 15bpp, 512x384 15bpp, 640x400 16bpp, 640x480 15bpp, 640x480 24bpp and 800x600 8bpp video modes. (3424d34e)
- Improved support for synchronizing to video modes with vertical refresh rate below 50Hz. In such case, if viewing in upscaled video mode (and not Passthrough mode), CRT Terminator will repeat frames to bring the refresh rate up to 50Hz minimum. (cc9c8f99)
- Improved support for synchronizing to Hercules Stingray Pro (ARK1000PV) VESA 1280x1024p 256c and 800x600 15bpp video modes. (541670e2)
- Improved support for synchronizing to Avance Logic ALG2302 SVGA Mode 45h 1024x768i 16bpp (2cdc668d)
- Improved support for synchronizing to Hercules Stingray Pro (ARK1000PV) 640x480 256c, 800x600 256c, 800x600 15bpp and 1280x1024i video modes. (a61359e2)
- Fixed a video synchronization issue when VGA borders were disabled from DOS I/O interface and current video mode was interlaced. (c759ea3f)
- Improved support for synchronizing to Hercules Stingray Pro (ARK1000PV) 640x480 24bpp video mode. (69783b0a)
- Added support in the DOS I/O interface to disable CRT Terminator video scanhalving in sub-register 40h bit 20h (fde6b447)
- Improved support for synchronizing to Hercules Stingray Pro (ARK1000PV) nonstandard 800x600 72Hz, 1024x768p 69Hz and 1024x768 74.6Hz video modes. (97337574)
- Optimized video RAM bandwidth usage to increase interlaced <= 8bpp video modes to output 60Hz progressivized refresh rate instead of 50Hz. (b20db3b3)
- Improved support for synchronizing to Hercules Stingray Pro (ARK1000PV) UniVBE 512x384 24bpp 640x480 24bpp video modes. (0038d0fd)
- Optimized video RAM bandwidth usage to enable synchronizing to Hercules Stingray Pro (ARK1000PV) UniVBE 1024x768 15/16bpp 43.7Hz video modes. (d33aa9eb, 8231ed56)
- Improved support for synchronizing to Hercules Stingray Pro (ARK1000PV) UniVBE 320x240 256c video mode. (73684707)
- Fixed video upscaler problems when operating in packed 16bpp video upscaling mode. (6c407807)
- Added support for precise vertical cropping of VGA border in interlaced modes that have a different number of border scanlines to crop from top and bottom edges in even and odd fields. (5f6904c1)
- Improved 15bpp vs 16bpp video signal detector to not get confused as easily between detecting the two. (7f67735a)
- Fixed support for synchronizing to Diamond SpeedStar 64 1280x1024i 256c video mode. (9aeb67c8)
- Improved support for synchronizing to various video modes on Avance Logic ALG2302. (b4e5f080, 59a30c1e, aad82b99)
- Improved support for synchronizing to various video modes on Tseng Labs Tseng ET4000AX. (9f7b2690, 8393a430)
- Improved support for synchronizing to 1280x1024i video mode on Diamond SpeedSTAR 24 Tseng ET4000AX. (dfc42fc6)
- Fixed border cropping on Paradise PVGA1A-JK 64x34 and 80x25 text modes. (81534c45)
- Fixed issues with synchronizing to Trident 8900D video modes. (b3efc3ce, 2d934299, 2b2374f8)
- Fixed an issue in Developer HUD printing the minus sign for negative horizontal back porch lengths. (141f71af)
- Improved video crop of Jazz Jackrabbit video mode to preserve more vertical pixels for better aspect ratio (circles more circular). (09c7cca6)
- Improved support for oddball Trident 8900C SVGA video modes like 512x480. (ab1e3510)
- Improved video cropping on Cirrus Logic CL-GD5428 text modes. (f403e666)
- Improved video mode synchronization on Trident 9000i and Jazz Multimedia Tseng ET6000. (86f1ecd2, 83085bfe)
- Improved support for S3 Powergraph X-24 (S3 86C801) video modes. (65312dd4, 5abfd300, 5eb8e3fc)
- Improved support for S3 ViRGE ST-325A in VESA and UniVBE operation. (068ca649, 09dd83ac)
- Improved support for S3 Vision864 (S3 86C864-P) video modes. (069115f5)
- Improved support for Realtek RTG3106 video modes. (93d706ab)
- Improved support for Wang Labs Chips&Techologies F82C452 video modes. (1a750c72)
- Improved support for Headland HT208 video modes. (b7767419)
- Improved support for Acumos AVGA2 and Matrox Millennium 2064W. (6c9b5d9b)
- Improved support for Acumos AVGA1 and ATI 28800-5. (882e8541)
- Improved support for ATI 28800-6. (464a30ff, ddd0dc1d)
- Fixed a video mode synchronization problem seen with ATI 28800-6, where transitioning from one high pixel clock video mode to another might cause the HDMI PLL lock to not be properly reinitialized. (84dd7e3a)
- Improved video mode support for Ahead V5000B and MiroVideo 12PD v2.00. (141d9a39, 5255fa4a)
- Added support for synchronizing to MiroVideo 12PD v2.00 VESA mode 1024x768i 16bpp 91.894 Hz (8d68f6c1)
- Improved support for Aska ZyMOS Poach 51 (Trident 8800CS) video modes. (39713cf6, b516301a)
- Improved support for Trident TVGA8816CSC2 video modes. (267d69ff)
- Fixed synchronization to Jaton Corporation CL-GD5446 SVGA Mode 6Ch 1280x1024i video mode. (f8d4f15a)
- Added special handling to Tseng ET4000/W32p big endian clocked hi color video modes. (3ee5458e)
- Improved the Developer HUD to show the built firmware version in the HUD directly. (3b73e590, a88d90e6, 33d85453, 17b6732f)
- Fixed an issue in scandoubling detection in interlaced video modes. (987bc258)
- Optimized video signal stability on high >75 MHz Feature Connector pixel clocks (6d63e44a)
- Disabled support for radical overclocked HDMI PLL values that will never be able to work. (098a7cb3)
- Fixed video mode synchronization to Hercules Stingray Pro (ARK1000PV) 800x600 256c 90Hz video mode. (21aed934)
- Optimized Feature Connector signal stability for >90MHz pixel clocks. (c41d4b2a, 6655c611)
- Improved Developer HUD to show with uppercase 'B' if current hi-color video mode is DDR signaled by the video card. (e09ec8c9)
- Lifted a color bit depth quantization limitation that occurred with DDR signaled video modes. (a512fa62)
- Improved color depth on DDR video modes on Trident TGUI9440 and WDC WD90C31-LR adapters. (a512fa62)
- Improved video upscaling stability when DOS VGA adapter is set to very high >100Hz refresh rates. (0e0b1e80)
- Fixed a pedantic quirk on handling VGA Palette Access Mask register (bc358d91)
- Optimized video signal stability when outputting 1920x1080 and 1920x1200 video modes. (c41d4b2a, 6776796a, 18a34a7e, 63933192)
- Optimized timing closure with a more efficient parallelized hill climbing timing optimizer. (aacfb10b)

The next batch of adapters will ship with all these fixes, and I'm setting up a way for existing owners to flash the updated firmware as well.

Of course, "if it ain't broke, don't fix it" applies, though we will have also good support for downgrading back to the first shipped firmware version so that migrating between the two will be easy for any possible comparison that may be needed, or if people will want to just play around.

Nice progress! Are you able to ship to everyone in queue or will there be a third batch later?

Is the process to flash the firmware going to involve a flashing tool?

janih wrote on 2025-01-08, 21:33:

Nice progress! Are you able to ship to everyone in queue or will there be a third batch later?

Thanks! The second batch will ship to everyone currently in the queue. There are still two cards left from that batch for anyone interested.

We haven't yet decided on a third batch.

keenmaster486 wrote on 2025-01-09, 01:10:

Is the process to flash the firmware going to involve a flashing tool?

Yes, that's right. The process will be to connect CRT Terminator to a modern PC with the USB-C port on the card, and then use a provided installer package to select and flash the firmware. The flashing tool will be Windows only at least for now.

The next set of CRT Terminator boards is now processing shipping. They will come with the new v2025.1.5 firmware flashed in them. Thanks to everyone who have had the patience to wait for this second shipping batch to come alive.

Still working on getting the firmware update documentation and tooling uploaded on the web site. I'll ping as soon as that is available.

Most CRT Terminators have been shipped. Heads up to everyone who ordered and have not yet received a tracking number: if you were one of the earlier people to order, my early order form had missed a phone number field. There are still a couple of you that I have tried to reach to ask for a number to attach to UPS delivery records. Please double check your email asap if I've sent you an email to inquire.

So if you have already paid for a board but haven't yet received an UPS tracking number (sent by PayPal notification), please check your emails.

Hello,
I received my CRT Terminator yesterday. Thank you very much. It works and provides a nice picture. Unfortunately the colors are wrong.
I have tested several ISA graphics cards and DVI monitors. On 286, 386 and 486 ISA systems. I have also tested different resolutions. On the 486 some colors are correct. But not all.
According to the manual, color problems may only occur with PCI or AGP, right?

Hi Nelson68k,

that is quite unexpected. What seems to be happening here is that CRT Terminator is not correctly observing the VGA adapter palette writes on the ISA bus, and the colors that it displays look like they are left in the default boot-up state, i.e. the default EGA palette. In some games, this palette may coincide with "correct" colors partially or in full, depending on which colors the software ends up overwriting.

What makes this situation unusual and surprising is that you are seeing the issue with ISA graphics cards, and on multiple different PC systems: on all motherboards that we have been aware of, if the VGA adapter is on the ISA bus, then CRT Terminator should always observe the palette updates as they occur.

If the VGA adapter is on the PCI/AGP bus, then by default the palette writes are not observed, and a separate configuration to enable palette snoop will be needed.

Here are some troubleshooting steps to take:
0. Download the software from https://github.com/juj/crt_terminator/tree/main/DOS/bin
1. Run PALANIM.EXE (from PALANIM.ZIP). If palette writes are observed properly, then the test should display vertical stripes where the colors cycle upwards in a scrolling animation. If the colors stay fixed and don't move, then palette writes are not observed properly.
2. Run SNOOP.EXE (from SNOOP.ZIP). If palette writes are observed incorrectly, SNOOP should be reporting something along the lines of "CRT Terminator is currently not observing RAMDAC palette writes." Are there any other errors/warnings that SNOOP.EXE might report?
3. Run "CRTT.EXE --snoop=on" (from CRTT.ZIP). Then re-run PALANIM.EXE and SNOOP.EXE (or SCROLL.EXE from SCROLL.ZIP, or any other VGA game, like Monkey Island) to see if that has an effect.
4. If the above does not help, then try running PALTSR.EXE (from PALTSR.ZIP), and then re-run PALANIM.EXE/SNOOP.EXE/SCROLL.EXE or any VGA game.

Do you have specific model information on the tested motherboards and VGA adapters?

Lastly, any chance there might be an overclock on the ISA bus clock speed? If so, try reducing the ISA bus to 10 MHz at maximum. (that is the speed at which I have run the CRT Terminator tests at)

Hi, thanks for your reply. I have tried the card in my 286 replica. everything works fine here. i am very positively impressed. here are a few screenshots.

Thanks for the hints. I will try to find the problems with the other boards.

Fantastic. Running SNOOP.EXE contains some diagnostics items that might give a clue, and it is also interesting to check how PALTSR.EXE might affect the situation.

One thing to clarify that I missed in previous comment is that ISA VLB graphics cards fall into the same category as PCI and AGP graphics cards: they'll need some solution for enabling palette snoop (e.g. a motherboard BIOS setting to enable palette snooping), or the PALTSR.EXE program can be used.

On PCI and AGP adapters specifically, "CRTT.EXE --snoop=on" is the recommended way to enable snooping. (that method won't work for ISA VLB cards)

The new firmware v2025.1.5 is now live!

Check out the new section Updating the Firmware in the manual for the installation flow and downloads.

In particular, I hope that the new firmware might be able to help with the issue that Keenmaster486 identified in this comment: Re: CRT Terminator Digital VGA Feature Card ISA DV1000 and allow you to use the automatic signal sampling phase setting again, instead of having to resort to the manual sampling phase settings.

Nice. Will there be a process to update the firmware on Linux?

keenmaster486 wrote on 2025-01-28, 22:53:

Nice. Will there be a process to update the firmware on Linux?

I would imagine it's more-or-less the same process; get libusb and openfpgaloader, use the latter to send the firmware (exact command lines are in the update script). Maybe it doesn't even need libusb on linux, if there's a default driver for FTDI.

I have not had plans on packaging a Linux installer, largely due to the large variance of different Linux distributions that is complicated to sort out and test.

The FPGA binary is indeed programmed with openfpgaloader like jmarsh mentions. So if using a Linux distro that has openfpgaloader apt-gettable/rpmble/pacmanable (I believe libusb gets/should get installed as dependency to openfpgaloader), then the installation amounts to a single invocation of openfpgaloader:

1openfpgaloader --cable ft232 --write-flash DV1000-2025.1.5.fs

My understanding is that beyond libusb and openfpgaloader, Linux world does not need a specific driver for the FTDI JTAG programmer present, like Windows does with the WinUSB business.

Let me know if there's challenges with getting the firmware updated, and which Linux distro you are on. Maybe I can get a test PC box and install that distro to see how the installer path would work out there.

I just thought I'd share my limited experience with testing for a few minutes today.

I've got an AST Bravo LC/33 (486 DX33) machine where there is a Cirrus Logic CL-GD5422 embedded directly on the motherboard with exposed feature connector header.

I was also getting palette wonkiness, and so downloaded the utilities and started testing. Attached is the result from running snoop. When attempting to run palanim, the screen flashes constantly on the Terminator output, but looks perfectly smooth from the VGA out.

This card was on the tested / working card list, but I'm not sure how AST may have mucked it up in their implementation.

I'm now considering acquiring some more ISA VGA cards just to see what other results I may be able to get out of this machine.

CircuitRewind wrote on 2025-02-01, 10:09:

When attempting to run palanim, the screen flashes constantly on the Terminator output, but looks perfectly smooth from the VGA out.

That sounds unexpected/something that I feel like I haven't seen before. Would you be able to record a video of what that looks like?