Flashy wrote on 2022-07-26, 18:55:

I have a Speedstar 24X REV A5 with the WDC chip and no BIOS ICs (I have a programmer and EPROMs). Can somebody dump me the BIOSes for this card? I could not find it anywhere yet. Google photo search shows cards with BIOS version 1.04, so it may be a complete different branch than those I found here. I tried one, no success.
Thanks!

I have this card:

I will dump the bioses tomorrow - it's the same data on both chips IIRC

also here is version 2.02 posted: Re: Looking for Diamond SpeedStar 24X WDC Bios files

keropi wrote on 2022-07-26, 19:02:

I have this card: […]
Show full quote

Flashy wrote on 2022-07-26, 18:55:

I have a Speedstar 24X REV A5 with the WDC chip and no BIOS ICs (I have a programmer and EPROMs). Can somebody dump me the BIOSes for this card? I could not find it anywhere yet. Google photo search shows cards with BIOS version 1.04, so it may be a complete different branch than those I found here. I tried one, no success.
Thanks!

I have this card:

I will dump the bioses tomorrow - it's the same data on both chips IIRC

also here is version 2.02 posted: Re: Looking for Diamond SpeedStar 24X WDC Bios files

My card looks the same, only without the original EEPROMs. Now I have success with 2.02 with just 1 BIOS IC in 8-bit mode (it did not start with an EEPROM in the other socket and that was all I had now), thanks for the link! I will save your version also for archiving and maybe testing for differences later.
This forum is such a great community!

Flashy wrote on 2022-07-26, 20:16:

My card looks the same, only without the original EEPROMs. Now I have success with 2.02 with just 1 BIOS IC in 8-bit mode (it did not start with an EEPROM in the other socket and that was all I had now), thanks for the link! I will save your version also for archiving and maybe testing for differences later.
This forum is such a great community!

sorry for the delay, I forgot about it
I am attaching the 1.04 dumps , both chips hold the same data
if you ever compare it to 2.02 please post your findings 😀

I have made a few tests.
Motherboard: GA-586ATS, CPU: P1 133MHz, RAM: 16M EDO; Video card: Diamond Speedstar 24X with WDC chipset and 1Mbyte video memory, 0WS jumper installed, 16 bit BIOS jumper installed.
Test programs: the DOSBENCH collection and VIDSPEED.
Long story short: I did not find any benchmark that gives different results with 1.04 and 2.02 video BIOS... (the BIOS files differ and HWINFO correctly detects the versions).
Removing the 0WS jumper makes everything slower by at least 10% (it depends on what the program is doing, up to about 25%), VIDSPEED shows that only writes are slower.

Flashy wrote on 2022-08-12, 10:56:

I have made a few tests. Motherboard: GA-586ATS, CPU: P1 133MHz, RAM: 16M EDO; Video card: Diamond Speedstar 24X with WDC chipse […]
Show full quote

I have made a few tests.
Motherboard: GA-586ATS, CPU: P1 133MHz, RAM: 16M EDO; Video card: Diamond Speedstar 24X with WDC chipset and 1Mbyte video memory, 0WS jumper installed, 16 bit BIOS jumper installed.
Test programs: the DOSBENCH collection and VIDSPEED.
Long story short: I did not find any benchmark that gives different results with 1.04 and 2.02 video BIOS... (the BIOS files differ and HWINFO correctly detects the versions).
Removing the 0WS jumper makes everything slower by at least 10% (it depends on what the program is doing, up to about 25%), VIDSPEED shows that only writes are slower.

thanks for the testing!
it would be interesting to find though if the 2.02 bios added some more VBE screenmodes - I think this will be the real difference
maybe scitech display doctor can be used, it reports before and after number of screenmodes when the tsr loads

Interestingly still no difference between the two BIOS versions.
Scitech Display Doctor (6.53 DOS version) reports 7 modes before and 7 after, but after running the Speedstar24X's own VESA TSR I get 3 new 15 bit modes (without the tsr 8 bit modes are the highest available). These new modes work well with the picture viewer "Sea", but SDD's test program flickers in double buffering test (where there should be no flicker according to the text on the screen) in all resolutions (not just the 3 new ones).

keropi wrote on 2022-08-12, 11:08:

thanks for the testing! it would be interesting to find though if the 2.02 bios added some more VBE screenmodes - I think this w […]
Show full quote

Flashy wrote on 2022-08-12, 10:56:

I have made a few tests. Motherboard: GA-586ATS, CPU: P1 133MHz, RAM: 16M EDO; Video card: Diamond Speedstar 24X with WDC chipse […]
Show full quote

I have made a few tests.
Motherboard: GA-586ATS, CPU: P1 133MHz, RAM: 16M EDO; Video card: Diamond Speedstar 24X with WDC chipset and 1Mbyte video memory, 0WS jumper installed, 16 bit BIOS jumper installed.
Test programs: the DOSBENCH collection and VIDSPEED.
Long story short: I did not find any benchmark that gives different results with 1.04 and 2.02 video BIOS... (the BIOS files differ and HWINFO correctly detects the versions).
Removing the 0WS jumper makes everything slower by at least 10% (it depends on what the program is doing, up to about 25%), VIDSPEED shows that only writes are slower.

thanks for the testing!
it would be interesting to find though if the 2.02 bios added some more VBE screenmodes - I think this will be the real difference
maybe scitech display doctor can be used, it reports before and after number of screenmodes when the tsr loads

Flashy wrote on 2022-08-13, 15:35:

Interestingly still no difference between the two BIOS versions.
Scitech Display Doctor (6.53 DOS version) reports 7 modes before and 7 after, but after running the Speedstar24X's own VESA TSR I get 3 new 15 bit modes (without the tsr 8 bit modes are the highest available). These new modes work well with the picture viewer "Sea", but SDD's test program flickers in double buffering test (where there should be no flicker according to the text on the screen) in all resolutions (not just the 3 new ones).

thanks for testing this too!
I wonder what the point is for the 2.xx bios version ... anyways this means I won't have to bother upgrading my cards 😀

I have discovered this thread unfortunately only now. A few years ago I tried to upgrade a Speedstar VGA to a Speedstar 24 and give it a Vesa-compatible bios and tried a lot.
Unfortunately it didn't work. With the Bios (6.01) and DAC (SS24) of a Speedstar 24 I got a picture but also a freeze during POST. I then discovered it depends not only on the DAC but also on the revision of the ET4000AX if higher colors are possible. Older ET4000AX chips cannot do higher color depths no matter what DAC they are combined with. So a newer chip version and probably sharper timings should be a difference between Speedstar VGA and Speedstar 24. That Tseng has diligently developed the chip further can be seen by the fact that the "Copper" demo only runs perfectly with a certain ET4000AX version. Here is the thread about it: Copper Demo
So probably adjustments to changes Tseng has made to their chips are also a reason for the different bios versions.

I found a real treasure on the net at the time: I found a complete copy of Diamond's (German) support website from February 1997. There is an interesting file in which it says which Bios versions are current for which card and the ISA Speedstar cards are also listed. Also drivers and software can be downloaded there, as if it was 1997.

http://edv-janssen.synology.me/diamond/

And here the direct link to the file with the bios versions:
http://edv-janssen.synology.me/diamond/driver … ile/bios_di.xls

By the way, in many of my ET4000AX cards the bios of a Targa Win 32K works perfectly. It is a late bios for the Tseng cards and supports VEB 1.2 iand many special text modes directly from the ROM. With the Speedstar unfortunately not, because the ROMs are somehow encrypted, as has already been written here.

DragonSphere wrote on 2022-08-22, 12:43:

I have discovered this thread unfortunately only now. A few years ago I tried to upgrade a Speedstar VGA to a Speedstar 24 and g […]
Show full quote

I have discovered this thread unfortunately only now. A few years ago I tried to upgrade a Speedstar VGA to a Speedstar 24 and give it a Vesa-compatible bios and tried a lot.
Unfortunately it didn't work. With the Bios (6.01) and DAC (SS24) of a Speedstar 24 I got a picture but also a freeze during POST. I then discovered it depends not only on the DAC but also on the revision of the ET4000AX if higher colors are possible. Older ET4000AX chips cannot do higher color depths no matter what DAC they are combined with. So a newer chip version and probably sharper timings should be a difference between Speedstar VGA and Speedstar 24. That Tseng has diligently developed the chip further can be seen by the fact that the "Copper" demo only runs perfectly with a certain ET4000AX version. Here is the thread about it: Copper Demo
So probably adjustments to changes Tseng has made to their chips are also a reason for the different bios versions.

I found a real treasure on the net at the time: I found a complete copy of Diamond's (German) support website from February 1997. There is an interesting file in which it says which Bios versions are current for which card and the ISA Speedstar cards are also listed. Also drivers and software can be downloaded there, as if it was 1997.

http://edv-janssen.synology.me/diamond/

And here the direct link to the file with the bios versions:
http://edv-janssen.synology.me/diamond/driver … ile/bios_di.xls

By the way, in many of my ET4000AX cards the bios of a Targa Win 32K works perfectly. It is a late bios for the Tseng cards and supports VEB 1.2 iand many special text modes directly from the ROM. With the Speedstar unfortunately not, because the ROMs are somehow encrypted, as has already been written here.

The TC6100AF marked chips will do high color. The TC6058AF will not

Don't mean to raise this from the dead too much, but I have this specific ET4000AX Speedstar card. Acquired it a few weeks ago at the recent VCFed swapmeet in NJ, sitting in a 386DX build that was pretty worse for wear, however this was the crown jewel that made me go for it. There doesn't seem to be much in the way of documentation for this specific kind of card, so I'm dropping it in here as this is where the party is (was!) at for ROM dumps of these very nifty cards.

it has BIOS 4.23 on it and the Music TR9C1710 DAC. Based on the latest datecodes on the board I can only assume this card was from early, early 1992. from what little I can gather about this particular card, this is the precursor to the Speedstar24, but if you waited a bit, you could upgrade it to a Speedstar24 with an upgraded ROM set and DAC... at least, that's what I've read from one of the sources I was able to dig up. I'm not sure what the extra mask ROM there is for, however. character set? I can't figure out which mask ROM type it really is to get my T48 to read it. I did dump the ROMs for the card otherwise, although I don't know if it matters much... I think this is a common ROM set.

Given I have recently bought a UV-C EPROM eraser, I'm going to try and upgrade the BIOS on this card to one of the later 5.x/6.x ROMs and see how that goes. Mine has an issue where higher resolutions than 640x480 are rather noisy, so I want to see if that helps any before I bite the bullet and throw money at the problem for another DAC. Mind you the highest res it'll likely ever see is going to possibly be 800x600@8bit under WFW311. 640x480@8bit is pretty much perfectly fine.
I also have a Weitek 8256 card here with a Sierra SC11483CV on it I could always yank off and drop in the socket, seeing as that seems to be another compatible DAC, albeit only a bit better than the one I'm working with now.

Currently this card is sitting in my PS/2 40SX, and was a much easier upgrade than yanking the motherboard and installing the extra ZIP-40 RAM for the onboard video... but I did need to upgrade the machine's BIOS before it would properly enumerate at boot. Doing this on the earlier March/April 1991 BIOS compared to the September 1991 BIOS is futile.

Sudos wrote on 2023-11-09, 13:31:

I'm not sure what the extra mask ROM there is for, however. character set?

X-ROM a.k.a. Translation ROM
Allows to use register-level CGA and Hercules modes with regular VGA monitors, ie. 31 kHz HSYNC.

Sudos wrote on 2023-11-09, 13:31:

I have this specific ET4000AX Speedstar card.

Wow! I also have this same card and I was also looking for best bios for it. I'd be very interested in knowing how it went for you. This card differs from the later speedstar24 by that it doesn't have the eeprom to store monitors refresh rates across reboots, so you have to reload them from autoecec.bat every time with a special utility from Diamond.

Grzyb wrote on 2023-11-09, 15:34:

X-ROM a.k.a. Translation ROM
Allows to use register-level CGA and Hercules modes with regular VGA monitors, ie. 31 kHz HSYNC.

Ah okay, that makes total sense then.

Zerthimon wrote on 2023-11-09, 20:12:

Wow! I also have this same card and I was also looking for best bios for it. I'd be very interested in knowing how it went for you. This card differs from the later speedstar24 by that it doesn't have the eeprom to store monitors refresh rates across reboots, so you have to reload them from autoecec.bat every time with a special utility from Diamond.

I did try the 5.0 and 6.01 ROMs earlier today, and with each I was getting random text corruption, each boot on random sets of characters... so either it's an incompatibility with the Music ramdac itself or some other issue that's gonna keep me on v4.23 until I get myself a better one... or try that other one I have, whenever that happens.

I wasn't aware there was proper software for this specific card. All the software I can find is for the Speedstar24. Mind posting where I can find this? having that plus whatever Diamond-sanctioned drivers for Windows might increase my chances of getting things happy further down the line.

Should also make mention that this is the card I have that I'd need to hot air the RAMDAC off of. Mine is in worse-for-wear shape as the zip-4o DRAM on it was absolutely trashed , but I did get it to turn on and honestly it's nothing to write home about... no faster than an ATI VGA Wonder-16 and probably a lot more stunted than such in that department by the looks of it. Those Weitek chips really shined later on as the base chip in the newer accelerator cards.

Sudos wrote on 2023-11-09, 20:26:

I wasn't aware there was proper software for this specific card. All the software I can find is for the Speedstar24. Mind posting where I can find this? having that plus whatever Diamond-sanctioned drivers for Windows might increase my chances of getting things happy further down the line.

This is what I'm using:

Zerthimon wrote on 2023-11-09, 21:05:

This is what I'm using:

Ah! That makes all the sense then that it'd be in there. I was under the impression that there was an older set of drivers for this card. For all I know, there still is.

I'm still puzzled about the text corruption. no amount of cleaning any of the pins changed the behavior, it was all down to the BIOS in the end... but I did have the card running in full-on turbo mode. maybe that's it? but I doubt it.

Sudos wrote on 2023-11-09, 20:26:

Should also make mention that this is the card I have that I'd need to hot air the RAMDAC off of. Mine is in worse-for-wear shape as the zip-4o DRAM on it was absolutely trashed , but I did get it to turn on and honestly it's nothing to write home about... no faster than an ATI VGA Wonder-16 and probably a lot more stunted than such in that department by the looks of it. Those Weitek chips really shined later on as the base chip in the newer accelerator cards.

Do you have the drivers for this card? All three existing Windows 3.1 drivers for the Weitek 5x86 that can be found on the net (two on Weitek's site and the third on Metropoli) seem to only support the 5086 and 5186 with 256-colour modes while having garbage output with the 5286, but 15-bit output works fine. It doesn't matter if I use DIP-20 DRAM, ZIP-40 DRAM or ZIP-28 VRAM on the card, it still produces garbage output.

DEAT wrote on 2023-11-13, 23:34:

Do you have the drivers for this card? All three existing Windows 3.1 drivers for the Weitek 5x86 that can be found on the net (two on Weitek's site and the third on Metropoli) seem to only support the 5086 and 5186 with 256-colour modes while having garbage output with the 5286, but 15-bit output works fine. It doesn't matter if I use DIP-20 DRAM, ZIP-40 DRAM or ZIP-28 VRAM on the card, it still produces garbage output.

Sadly I do not. given I wasn't able to find any drivers for it either, and the fact it's used in more robust cards such as one of the power 9000 cards as a DOS compatibility chip, I don't think you're going to find them anytime soon. I gave up and harvested the zip-4o DRAM off mine and the EPROM for other uses. However, given mine is 2.0 and there's also a GAL on it, I dumped both of those beforehand. They're attached if needed.

I got the RAMDAC off that weitek card, cleaned up the pins and dropped it in. no BIOS change, just dropped it in. now I can do 32768 colors! I've found the Samsung DAC runs a bit warmer than the Music DAC... probably because it can do more at likely the same node process. Might adhere a heatsink to it... but the important thing here is, I have as close to a Speedstar24 as I'm going to get without messing around with more ROM stuff. WFW311 with some nice wallpapers never looked so pretty.
The only problem I see is the picture can be a bit on the shaky side, very hard to spot from a distance but you can definitely tell once you're up close. It was like this prior as well but not as bad with the Music DAC so I'm betting it has to do with the tantalum capacitors on the card and the fact that it's coming up on 32 years old. The problem with replacing capacitors on this card is going to be space constraints, but I've done crazier stuff before. I'm no stranger to hiding capacitors on the backside and hot-snotting them down if it means a cleaner picture at the end of the day, and I'm one of those people that will demand perfection when something silly like this comes up.

I'm still more than happy regardless, and while I wait to do my end-of-year cap order, I'm gonna push this RAMDAC and make sure it's a better fit than the old one in terms of looking for snow artifacts in DOS games and such. I didn't see any when I loaded up Pinball Fantasies but that isn't much of a test. I have a couple others to go through before I'll be convinced it's fine.

Sudos, I have the same card as you and I find what you do with it very exciting.

I just wanted to check Diamond's information about the bios versions again and found that unfortunately the website with the backup of the Diamond homepage from 1997 that I once posted in this thread is no longer online.
Here is the information on the Speedstar again, which I luckily saved:

1boardname	                           revision	   current BIOS	                    
2SpeedStar 24 ISA                       C1          6.01
3SpeedStar VGA / Plus / Hi-Color ISA    C1, C6      3.10 or 4.13 (long card)
4SpeedStar VGA / Plus / Hi-Color ISA    C1, C6      3.24 or 4.23 (short card)
5SpeedStar VGA / Plus / Hi-Color ISA    C1, C6      6.01 (only Hi-Color)

This was the info for Diamond support staff in 1997 on the most current BIOS versions and hardware revisions.
Here is the complete file for all diamond cards

So version 6.01 is the latest BIOS version for cards with 24-Bit DAC.
That 3.x or 4.x versions can be the most current versions of the other (VGA / Plus) cards can actually only be related to the DAC or chip used. What do you think?