Hello !

I have a Expertcolor DSV5357 that suffers from this problem. Is it possible to patch its bios ? (there it is

) I'd like to know how you patch these bioses as well. Maybe we could make a automatic patching utility for that ?

I'm sure I have other cards with that problem but I haven't tested them.

I was actually just wondering about this as well the other day, as I purchased another GX2 in the hopes that it would be correct, which it wasn't. While the small tools provided work fine under DOS, you can't use them in Windows. If there was a reliable way to patch each VBIOS (from what I gather, a checksum would need to be recalculated as well?) then it would be at the correct levels at boot and you wouldn't even have to worry about it. I don't know anything about doing this kind of thing, but maybe someone else can find a way? This is probably a tall order but maybe it's possible.

I bought two of S3 Trio 3D/2Xs recently and they both exhibit the same issue. In text mode only, blacks are washed out and there are vertical bars when viewed on an LCD. The vertical bars do not appear on my CRT but the blacks are still washed out.

I saw a video by Phil where he talked about this issue on the S3 Trop64V+ but I didn't realize that the problem extended to other models as well. I've included a photo of some markings on the back from the manufacturer which suggests that both of these cards were manufactured around the same time or are simply an identical revision.

Here is a link to Phil's video:

https://youtu.be/N_5suEoXTc8

I can confirm that the following workaround successfully corrects the issue temporarily until the next reboot:

S3 AGP Cards (and possibly others) Too Bright

If I want this to be a permanent fix then I would need to replace the BIOS chip. What would be the cheapest chip programmer that would be able to do this? I don't have one.

The BIOS chip on these beneath the sticker are a A276308-70 which I think are only one-time programmable (as indicated by OTP on the datasheet). What kind of chip would be a suitable replacement?

Hi Kahenraz, have you found a bios and load it to shadow or you did the 3c5 register trick to fix your luminosity level ?

I have some w27c512-45 that *might* be compatible, if you have the bios and dont live too far maybe I can just snail mail you a programmed eprom. I found a s3 trio 3d 86c365 on my spair and the bios is an an atmel at27c512r. Otherwise to answer your question I use xgecu tl866ii plus and it seem to support your amic chip (http://www.autoelectric.cn/MiniPro/TL866II_List.txt). My only regret about that thing is that I did not bought it earlier.

I used the debug 3c5 register trick. Is there a way to load one of the BIOSes in this thread into shadow ROM to verify compatibility?

Just use s3black.com from this post Re: S3 AGP Cards (and possibly others) Too Bright in autoexec.bat.
This is all you need 😀

Kahenraz wrote on 2021-08-20, 06:00:

I used the debug 3c5 register trick. Is there a way to load one of the BIOSes in this thread into shadow ROM to verify compatibility?

Right, just look at page 4 of this thread, user wbc have a file sbvl.zip attached. You can use it to load temporarily a bios into your shadow ram (make sure it's enabled in your mainboard bios, it's often called video shadow enabled/disabled), hopefully one if the .cps (triton, sis...) in the archive will be compatible with your chipset. Once you find a bios that work for you, you can burn it on an eprom as a permanent solution and move on with your life 😉

I have the S3 Trio3D/2X both in PCI and AGP versions and they're great for DOS. I'm currently using it on my Socket 7 machine.

Using S3BLACK.COM in AUTOEXEC.BAT as zyga64 said works as intended, however there is another way using a utility called CONFIX which allows to use the in and out instructions in CONFIG.SYS and this is the one that I use since it does the workaround prior to executing AUTOEXEC.BAT.

1DEVICE=C:\CONFIX\CONFIX.SYS OUT 3C4 08
2DEVICE=C:\CONFIX\CONFIX.SYS OUT 3C5 06
3DEVICE=C:\CONFIX\CONFIX.SYS OUT 3C4 27
4DEVICE=C:\CONFIX\CONFIX.SYS OUT 3C5 00
5DEVICE=C:\CONFIX\CONFIX.SYS OUT 3C4 08
6DEVICE=C:\CONFIX\CONFIX.SYS OUT 3C5 00

I have an thing for this card which allows it to use UniVBE. This is because while it has an amazing VBE 2.0 implementation it has some bugs that are not present on the previous S3 cards when using VESA video modes on certain programs.

Best would be if anyone can use the right value on the bios, I did some test but I was unable to locate the exact location :\
In that way it would be possible to update the bios without using any additional dos command.

Nemo1985 wrote on 2021-08-20, 16:50:

Best would be if anyone can use the right value on the bios, I did some test but I was unable to locate the exact location :\
In that way it would be possible to update the bios without using any additional dos command.

It depends of the S3 model as some has the BLANK pedastal located on a different register.

S3 ViRGE based cards (except GX2 versions) seems to have this located at SR1A bit 5 (sequence register, from the S3 ViRGE/VX programming datasheet). Later cards have this parameter located at SR27 bit 3.

Edit: Didn't notice that wbc has posted a similar information prior in this same thread.

The DEBUG command for the setting off SR1A bit 5 is this:

1o 3c4 08
2o 3c5 06
3o 3c4 1a
4i 3c5
5<subtract 20h from given value>
6o 3c5 <resulting_value>
7q

Gmlb256 wrote on 2021-08-20, 17:04:

It depends of the S3 model as some has the BLANK pedastal located on a different register.
S3 ViRGE based cards (except GX2 versions) seems to have this located at SR1A bit 5 (sequence register, from the S3 ViRGE/VX programming datasheet). Later cards have this parameter located at SR27 bit 3.

OMG that would be so awesome if it work ! I was not successful at finding those values... We could fix pre vesa 2.0 bios that still have the font maxis game used !! And we could fix latest s3 virge bios and dont have to use random version that dont have the issue !!! I'll check it out later on and get back to you about it. Thanks

Gmlb256 wrote on 2021-08-20, 17:04:

Nemo1985 wrote on 2021-08-20, 16:50:

Best would be if anyone can use the right value on the bios, I did some test but I was unable to locate the exact location :\
In that way it would be possible to update the bios without using any additional dos command.

It depends of the S3 model as some has the BLANK pedastal located on a different register.

S3 ViRGE based cards (except GX2 versions) seems to have this located at SR1A bit 5 (sequence register, from the S3 ViRGE/VX programming datasheet). Later cards have this parameter located at SR27 bit 3.

That's a good catch, despite the different location of supported cards, modify a s3 bios is quite straight forward, there is a sumcheck that's quite easy to calculate.
I remember I was able to modify the frequencies without issues a year back or so.

From a mention of this thread I curiously read here.

Kahenraz wrote on 2021-08-20, 02:47:

In text mode only, blacks are washed out and there are vertical bars when viewed on an LCD. The vertical bars do not appear on my CRT but the blacks are still washed out.

I find it just awesome that nobody seems to look at the possibility that this phenomenon could be hardware related.

As I am very flicker-sensitive, I already in 1989 overclocked my Tseng ET3000 by exchanging some xtal oscillators to get higher refresh rates.
When I did that, I observed vertical stripes, just like @Kahenraz showed in a photo in his post above.
The reason was obvious: the supply voltages' surges fed back to the output signal.
The solution was easy and straightforward: improve the decoupling by adding some ceramic caps.

I observed this phenomenon also with some other graphics cards which I overclocked for better refresh rates, and adding caps always fixed the issue.
In some cases it was also necessary not only to add caps to the DAC section, but also to the CRTC and RAM sections, in particular when white/colored vertical bars appeared which are otherwise typical of bad video ram.

I just do not understand why nobody else seems to think of this simple cause and remedy.

retardware wrote on 2021-09-21, 15:12:

From a mention of this thread I curiously read here. […]
Show full quote

From a mention of this thread I curiously read here.

Kahenraz wrote on 2021-08-20, 02:47:

In text mode only, blacks are washed out and there are vertical bars when viewed on an LCD. The vertical bars do not appear on my CRT but the blacks are still washed out.

I find it just awesome that nobody seems to look at the possibility that this phenomenon could be hardware related.

As I am very flicker-sensitive, I already in 1989 overclocked my Tseng ET3000 by exchanging some xtal oscillators to get higher refresh rates.
When I did that, I observed vertical stripes, just like @Kahenraz showed in a photo in his post above.
The reason was obvious: the supply voltages' surges fed back to the output signal.
The solution was easy and straightforward: improve the decoupling by adding some ceramic caps.

I observed this phenomenon also with some other graphics cards which I overclocked for better refresh rates, and adding caps always fixed the issue.
In some cases it was also necessary not only to add caps to the DAC section, but also to the CRTC and RAM sections, in particular when white/colored vertical bars appeared which are otherwise typical of bad video ram.

I just do not understand why nobody else seems to think of this simple cause and remedy.

Getting a card from a decent manufacturer (seems that not many is getting this when buying one) and avoiding noname ones can help as image quality may vary. CRT monitor were commonly used back then so these stripes that Kahenraz showed weren't noticeable until recently when LCD monitors became mainstream.

Besides the S3 Trio3D/2X cards I currently have, none of my other S3 cards suffers from the brightness bug. I actually don't recommend a Trident card for anything, there are much better cards in terms of performance and their VESA implementation is the worst of all the cards I have tried. Only the VGA compatibility makes for it though.

Sorry for the necro but I wanted to chime in with some relevant info.

wbc wrote on 2015-10-05, 14:35:

(my previous post has not been sumbitted, dunno why)
For S3 Trio64V2/DX or GX you can try this video BIOS (in attachment) that does not have brightness issue. It is also capable of VBE 2.0 support (no more TSR's like UNIVBE or S3VBE20 in memory! well, almost 😀)

but be aware of card with this bios + Intel 440LX\EX or 440BX\ZX motherboards, seems that they crash Windows.

I just tried this bios on my PCI cardex trio64v2/dx. Original BIOS was version 1.01.05, it had the grey blacks issue, VBE 1.2, plus it was glitchy on some games and in win95. Bought a replacement chip (original was a one time program EPROM, replacement is AT29C256 part), used a programmer to write this bios on it, switched the chips and it works great. VBE 2.0 enabled, no more glitches, and the bright blacks issue is completely gone. Thank you 😀

I kept the original in case it'd ever be needed or this one has a serious problem I haven't come across again. Attaching it here in case someone out there needs it for some reason (maybe tried a flash and it didn't work for theirs so they need it back again). It does have the brightness issue so a debug command will be needed to correct it.

wbc wrote on 2016-06-02, 13:39:

patched video BIOSes for S3 Trio3D/2X with BLANK pedestal disabled (in attachment). Tested on PCI card with 4MB of SDRAM, should work with others. MCLK for SDRAM\SGRAM = 90 MHz

This worked fine on my S3 Trio3D/2X AGP. I flashed 362_C10E.BIN to an ATMEL AT27C512R OTP DIP28. I tested it first with the Shadow Video BIOS Loader, linked earlier in this thread to test it before flashing.

For some reason though, this BIOS, when flashed, only works on the two of my three cards cards. The two it works on originally had version 2.0B.06.P.

My third card already had version 2.0C.10, but a different binary when compared as a diff. This card works fine with the modified BIOS when it's loaded from shadow RAM, but not during boot when from my flashed ATMEL chip. When I try to use this it, the system will just beep in error, reporting that no video card is detected.

Does anyone know what the problem is with this BIOS or chip? I've attached a dump of the original BIOS files, for comparison.

The two cards where it worked fine use 86C362 chips. The one where it does not actually uses an 86C368 chip, but they are all labeled S3 Trio3D/2X. Interestingly, the original BIOS chip from the 86C368 works fine in the 86C362, but not the other way around.

I worked on this today, reading chips, burning chips, and comparing the results between my Trio3D/2X cards. The similarity between all cards is that they all came with 64KB ROM chips, despite the BIOS only being 32K.The difference between my two boards is in how they read the BIOS. For the brown ones, the original BIOS was written as 32K of ROM + 32K of 0xFF; I can write a 32K ROM to a new chip and anything else to the second half and it will work fine. This is NOT the case for the green PCB Trio3D/2X! For this one, I think it reads from the second half of the chip? In all instances where I only wrote 32K to the chip, it remained unbootable. What it wants is either an identical 32K+32K or just 32K in the second half (I ran out of ROM chips to test only writing to the second half).

In my case, I had to take the BIOS and duplicate it into a 64K binary and then write that. These 32K+32K chips worked fine in my brown boards. So, from my experience, the safest and most compatible way to burn one of these is to use the entire 64K and duplicate the BIOS at both the top and bottom half of the chip. This allows it to work within either card, no matter which end of the chip it reads from.

As a side effect, I have 6 extras of these chips burned with the S3 "bright fix" and 4 OEM chips without it (2.0B.06.P x 2 and 2.0C.10 x 2). PM me if you need one or want one of the original OEM chips.

I bought another one of these recently and found that it came with the BIOS on a 32K ROM chip instead of a 64K one. I used one of the 64K chips I burned previously and it worked just fine.

I guess it makes sense, since the ROM is only 32K that they might have used these in at least some cards, based on supply. If you want to flash one of these yourself, it seems to make sense to buy a 32K chip instead, that way you won't have to worry about making sure the ROM is copied into both the top and bottom parts of the memory.

This chip is also an S3 2.0C.10 BIOS, like the rest.

Cross linking here to my other thread. If you have an onboard S3 ViRGE, you may need to workaround this issue by other means.

Re: Intel TC430HX Socket 7 motherboard repair

Does anyone know why the utilities mentioned previously in this thread, such as debug, s3bright, s3black, and s3vbefix, don't seem to work with this onboard ViRGE?

Jepael wrote on 2016-06-02, 18:58:

elianda wrote:

Could you compile a small tool as well that allows to set the bit 'ON' and 'OFF' from plain DOS on the fly?

Here are two simple DOS .com programs to either turn the bit on or off, but I haven't tested these. I used NASM to compile these so I haven't tested them on S3 card. These are so simple you could just do these in debug and save them.

uau! the BLACK on a no brand Trio3d/2x
thank you so much!! 😀