LCD monitor?

wierd_w wrote on 2025-07-28, 21:21:

LCD monitor?

Or maybe the VGA cable

wierd_w wrote on 2025-07-28, 21:21:

LCD monitor?

Already tried with a CRT a while back and the grainy graphics were still there, sadly.

darry wrote on 2025-07-28, 21:27:

Or maybe the VGA cable

Nope, already tried with a different one, plus the issue doesn’t happen with other PCs using that same vga cable 🙁

Maybe something in the PC is causing this.

Maybe something wrong with the PSU or motherboard (out of spec capacitors ?) or an environmental factor like a source of low frequency RF noise in a bad quality switching transformer .

Basically, what I think this looks like (and looks like Darry is thinking similar)

This looks like an analog to digital conversion anomaly, (from a marginal, or not quite in spec) analog VGA signal.

I've seen similar happen a lot on lcd monitors with less than stellar vga support, when the Hsync/ scanline timing signal rate is not absolutely perfect.

On good monitors, similar can happen if the analog signal is degraded in some way, putting the signal into the 'not really in spec' category, which is where I think Darry is going.

The analog circuitry inside a vga crt probably wont have any troubles here.

if you have an analog CRT, I would suggest you test against it, to see if the issue persists. If it does persist, you might have a very noisy ground plane, causing bus noise, and ultimately causing video signal noise. (sound cards on the isa bus should also have odd noises if this was the case though.)

If you have an oscilloscope, we could directly test for ground noise.

As a 'throw at the wall and see what sticks' solution, try adding ferrite beads to the power cables of both the PC and the monitor.

wierd_w wrote on 2025-07-28, 21:38:

Basically, what I think this looks like (and looks like Darry is thinking similar) […]
Show full quote

Basically, what I think this looks like (and looks like Darry is thinking similar)

This looks like an analog to digital conversion anomaly, (from a marginal, or not quite in spec) analog VGA signal.

I've seen similar happen a lot on lcd monitors with less than stellar vga support, when the Hsync/ scanline timing signal rate is not absolutely perfect.

On good monitors, similar can happen if the analog signal is degraded in some way, putting the signal into the 'not really in spec' category, which is where I think Darry is going.

The analog circuitry inside a vga crt probably wont have any troubles here.

if you have an analog CRT, I would suggest you test against it, to see if the issue persists. If it does persist, you might have a very noisy ground plane, causing bus noise, and ultimately causing video signal noise. (sound cards on the isa bus should also have odd noises if this was the case though.)

If you have an oscilloscope, we could directly test for ground noise.

As a 'throw at the wall and see what sticks' solution, try adding ferrite beads to the power cables of both the PC and the monitor.

Thanks for the response, I’ve updated the first post with a list of things I tried, already tried a CRT monitor and even moved between countries since I first had this issue (so its not something in my house in particular causing interference).
The only constant in all these years is the PSU, which is something I plan on replacing this week. If that doesn’t solve it, then maybe it has to do with capacitors, thing is, this is happening for over 12 years and it never got better or worse… wouldn’t a bad capacitor make the issue get worse over time?
Also, the issue only happens in an horizontal section of the screen, which may change sometimes (I think it depends on the refresh rate, but not sure). In case of Outlaws, the grainy effect only happens when I have one of the options highlighted, if nothing is selected, there is no grainy effect at all.

A poorly considered ground plane for the motherboard would have issues with 'electronically resonant' RF noise sources saturating it, and then those signals causing changes in effective voltage on analog circuits (like the stereo amp in a soundcard, or the analog signals generated by the vga card, using that ground.)

Ferrite beads are a cheap way to clamp down transient RF signal bleed on cables.

A dodgy psu in the computer could produce such RF from out of spec caps or coil whine.

Similar could happen with the motherboard itself from failing or not fully in spec voltage regulators, failing caps, or bad grounding.

gaffa2002 wrote on 2025-07-28, 21:51:

Thanks for the response, I’ve updated the first post with a list of things I tried, already tried a CRT monitor and even moved b […]
Show full quote

wierd_w wrote on 2025-07-28, 21:38:

Basically, what I think this looks like (and looks like Darry is thinking similar) […]
Show full quote

Basically, what I think this looks like (and looks like Darry is thinking similar)

This looks like an analog to digital conversion anomaly, (from a marginal, or not quite in spec) analog VGA signal.

I've seen similar happen a lot on lcd monitors with less than stellar vga support, when the Hsync/ scanline timing signal rate is not absolutely perfect.

On good monitors, similar can happen if the analog signal is degraded in some way, putting the signal into the 'not really in spec' category, which is where I think Darry is going.

The analog circuitry inside a vga crt probably wont have any troubles here.

if you have an analog CRT, I would suggest you test against it, to see if the issue persists. If it does persist, you might have a very noisy ground plane, causing bus noise, and ultimately causing video signal noise. (sound cards on the isa bus should also have odd noises if this was the case though.)

If you have an oscilloscope, we could directly test for ground noise.

As a 'throw at the wall and see what sticks' solution, try adding ferrite beads to the power cables of both the PC and the monitor.

Thanks for the response, I’ve updated the first post with a list of things I tried, already tried a CRT monitor and even moved between countries since I first had this issue (so its not something in my house in particular causing interference).
The only constant in all these years is the PSU, which is something I plan on replacing this week. If that doesn’t solve it, then maybe it has to do with capacitors, thing is, this is happening for over 12 years and it never got better or worse… wouldn’t a bad capacitor make the issue get worse over time?
Also, the issue only happens in an horizontal section of the screen, which may change sometimes. In case of Outlaws, the grainy effect only happens when I have one of the options highlighted, if nothing is selected, there is no grainy effect at all.

Please don't feel you need to try a different planet. 😀

More seriously, the best approach, IMHO, once you've tried another PSU, is to strip down that PC to the bare minimum and start swapping parts, one by one.

It's tedious, I know, but the process of elimination should allow you to narrow it down.

Best of luck to you.

darry wrote on 2025-07-28, 21:58:

Please don't feel you need to try a different planet. :) […]
Show full quote

gaffa2002 wrote on 2025-07-28, 21:51:

Thanks for the response, I’ve updated the first post with a list of things I tried, already tried a CRT monitor and even moved b […]
Show full quote

wierd_w wrote on 2025-07-28, 21:38:

Basically, what I think this looks like (and looks like Darry is thinking similar) […]
Show full quote

Basically, what I think this looks like (and looks like Darry is thinking similar)

This looks like an analog to digital conversion anomaly, (from a marginal, or not quite in spec) analog VGA signal.

I've seen similar happen a lot on lcd monitors with less than stellar vga support, when the Hsync/ scanline timing signal rate is not absolutely perfect.

On good monitors, similar can happen if the analog signal is degraded in some way, putting the signal into the 'not really in spec' category, which is where I think Darry is going.

The analog circuitry inside a vga crt probably wont have any troubles here.

if you have an analog CRT, I would suggest you test against it, to see if the issue persists. If it does persist, you might have a very noisy ground plane, causing bus noise, and ultimately causing video signal noise. (sound cards on the isa bus should also have odd noises if this was the case though.)

If you have an oscilloscope, we could directly test for ground noise.

As a 'throw at the wall and see what sticks' solution, try adding ferrite beads to the power cables of both the PC and the monitor.

Thanks for the response, I’ve updated the first post with a list of things I tried, already tried a CRT monitor and even moved between countries since I first had this issue (so its not something in my house in particular causing interference).
The only constant in all these years is the PSU, which is something I plan on replacing this week. If that doesn’t solve it, then maybe it has to do with capacitors, thing is, this is happening for over 12 years and it never got better or worse… wouldn’t a bad capacitor make the issue get worse over time?
Also, the issue only happens in an horizontal section of the screen, which may change sometimes. In case of Outlaws, the grainy effect only happens when I have one of the options highlighted, if nothing is selected, there is no grainy effect at all.

Please don't feel you need to try a different planet. 😀

More seriously, the best approach, IMHO, once you've tried another PSU, is to strip down that PC to the bare minimum and start swapping parts, one by one.

It's tedious, I know, but the process of elimination should allow you to narrow it down.

Best of luck to you.

Hahaha! Well, I’m a desperate man.
Thanks guys for the tips, I really hope the PSU is the cause, although I have used it in a different PC and there was no such issue. Most likely is something on motherboard, will also try the ferrite beads suggestion and checking the components one by one.

Make sure the chasis is clean, and free of corrosion, where the PSU contacts it. (Eg, 'clean continuity of the communal ground with the PSU')

wierd_w wrote on 2025-07-28, 22:12:

Make sure the chasis is clean, and free of corrosion, where the PSU contacts it. (Eg, 'clean continuity of the communal ground with the PSU')

Already tried with a different chasis, even tried plugging everything on top of a cardboard box and the issue is still there, happening in the same place.
The interference is also quite stable, no matter if I move or wiggle cables around, there is no change at the screen at all.

This sort of looks like palette updates "snow" when the video card is showing previous pixel in where update is happening (which is what the ATI Rage chip is doing)... but the game shouldn't be doing it all the time like that (but I am unfamiliar with that game). It should be noticable during things like fade ins and outs etc. when normally palette is being updated.

Tiido wrote on 2025-07-29, 15:02:

This sort of looks like palette updates "snow" when the video card is showing previous pixel in where update is happening (which is what the ATI Rage chip is doing)... but the game shouldn't be doing it all the time like that (but I am unfamiliar with that game). It should be noticable during things like fade ins and outs etc. when normally palette is being updated.

I think that, too. The effect is usually noticeable when games are fading in/out or flashing. But the effect doesn’t necessarily happen on the area which has the color change, it happens in a fixed area of the screen.
Today I tried the bare minimum (CPU, RAM and HDD only) and the issue is still there, so it’s not related to any other component. Only thing that changes it is by using another video card which greatly reduces the effect, but its still there…
A new PSU should arrive today or tomorrow, this is the only component I never changed throughout the years.

Where exactly the effect happens is purely down to the exact point of a plette update and often PC games are not at all synchronized with video output as things are on game consoles. So on PC the snow can be anywhere on the frame. It won't only affect the particular color in the palette being updated but the pixel that is being shown when the update is happening. That pixel gets its color from the palette, and when palette is getting written into the memory becomes unavailable for the image output process and there are two things the video output can do :
* Show the data being written, which will cause the actual snow effect the problem is named after. Early video cards suffered from this in particular
* Repeat previous pixel, which gives you artifacts still but they are less offensive.

Better option is to use dual ported memory but that is an expensive thing (in cost and also silicon realestate) and generally not done because of it.

Another video card changing the behaviour strongly points to this being the issue.

Tiido wrote on 2025-07-29, 15:27:

Where exactly the effect happens is purely down to the exact point of a plette update and often PC games are not at all synchron […]
Show full quote

Where exactly the effect happens is purely down to the exact point of a plette update and often PC games are not at all synchronized with video output as things are on game consoles. So on PC the snow can be anywhere on the frame. It won't only affect the particular color in the palette being updated but the pixel that is being shown when the update is happening. That pixel gets its color from the palette, and when palette is getting written into the memory becomes unavailable for the image output process and there are two things the video output can do :
* Show the data being written, which will cause the actual snow effect the problem is named after. Early video cards suffered from this in particular
* Repeat previous pixel, which gives you artifacts still but they are less offensive.

Better option is to use dual ported memory but that is an expensive thing (in cost and also silicon realestate) and generally not done because of it.

Another video card changing the behaviour strongly points to this being the issue.

You are right, thanks! You seem to have found the reason why the snowing is happening at least (I'll start using the correct name from now on) 😀. Another game in which the snow happens is Virtua Fighter 2: During the fights, because the health bars have a flashing effect similar to Outlaws main menu (achieved by changing the color palette each frame). Basically any game which has constant palette updates going on will have the effect until the updates stop. Also in Virtua Fighter 2, there is a lot of snow effects in the options menu, probably because the palette is constantly updated to make the selected option blink.
So, it's very unlikely the new PSU will change anything 🙁

No idea if there is any solution for that, or if that's simply an issue with the Ati Rage Pro that I'll have to live with. Thats too bad because this card has some decent direct 3D support for games, much better than the Diamond Stealth infamous 3D (De)Accelerator (which looks much better with very little "snow", despite still having it).

Oddly enough, if I put Outlaws in windowed mode, the issue goes away, but I think that's just because it may use a different method for rendering the flashing effect when windowed.

Replaced the PSU, nothing changed 🙁.
It seems its either an issue with my onboard ATI Rage Pro or plain incompatibility with the games that modify the color palette for visual effects.
Well, guess I'll stick to the Diamond Stealth 2ooo as the main card for this system. It's better for DOS game compatibility with the tradeoff of losing direct3d and OpenGL support in windows and even being a bit slower in directdraw mode compared to the Rage Pro.
Thats ok since the main focus of this machine is DOS and some very early Windows 95 games anyway. Building retro PCs like this makes me remember how fast things moved back then... you can perfectly play most games released until 1996 with a system like this, but games from 1998 or even 1997 start to become painfully slow.

duke3d is another good example of this, the health pickup and damage screen tint effects cause those snow effects on a lot of cards.

Tiido wrote on 2025-07-29, 15:27:

Better option is to use dual ported memory but that is an expensive thing (in cost and also silicon realestate) and generally not done because of it.

so according to that, should (could) using cards with VRAM/WRAM alleviate this issue? potentially SGRAM as well, although since it was optional on some cards its features may not necessarily have been used.

comparing an s3 864/868 vs. 964/968 might be useful in this regard.

auron wrote on 2025-08-02, 11:05:

duke3d is another good example of this, the health pickup and damage screen tint effects cause those snow effects on a lot of ca […]
Show full quote

duke3d is another good example of this, the health pickup and damage screen tint effects cause those snow effects on a lot of cards.

Tiido wrote on 2025-07-29, 15:27:

Better option is to use dual ported memory but that is an expensive thing (in cost and also silicon realestate) and generally not done because of it.

so according to that, should (could) using cards with VRAM/WRAM alleviate this issue? potentially SGRAM as well, although since it was optional on some cards its features may not necessarily have been used.

comparing an s3 864/868 vs. 964/968 might be useful in this regard.

The Diamond Stealth 2000 that OP has is based on a S3 Virge 325 with EDO RAM and a 135MHz integrated RAMDAC. I owned and used a card base on that same chip back in the day and I was playing Duke Nukem 3D a lot, mostly, AFAICR, at 512x384 (using Dietmar Meschede's S3 VBE/Core 2.0 utility) and I do not recall seeing the effect that you mention on a 15" CRT. AFAICR, my card was clocked at 50 or 55MHz.

The ATI Rage Pro typically used SDRAM clocked at 75MHz, if I am not mistaken.

OP mentioned that the issue is more less apparent on the Diamond Virge based card than on the ATI Rage Pro one. If the issue was purely related to memory bandwidth, I would have expected the opposite.

darry wrote on 2025-08-02, 12:29:

OP mentioned that the issue is more apparent on the Diamond Virge based card than on the ATI Rage Pro one. If the issue was purely related to memory bandwidth, I would have expected the opposite.

i'm understanding OPs posts the other way around - his virge card seems to show the issue less than the rage pro in the games that he tested.

the other thing is that OP seems to be talking only about windows games, while duke3d is DOS. in DOS ati cards generally force vsync, which will make them less affected than nvidia cards that don't, for instance. according to the (hardly comprehensive) notes i made on this, the TNT2 with SGRAM shows the issue in duke3d while the voodoo3 with SDRAM doesn't. but the voodoo3 has similar looking but consistent snow issues in VGA games that have also been discussed in several threads here.

so although the effect can look similar in different games, there are a lot of other factors going on here.