I also have an ET4000AX based card that exhibits the same "issue" so I'd be curious to know the explanation for it 😀

I've seen a similar behaviour on Trident cards when there is a palette change without waiting for the VSync. I'm pretty sure it's the same issue, nothing you can do about as is a problem on the video chip design.

ViTi95 wrote on 2022-12-19, 09:24:

I've seen a similar behaviour on Trident cards when there is a palette change without waiting for the VSync. I'm pretty sure it's the same issue, nothing you can do about as is a problem on the video chip design.

Actually, as the "noise" is just at the top of the screen, this is a palette change (all 256 colors) with waiting for VSync. It seems the VSync interval isn't long enough to permit the complete palette change to happen during the blanking interval. A palette change is around 750 8-bit output cycles. You might be able to mitigate the problem by makeing 8-bit output cycles faster. If the BIOS setup of your machine allows you to choose "8 bit I/O recovery time", choose the lowest number.

Also, as it seems, not all DAC chips are susceptible to this problem, so if you have a different VGA card and the DAC chip is socketed, you might try swapping the DAC chip to see if the problem disappears.

mkarcher wrote on 2022-12-19, 18:36:

ViTi95 wrote on 2022-12-19, 09:24:

I've seen a similar behaviour on Trident cards when there is a palette change without waiting for the VSync. I'm pretty sure it's the same issue, nothing you can do about as is a problem on the video chip design.

Actually, as the "noise" is just at the top of the screen, this is a palette change (all 256 colors) with waiting for VSync. It seems the VSync interval isn't long enough to permit the complete palette change to happen during the blanking interval. A palette change is around 750 8-bit output cycles. You might be able to mitigate the problem by makeing 8-bit output cycles faster. If the BIOS setup of your machine allows you to choose "8 bit I/O recovery time", choose the lowest number.

Also, as it seems, not all DAC chips are susceptible to this problem, so if you have a different VGA card and the DAC chip is socketed, you might try swapping the DAC chip to see if the problem disappears.

I have a trident card but I can't find any DAC chip unlike on the ET4000AX so I assume it isn't socketed. The bios has no 8-bit option, it is a really old bios. It doesn't even support auto detect HDD or LBA. It is N418 motherboard, I have attached a picture of the bios chip on the motherboard if it is useful.

zconnect wrote on 2022-12-19, 20:15:

I have a trident card but I can't find any DAC chip unlike on the ET4000AX so I assume it isn't socketed. The bios has no 8-bit option, it is a really old bios. It doesn't even support auto detect HDD or LBA. It is N418 motherboard, I have attached a picture of the bios chip on the motherboard if it is useful.

I guess your Trident card is using the 9000i chip, then. The "i" is for "integrated", as the 9000i integrates the clock generator and the DAC into the main VGA chip.

The BIOS chip on your board has an AMI sticker, so likely it contains an AMI BIOS. If you manage to obtain a copy of AMISETUP (a 3rd party tool), you can adjust chipset options the mainboard vendor decided to not display to the user in the BIOS setup screen. Possibly you can find a wait state / recovery time option there. If not, you are likely out of luck, as the there seems to be no publicly available datasheet for the ALi m1219 chipset.

The artifacts you show were already present on some 90's computers when they were new, so they are not that something broke over time and needs to be fixed to be back in "original shape", but this is a quirk people lived with in the 90's, so I suggest that you just accept it as a part of the "retro experience".