Blurriness in the video signal can be caused by poorly designed filters in the cards output stage. Those filters are located right next to the VGA connector and usually consist of 2 capacitors and an inductor for each of the RGB lines. These filters are used to eliminate high frequencies from the VGA signal in order to reduce RF emissions. You can increase the VGA signal sharpness by 'improving' the output filter and probably violate RF emissions standards.

A dirty but effective method is:
For each of the RGB lines, remove the 2 capacitors and replace the inductor with a wire bridge.

Kahenraz wrote on 2026-09-02, 23:15:

The problem I have with these cards, however, is the very poor VGA signal quality. I have tried both PCI and AGP Rage IIc, Rage Turbo, and RageXL. Higher refresh rates are also exceptionally blurry.

That's odd. I also have a Rage XL (specifically, this one) and its VGA output quality is very good. In fact, I was surprised how crisp the desktop looked at 1600x1200 on that card when using an LCD monitor.

Dorunkāku wrote on 2026-09-03, 06:39:

and replace the inductor with a wire bridge.

That’s exactly the issue.
This isn't an inductor, it's a ferrite bead—with rising frequency, this component dissipates the signal as heat. Kind of a half-baked inductor and a hyper-resistor.
They became a trend back then and were put everywhere willy-nilly.

It's best to keep one capacitor and remove the second one along with the short ferrite bead, so that at least some filtering remains.

I don't know what difference in image quality a inductor or a ferrite bead would make, but I found this in the CL-GD5446 Databook:
"There must be π-LC filters on each of the RGB lines, as shown in the reference designs. The
recommended component values are 10 pF for the capacitors. The inductor is a ferrite bead,
with 10–20 Ω impedance at 100 MHz.
There is a trade-off involved in the selection of these component values. Obtaining crisp pixels
on screen requires that the rise and fall times be as short as possible. However, to obtain
acceptable emissions testing results, relatively slow rise and fall times are preferred. As the
pixel rates get higher, there will be less margin between these two conflicting requirements.
The component values recommended above represent the Cirrus logic recommendation as of
this writing. Filter components must be placed as close as possible to the VGA DB15 connector."

Dorunkāku wrote on 2026-09-03, 08:38:

I don't know what difference in image quality a inductor or a ferrite bead would make

I’m sorry, but there’s a whole load to say on this topic.

If you’re interested, ask Google AI three consecutive questions in a single window.
The answers match what I know, so it’s not just him imagination.

1. Problems of using ferrite beads for filtering analog RGB- video 120MHz signals in a CLC-filter
2. reconstruction filter for videoDAC aliasing
3. reconstruction filter for videoDAC with using ferrite bead

The general idea is that, as always, they were looking for a way to make it cheaper.
It works out cheaper with ferrite beads.
But with ferrite beads, you can’t just rely on calculations in the CAD systems of the time; you need to check how it actually turns out – that is, look at the monitor with your own eyes.
Which they didn’t do this time.

But because this is, amongst other things, a post-RAMDAS reconstruction filter, at least some filtering (capacitance) must remain there.
In other words, it is not advisable to remove both capacitors.

Joseph_Joestar wrote on 2026-09-03, 06:50:

Kahenraz wrote on 2026-09-02, 23:15:

The problem I have with these cards, however, is the very poor VGA signal quality. I have tried both PCI and AGP Rage IIc, Rage Turbo, and RageXL. Higher refresh rates are also exceptionally blurry.

That's odd. I also have a Rage XL (specifically, this one) and its VGA output quality is very good. In fact, I was surprised how crisp the desktop looked at 1600x1200 on that card when using an LCD monitor.

I have the same experience as you, so its strange that the OP says its not very good.
I have the same card as you, it came bundled with a motherboard I bought in the early 2000s and its been with me ever since.
Its one of the few pieces of hardware thats always been there with me... Software its Win2k Pro, hardware its that RageXL.
It feeds a Dell 2007FPb very well indeed at its native 1600x1200 resolution.

To the OP I would ask if its a nice quality cable thats being used before playing with the hardware in the card.

ElectroSoldier wrote on 2026-09-04, 13:34:

I have the same experience as you, so its strange that the OP says its not very good. I have the same card as you, it came bundl […]
Show full quote

Joseph_Joestar wrote on 2026-09-03, 06:50:

Kahenraz wrote on 2026-09-02, 23:15:

The problem I have with these cards, however, is the very poor VGA signal quality. I have tried both PCI and AGP Rage IIc, Rage Turbo, and RageXL. Higher refresh rates are also exceptionally blurry.

That's odd. I also have a Rage XL (specifically, this one) and its VGA output quality is very good. In fact, I was surprised how crisp the desktop looked at 1600x1200 on that card when using an LCD monitor.

I have the same experience as you, so its strange that the OP says its not very good.
I have the same card as you, it came bundled with a motherboard I bought in the early 2000s and its been with me ever since.
Its one of the few pieces of hardware thats always been there with me... Software its Win2k Pro, hardware its that RageXL.
It feeds a Dell 2007FPb very well indeed at its native 1600x1200 resolution.

To the OP I would ask if its a nice quality cable thats being used before playing with the hardware in the card.

Those are good ones and also 640x480 looks clean to me.

Joseph_Joestar wrote on 2026-09-03, 06:50:

Kahenraz wrote on 2026-09-02, 23:15:

The problem I have with these cards, however, is the very poor VGA signal quality. I have tried both PCI and AGP Rage IIc, Rage Turbo, and RageXL. Higher refresh rates are also exceptionally blurry.

That's odd. I also have a Rage XL (specifically, this one) and its VGA output quality is very good. In fact, I was surprised how crisp the desktop looked at 1600x1200 on that card when using an LCD monitor.

My RageXL PCI is actually a little different. The PCI version has an additional little black component at the VGA output.

Does this make the signal better or worse? What is it for?

marxveix wrote on 2026-09-04, 14:12:

ElectroSoldier wrote on 2026-09-04, 13:34:

I have the same experience as you, so its strange that the OP says its not very good. I have the same card as you, it came bundl […]
Show full quote

Joseph_Joestar wrote on 2026-09-03, 06:50:

That's odd. I also have a Rage XL (specifically, this one) and its VGA output quality is very good. In fact, I was surprised how crisp the desktop looked at 1600x1200 on that card when using an LCD monitor.

I have the same experience as you, so its strange that the OP says its not very good.
I have the same card as you, it came bundled with a motherboard I bought in the early 2000s and its been with me ever since.
Its one of the few pieces of hardware thats always been there with me... Software its Win2k Pro, hardware its that RageXL.
It feeds a Dell 2007FPb very well indeed at its native 1600x1200 resolution.

To the OP I would ask if its a nice quality cable thats being used before playing with the hardware in the card.

Those are good ones and also 640x480 looks clean to me.

Mine spent most of its long years doing 1024x768

marxveix wrote on 2026-09-04, 14:12:

ElectroSoldier wrote on 2026-09-04, 13:34:

I have the same experience as you, so its strange that the OP says its not very good. I have the same card as you, it came bundl […]
Show full quote

Joseph_Joestar wrote on 2026-09-03, 06:50:

That's odd. I also have a Rage XL (specifically, this one) and its VGA output quality is very good. In fact, I was surprised how crisp the desktop looked at 1600x1200 on that card when using an LCD monitor.

I have the same experience as you, so its strange that the OP says its not very good.
I have the same card as you, it came bundled with a motherboard I bought in the early 2000s and its been with me ever since.
Its one of the few pieces of hardware thats always been there with me... Software its Win2k Pro, hardware its that RageXL.
It feeds a Dell 2007FPb very well indeed at its native 1600x1200 resolution.

To the OP I would ask if its a nice quality cable thats being used before playing with the hardware in the card.

Those are good ones and also 640x480 looks clean to me.

A resolution of 640×480 is practically TV standard; it will always work fine. You’d have to try very, very hard to mess it up.
640*480*60HZ= 18.4MHz
1024*768*100Hz=78.7MHz
These aren’t just different galaxies — they’re different universes.
Nowadays, a frequency of 20 MHz is practically regarded as a DC signal from a printed circuit board design perspective.

shevalier wrote on 2026-09-05, 07:57:

A resolution of 640×480 is practically TV standard; it will always work fine. You’d have to try very, very hard to mess it up.
640*480*60HZ= 18.4MHz

Apparently not so hard... 640x480 uses H/V timings of 800 clocks x 525 lines, giving a progressive pixel clock of 25.175Mhz. A "TV standard" signal would be half of that because it would be interlaced.

I have been digging into this more deeply and have come to the conclusion that it isn't the VGA signal quality but some weird artifacting by the chip or the driver. I found the same behavior between the RageXL PCI, AGP, Rage Turbo, and even the Rage IIc in my testing. Maybe this is some bug that has always existed across the early Rage series of cards.

This bug only exists when in 256-color mode and is confirmed in pure DOS, Windows 3.1, and Windows 98. It can be easily induced in Windows by running SimCity 2000 and setting the speed to maximum.

This is unfortunate, as the Rage series has been a good fallback card with drivers for both Windows 3.1 and Windows 98. I tend to use 256 colors in Windows 3.1 and it is also a very common display mod in DOS, so it is definitely very frustrating to have discovered this.

Also, ignore the graininess and minor pixel flickering in the Quake example. This is due to capturing from VGA and then trying to flatten an imperfect capture to 256 colors. The bug I am describing is visible within the ammo bar, specifically.

Kahenraz wrote on Today, 05:56:

I have been digging into this more deeply and have come to the conclusion that it isn't the VGA signal quality but some weird ar […]
Show full quote

I have been digging into this more deeply and have come to the conclusion that it isn't the VGA signal quality but some weird artifacting by the chip or the driver. I found the same behavior between the RageXL PCI, AGP, Rage Turbo, and even the Rage IIc in my testing. Maybe this is some bug that has always existed across the early Rage series of cards.

This bug only exists when in 256-color mode and is confirmed in pure DOS, Windows 3.1, and Windows 98. It can be easily induced in Windows by running SimCity 2000 and setting the speed to maximum.

This is unfortunate, as the Rage series has been a good fallback card with drivers for both Windows 3.1 and Windows 98. I tend to use 256 colors in Windows 3.1, so it is definitely very frustrating to have discovered this.

Yeah I dont remember anything like that on mine happening.
But mine wasnt ever really used for games. But ATi didnt have the best rep regarding drivers.

I've used the RageXL for years and never noticed it before. I remembered seeing some glitching when running SimCity 2000 in Windows 3.1, but I always thought it was the signal quality going to my monitor or something to do with the drivers. I wasn't until recently that I just started testing this further with a capture card and tracked it down to any 8-bit color mode, regardless of the operating system.

The glitching is very minor and not always obvious. It's possible that it was an oversight at ATI and then no one bothered to thoroughly test 8-bit colors after some point in development.

To me this looks like Linear Framebuffer (LFB) quirks. Used to drive me nuts at some point until I figured what it was. Take a look within the BIOS settings on your board and see if there's an option to disable it.

Garrett W wrote on Today, 07:06:

To me this looks like Linear Framebuffer (LFB) quirks. Used to drive me nuts at some point until I figured what it was. Take a look within the BIOS settings on your board and see if there's an option to disable it.

There is nolfb.com, can it fix it or what solution worked for you if any?