it's working 😀

The Geforce 4mx440 had a dying fan so I replaced it with a passive heatsink, from another mx440 that was junk (64bit bus)

The output was also quite blurry @75hz on Windows, so I searched a bit and found this topic talking about a mod to remove too much filtering on the vga lines : http://www.overclock.net/forum/69-nvidia/1442 … ce-8400-gs.html

I did the same thing as this guy : removing the capacitor and shunting the coil for each R,G,B output lines, except my card has two of them for each line (2x capacitors, 2x coils) :

And... whoa it works ! Now the VGA output is perfectly sharp. I don't know why those components were put there if they degrade the image quality.

HEAVILY BOTTLENECKED 🤣

PII, GF2, ISA soundcard, makes a good "superfast DOS SVGA gamming PC" 😀

Yeah ! I need to try more DOS games.

I ran 3D Mark again, this time with a Pentium III 650 (coppermine)

The result is 2.3 times better than with the PII 450. While the PIII clock being only 1.44 times faster. It's efficiency per cycle is much better

Nice work. Did you try just removing the capacitors to see if that fixed the issues without jumping out the coils?

Nope I didn't tried that. I'm not sure it's relevant to leave only the coils (a coil alone just delays the signal without modifying it, right?)

Those components were there to prevent RF EMI picked up by your monitor cables (which act like antennas picking up RF) from entering your video card.
A strong enough radio transmission (walking talkie, passing cop car, service truck w/radio) can now possibly fry your video card. - In theory at least.
.
For that matter a cell phone could do it.
I once worked on a system that rebooted anytime a cell phone rang within 5 feet of it.
.

Eh, that makes sense. A better solution would be to modify the filter instead of removing it, except if the frequencies it has to filter are the same as those we need to get a good image quality...

okenido wrote:

Eh, that makes sense. A better solution would be to modify the filter instead of removing it, except if the frequencies it has to filter are the same as those we need to get a good image quality...

No, I think RF is much higher frequency. - I'm out of practice with that sort of thing and I could be wrong.

Other cards seem to pull it off just fine. Clearly there is a compromise that works.
.

The caps should help any voltage drops from the signal imo

Then coils in general are to help resist and current/voltage change

Many higher quality cables have shielding and ferrite beads on them. These items are to help with rf/emi

I would suspect the caps create resonence in certain frequncies that can show up as ringing in the picture.

RogueTrip2012 wrote:

The caps should help any voltage drops from the signal imo

Then coils in general are to help resist and current/voltage change

Yes caps do that too but caps have multiple jobs simultaneously.
As part of an LC filter the inductor resists the current changes while the cap shunts the noise (EMI/Ripple) to ground.
The L & C values determine which frequency (ranges) get dumped to ground and which don't.

RogueTrip2012 wrote:

Many higher quality cables have shielding and ferrite beads on them. These items are to help with rf/emi

Those help a lot.

RogueTrip2012 wrote:

I would suspect the caps create resonence in certain frequncies that can show up as ringing in the picture.

I think that can happen with poorly chosen values of L & C and I think that might be part of the problem with this card.
.

🤣, earlier I was thinking why they don't use a diode or zener for shunting but was forgetting we are only using 1v p-p signals most of the time.

At sometime I'd like to look at a "good" design vs. a not "good" design. Seems like alot of manufacturers cheapened out during Geforce 1 to 4 series from my previous readings. I had thought they fixed most around the Geforce 5 series but that 8xxx shows a different story aside from poor cable quality and high res/refresh rates.