Any settings in BIOS, like they can be found in Notebooks ("normal/dim/bright")?
Otherwise I´d suspect a hardware issue around the graphics chip = northbridge.

No settings for that, it's just regular AWARD BIOS of the time (I flashed it to latest version, but it did not change anything).
I don't think a nothbridge is a culprit, I suspect something between northbridge and VGA connector.
Or maybe shorted connector? Shorted cap? Screwy resistor?
Nothing is obviously hot, so I keep on searching through SMD components.

Exactly - that´s why I wrote "around the chip", although the chip itself *could* be the culprit in the end.
I would focus on this area first:

Yeah, I'll give a multimeter a go over this.
Not that this graphics is a thing worth saving, Riva Vanta TNT2 put in PCI slot makes mucha better job, but I just want that board to be 100% fine.

vga output is accomplished thru DACs, and those require VREF (reference voltage to set the proper levels). just an example https://www.analog.com/media/en/technical-doc … ets/adv7120.pdf
what are the markings on D15 diode?

Here it is.
Besides, it looks to me like it is not getting enough output voltage.

sorry, D16 😜 maybe also D11 - markings are on the side

Maybe a thing to consider is that the ESR for those caps has gone high for high frequencies., even though they look unbloaty.

vga signal doesnt cross electrolytic caps 😀 only inductors with pico range caps going to ground (low pass filter) and clamping diodes for some ESD protection.
Sadly I dont think I have diagrams for such old motherboards 🙁 VT8501 chipset has integrated Trident Blade3D core. If VIA just plopped Trident design into their own then Vref is internal so there is nothing to be done here.

D16 is 'MS 20' (Schottky diode?)
D11 is 'TL 431A 9E33' (voltage reference?)

smoke86 wrote on 2022-10-19, 19:38:

I have a problem with a Lucky Star 5MMVP4 board - onboard graphics adapter is very (and I mean VERY) dim - to the point I cannot see anything during POST.
...

Hi,
There's a schematic diagram of a similar mainboard ( GA-5VMM VIA MVP4 NORTH BRIDGE VT82C501 SOUTH BRIDGE VT82C686 ).
Maybe it'll help

smoke86 wrote on 2022-10-20, 09:25:

D16 is 'MS 20' (Schottky diode?)
D11 is 'TL 431A 9E33' (voltage reference?)

can you be this lucky? 😀 measure voltage on D11, it can be the source for our video DAC Vref. If its blown and not providing set voltage all teh outputs will be very low or zero. Some random reference voltage tutorial for beginners https://www.youtube.com/watch?v=NecbHbGAkkc
Its also possible its a part of some power supply together with the Q4 mosfet and has nothing to do with Video 🙁

Edit:

Roman555 wrote on 2022-10-20, 15:32:

There's a schematic diagram of a similar mainboard ( GA-5VMM VIA MVP4 NORTH BRIDGE VT82C501 SOUTH BRIDGE VT82C686 ).

This is fantastic! 👍 D11 together with Q4 might be providing dedicated analog supply for VT8501. On your diagram Gigabyte opted to just powering Analog section from single LDO providing Digital 2.5V supply.

So smoke86, measure D11 (pin 1 reference, its the one closest to the battery) and Q4 (drain and source, middle pad and pin closest to D11) voltage

D11 reference to (ceb703al, which looks like mosfet) source measured REALLY close to 0 (something like 0.009V). Does it means D11 is blown?

How is D11 wired? are pins 1 and 3 connected to each other? I would solder it out and measure out of circuit
if you want to quickly check if thats it without waiting for parts you can look for another TL431 in old junk power supplies

Edit: wait. I didnt mean "D11 reference to source". I meant measure all with reference to Ground. so
- D11 pin 1 to ground voltage
-Q4 drain to ground
-Q4 source to ground

- D11 pin 1 to ground voltage - >2,47V
-Q4 drain to ground - > 4,76V
-Q4 source to ground - > 2,49V

smoke86 wrote on 2022-10-20, 18:55:

-Q4 source to ground - > 2,49V

Looks o.k.

Sorry, drain to ground is 4,46, not 4,74V

Yes, sadly looks ok. You could maybe try to follow Q4 source track/vias all the way to the chip, maybe there is a fuse/inductor in the way that is blown/chipped.
other than that set meter to diode mode and (on powered off system) check 5 diodes there. Its highly unlikely, but theoretically possible that all ESD protection clamping diodes are blown?

Diodes are ok.
Cannot see anything crooked yet.
Here's a photo of that area.