Do you have a PCI card laying around?
Perhaps worth a shot having that one installed in parallel. If possible change the primary GPU to PCI.

What kind of equipment do you have available?

If you disconnect the external power, the card should prompt a warning message on the screen.

Though the gold-fingers look okay. Still though, can you give it a shot to clean them with some paper towel plus alcohol?

Also, on the backside, upper area where the external power input is located, it appears that there is some "damage" to the large trace? Or is this just some gunk?

byte_76 wrote on 2025-06-23, 19:07:

Obviously the heatsink is there for a reason so I don’t want to assume that a component is faulty just because it gets hot.
What are the function of the components under that small heatsink plate?

turn electricity into heat, bringing the warm death of the universe closer

Download the datasheet to sc1175, according to the reference design, everything will be clear

byte_76 wrote on 2025-06-23, 19:07:

Obviously the heatsink is there for a reason so I don’t want to assume that a component is faulty just because it gets hot.
What are the function of the components under that small heatsink plate?

Mosfets for the SC1175 that provides the GPU power rail.

I have an agp tester which has LED’s to indicate connection between the slot and the core.
All LED’s light up as they should so it tests fine in that respect.

I washed the card with IPA.

That spot in the top corner by the external power is some sort of hard yellow glue. It doesn’t wash off but the traces are not damaged.

If there is a blown mosfet, is it likely that the core would have been damaged?
I would assume that a blown mosfet would create a short that should prevent the PC power supply from starting but that isn’t happening here.

I have a hot air station so I can try to remove the mosfets.

I'd start with some measurements. The card is sharing the same layout as 9700s. Check the voltages on the large caps to see if the GPU is powered.

The only thing I see is that JU2 (4-pin fan connetor) does not seem to be soldered on the back side of the board, but that is a minor issue for now; you have a fan I assume?

byte_76 wrote on 2025-06-23, 20:20:

I have an agp tester which has LED’s to indicate connection between the slot and the core. All LED’s light up as they should so […]
Show full quote

I have an agp tester which has LED’s to indicate connection between the slot and the core.
All LED’s light up as they should so it tests fine in that respect.

I washed the card with IPA.

That spot in the top corner by the external power is some sort of hard yellow glue. It doesn’t wash off but the traces are not damaged.

If there is a blown mosfet, is it likely that the core would have been damaged?
I would assume that a blown mosfet would create a short that should prevent the PC power supply from starting but that isn’t happening here.

I have a hot air station so I can try to remove the mosfets.

Oddly I'm troubleshooting a reference design 9700 as well which is missing one of those dual-MOSFETs. I found a box of junk cards that I put away 15 years ago and I have no idea at all of why I left this Hercules 3D Prophet 9700 like that 😀

The 256-bit Radeon 9500 is quite a rare find I think? I wanted one back in the day to softmod but in the end got a Radeon 9700 instead. Since your card is a reference design as well you can refer to the 9700 schematic to figure out what's what in the power circuitry although still some guessing is required because the silkscreen doesn't include many component names. Just search for "105-942XX-30_11" and you should find it

byte_76 wrote on 2025-06-23, 20:20:

It doesn’t wash off but the traces are not damaged.

really?

diagnotsic steps:
1. check voltages for gpu and vram
2. boot with isa/pci vga card and check firmware on card with atiflash

got all voltages, atiflash no detect ? dead core
atiflash detect, valid firmware? dead core -)

most of the R300 cores are dead by now. bad cooling doing bad job

shevalier wrote on 2025-06-24, 03:29:

byte_76 wrote on 2025-06-23, 20:20:

It doesn’t wash off but the traces are not damaged.

really?

I think it's worth checking continuity, just for being sure. This would be the 5VEXT line, which powered on half of the SC1175. The SC1175 combines two different input rails to generate VDDC (GPU Core voltage). It uses the external 5V and the 3.3V from the AGP slot connector.

I wonder however if even a cut in that trace would lead to the behavior we see. The card has a detection for the external power by sensing on the 5VEXT line. The circuitry is composed of U93 (front side of the PCB, next to the crystal oscillator in the lower part near the AGP connector) and a couple of resistors.

Due to its position, I assume the detection happens somewhere in the region of the SC1175 and not directly at the connector. Thus leading me to believe the card would just always prompt the "External Power Missing" warning on the screen, if the trace was broken (even with the power plugged in).

Furthermore, the card should still "run". The SC1175 would just generate the core voltage rail from the 3.3V of the AGP slot, which would be enough to get something on the screen.

For starters I'd say measure some of the output voltages of all onboard regulators.
We have the following voltages that are "most interesting":

VDDC - Core voltage
MVDDC - Memory voltage
VDDQ - Memory I/O voltage
VTT - Memory termination voltage

This techpowerup volt-mod page has the right measure points for you: https://www.techpowerup.com/articles/57
(Please note they refer to VTT as VREF).

I have a 9500 256Bit roaming around as well. I'll do some measurements while it's running later on so you have exact reference values.
I will also attach a thermal probe to the small heat plate on the back so we got some reference values for temperature of that during normal operation.

Cheers

Sounds like a dead mosfet.
Classic short to ground triggers the PSU protection.

Get a multimeter and measure the resistance from the 12 and 5v pins to the backplate/ground.

The spot on the top is just a blob of glue. It’s hard and doesn’t come off with IPA.
I scraped a bit of it away with my tweezer to see if there were damaged traces underneath but there isn’t.
Confirmed with my digital microscope.

That connector on the top corner is not a fan connector, it’s the auxiliary power connector and it is soldered properly on all 4 pins.

When I install the card into a PC with the auxiliary power connected, I don’t get POST, so I can ‘t boot to any OS with a secondary card.
If I disconnect the auxiliary power, the Pentium 3 PC boots with onboard display but the card is not detected.

The card does not seem to trigger PSU protection.

I’ll try to do some of the other checks for voltage as suggested.

As mentioned, these 256-bit Radeon 9500 are quite rare, so I’d like to fix it if I can.

Check for shorts first.
I'd put money on the 12v molex pin being short.

I suggest not to breed entities, but to wait until the author restores all damaged traces

myne wrote on 2025-06-24, 05:59:

Check for shorts first.
I'd put money on the 12v molex pin being short.

With a dead short the short circuit protection of the PSU should immediately activate and the system should not power up with the external power connector plugged in. However, looking at the initial post, it does.

tehsiggi wrote on 2025-06-24, 06:04:

myne wrote on 2025-06-24, 05:59:

Check for shorts first.
I'd put money on the 12v molex pin being short.

With a dead short the short circuit protection of the PSU should immediately activate and the system should not power up with the external power connector plugged in. However, looking at the initial post, it does.

Correct, it powers up.

I have checked for shorts on the connector but it isn’t shorted.

That mark behind the core is probably just some thermal paste but I’ll check when I get home. (At work at the moment)

Oh, I misread.

byte_76 wrote on 2025-06-24, 05:57:

The spot on the top is just a blob of glue. It’s hard and doesn’t come off with IPA. I scraped a bit of it away with my tweezer […]
Show full quote

The spot on the top is just a blob of glue. It’s hard and doesn’t come off with IPA.
I scraped a bit of it away with my tweezer to see if there were damaged traces underneath but there isn’t.
Confirmed with my digital microscope.

That connector on the top corner is not a fan connector, it’s the auxiliary power connector and it is soldered properly on all 4 pins.

When I install the card into a PC with the auxiliary power connected, I don’t get POST, so I can ‘t boot to any OS with a secondary card.
If I disconnect the auxiliary power, the Pentium 3 PC boots with onboard display but the card is not detected.

The card does not seem to trigger PSU protection.

I’ll try to do some of the other checks for voltage as suggested.

As mentioned, these 256-bit Radeon 9500 are quite rare, so I’d like to fix it if I can.

Thank you, those reference voltages and points where you tested would be very helpful.

@myne, no problem, I don’t have much knowledge on electronics repair so all advice and suggestions from all members that comment are appreciated.

Hey there, as promised, here are some results.

I did measure under idle and load.

Temperatures:

Ambient: 24°C
Cooling plate: 46°C in idle, 50°C under load.

Regulator output voltages:

Voltages were basically the same under load and idle.

VDDC: 1.505V
MVDDC: 2.847V
MVDDQ: 2.498V
VTT: 1.255V

I added a drawing that shows you where I measured which voltage.
All voltages were using the slot bracket as ground reference (using a clip). That'd be good enough..

Power input:

Average values taken, you could easily measure the EXT currents using a current clamp, if available.

IDLE:
3.3V AGP: 1.456A / 4.663W
5V AGP: 0.502A / 2.571W
12V AGP: 0.164A / 2.018W

5V EXT: 2.311A / 11.742W
12V EXT: 0.213A / 2.629W

LOAD:
3.3V AGP: 1.925A / 6.169W
5V AGP: 0.612A / 3.131W
12V AGP: 0.164A / 2.020W

5V EXT: 2.498A / 12.660W
12V EXT: 0.267A / 3.229W