Only 5 attachment per post. Here the last 2 pictures.

On the 2nd picture I measured the cap that was loose (C519) and it measured 1.2 uF. Isn't that kind of high for a 0402 cap? I'm asking because it doesn't look particularly damaged but it gave me a hard time to get a reading out of it.
I also took off and measured C456 and it's 120 nF. Judging by the pattern I'm assuming that the missing C471 is also 120 nF

On picture 5, i took off cap C614 off and it had the one pad loose. I tried to measure it and i got all sorts of values ranging from 50nF to 150 nF. I don't know if I can trust these measurements but it smells like a 100 nF cap to me.

Regarding U7 I measured the voltage of the pins and looks like some sort of 5V to 3.3V converter that is working properly. Attached picture with voltage.

My smartphone camera isn't the new hotness it once was. Hope you can make out what you're looking for.

.. and finally U7

Thanks mate for the effort. Unfortunately, only resistors and tantalum smd caps have markings on them. I replaced the broken capacitor near the BIOS chip with a 100 nF and now the card posts fine and boots into windows. However, driver install returns an error that no ATI card was found. This maybe a software glitch since I frequently change cards and install new drivers on this platform. Or maybe the value of the cap is wrong. Since it made the card work I'm guessing it's a series capacitor and the value is much more relevant. Or maybe the other missing components are the problem.

Yeah could only show the colors of the missing caps really. And measuring them in place with just a multimeter won't tell you much.

Booting is progress. Keep at it!

Retronerd878 wrote on 2026-06-07, 09:56:

Thanks mate for the effort. Unfortunately, only resistors and tantalum smd caps have markings on them. I replaced the broken capacitor near the BIOS chip with a 100 nF and now the card posts fine and boots into windows. However, driver install returns an error that no ATI card was found. This maybe a software glitch since I frequently change cards and install new drivers on this platform. Or maybe the value of the cap is wrong. Since it made the card work I'm guessing it's a series capacitor and the value is much more relevant. Or maybe the other missing components are the problem.

Hi,

If you think the problem is with the Windows Driver Stack not recognizing the card properly then you can try to check how OSes see the card:

1. On newer Windows versions where Powershell is available, you can use the "Get-PnpDevice" command to check how the device is recognized.
2. A system with a Radeon HD3850 AGP is most likely capable to boot Linux from a pen drive then you can run the command "lspci -v" to get how the system sees the card. (Unlike what its name suggest it should also lists PCIe and AGP cards)
3. If you have an older version of Windows like Windows XP you can use the "wmic" command to list devices, but I prefer to do this check with newer versions of Windows that has Powershell or Linux as the wmic command CLI interface is not very intuitive in my opinion.

The card is recognized correctly in apps like HW Info or GPU-Z. Think I need a fresh install of Windows XP to rule out the driver issue.

Found a cheap PCI-E version of the card and got it just to measure the components I'm missing. Values updated in the first post