damienpayne wrote on 2022-10-06, 07:52:

How do I identify witch specific chip is the culprit?
I'd rather not remove all eight of them.

I always find these questions funny...
Pinpointing/isolating the actual faulty memory IC is not something that can be taught in a "crash course forum post". 😀
In a way, this is like asking a surgeon to teach you how to diagnose a brain related illness (and, why not, also perform the brain surgery itself) during a night out drinking. 😀
Bottom line, this is not an exact science, there are MANY ways in which a memory chip can fail. Some are certainly easier to diagnose than others, but you still require extensive electronics knowledge, proper equipment (like an oscilloscope) and last but not least: a GREAT deal of experience.

I'm only an electronics hobbyist (so very far from the person I described above), which is why my repairs are mostly based on trial and error (and, even so, most of the time they are successful - though they do take longer).

So, at best, I can give you a few suggestions that will make the trial and error repairing attempt a bit easier:
- are those artifacts present during POST/in 2D mode? If yes, then you need to focus on the four memory ICs that belong to the primary VSA-100 chip (that would be the chip on the left, closest to the bracket).
- are the artifacts only occurring when trying to run a game and it otherwise works great in 2D mode? Then you can force single chip mode from the driver control panel and see if that fixes the issue. If it does, then you need to focus on the four memory ICs that belong to the secondary VSA-100 chip (the one on the right).

Once you've narrowed things down to a certain group of four memory ICs, then you can apply the very scientific "eeny, meeny, miny, moe" concept, and replace one IC at a time. If you're a lucky guy, the very first one you replace will be the faulty one. If you're unlucky, well... you know. 😁
Good luck!

What bloodem said.

But also, as a really quick/easy step to diagnose - have you tried lowering the memory clock and/or increased cooling on the memory chips to see if this is simply the hardware operating close to its limit?

bloodem wrote on 2022-10-06, 08:43:

I always find these questions funny... Pinpointing/isolating the actual faulty memory IC is not something that can be taught in […]
Show full quote

damienpayne wrote on 2022-10-06, 07:52:

How do I identify witch specific chip is the culprit?
I'd rather not remove all eight of them.

I always find these questions funny...
Pinpointing/isolating the actual faulty memory IC is not something that can be taught in a "crash course forum post". 😀
In a way, this is like asking a surgeon to teach you how to diagnose a brain related illness (and, why not, also perform the brain surgery itself) during a night out drinking. 😀
Bottom line, this is not an exact science, there are MANY ways in which a memory chip can fail. Some are certainly easier to diagnose than others, but you still require extensive electronics knowledge, proper equipment (like an oscilloscope) and last but not least: a GREAT deal of experience.

I'm only an electronics hobbyist (so very far from the person I described above), which is why my repairs are mostly based on trial and error (and, even so, most of the time they are successful - though they do take longer).

So, at best, I can give you a few suggestions that will make the trial and error repairing attempt a bit easier:
- are those artifacts present during POST/in 2D mode? If yes, then you need to focus on the four memory ICs that belong to the primary VSA-100 chip (that would be the chip on the left, closest to the bracket).
- are the artifacts only occurring when trying to run a game and it otherwise works great in 2D mode? Then you can force single chip mode from the driver control panel and see if that fixes the issue. If it does, then you need to focus on the four memory ICs that belong to the secondary VSA-100 chip (the one on the right).

Once you've narrowed things down to a certain group of four memory ICs, then you can apply the very scientific "eeny, meeny, miny, moe" concept, and replace one IC at a time. If you're a lucky guy, the very first one you replace will be the faulty one. If you're unlucky, well... you know. 😁
Good luck!

In my ignorance I kind of hoped for some kind of "follow these 5 easy steps". 😉
The artefacts are present from POST, i have not bothered to try any acclerated software since the card is "broken". Did not want to put any stress on my damaged baby.

A huge thank you for your reply! got me a lot wiser to this dilemma.

leonardo wrote on 2022-10-06, 09:03:

What bloodem said.

But also, as a really quick/easy step to diagnose - have you tried lowering the memory clock and/or increased cooling on the memory chips to see if this is simply the hardware operating close to its limit?

No, I have not tried that. Did not bother to keep on using the card in a damaged state.
Thanks for helping! 😀