I'd start by cleaning the card, if thes silver past then it could be shorting out something
Also post of photo of it artifacting

Jade Falcon wrote:

I'd start by cleaning the card, if thes silver past then it could be shorting out something
Also post of photo of it artifacting

Artifacts in DOS is a very bad sign. If anyone has actually brought a card back from this state (and it wasn't a simple cooling issue), I'd love to know. Any cards I've come across that did this were pretty much given up for dead.

Reputator wrote:

Artifacts in DOS is a very bad sign. If anyone has actually brought a card back from this state (and it wasn't a simple cooling issue), I'd love to know. Any cards I've come across that did this were pretty much given up for dead.

That's not DOS. That's my Dimension 4600s IDE self configuration screen that runs at every start up. It's part of the system BIOs.

Ok. I tried to clean that stuff off with hot water and a paper towel but I wasn't able to get the stuff off the PCB and there's still a slight silver tint to the memory chips from there being a slim layer stuck ontop I can't get off.

Any idea how to get this shit off the PCB without removing SMD components the unpleasant way?

TheAbandonwareGuy wrote:

Ok. I tried to clean that stuff off with hot water and a paper towel but I wasn't able to get the stuff off the PCB and there's still a slight silver tint to the memory chips from there being a slim layer stuck ontop I can't get off.

Any idea how to get this shit off the PCB without removing SMD components the unpleasant way?

If it's like AS, I never figured out an ideal way and I gotten to use the white (and way easier to remove) TIM way early in my retro computer hobby (it wasn't even called retro then). Paper towels would kinda remove lots of it and smear out the rest over a larger area, my god what nasty stuff that is 😵

I had also looked at the pics last night, but was too tired to post a reply but I didn't see anything out of the ordinary when it comes to visible damages. But I could of course have missed something.

Get all the silver paste off the card before powering it on again. If its silver or AC5 or something it could be shorting out stuff and cause more problems.
Try a wet nap or q tip dipped in alcohol.

To optimally remove thermal paste is to use citrus based solutions. Here's are some examples:
http://www.ebay.com/itm/Akasa-AK-TC-125ml-Bot … %3D182128943811
http://www.ebay.com/itm/New-Arctic-Silver-ACN … %3D112479693962

These type of cards came originally with silicon based thermal pads not silver paste for the memory modules. As seen here http://www.ninjalane.com/reviews/video/ultra1 … 00xp/page2.aspx
As Jade Falcon mentioned the silver based thermal paste can be capacitive or conductive. See details on one such manufacturer of thermal paste http://www.arcticsilver.com/as5.htm
Before putting it back together I'd get proper thermal pads like these:
http://www.ebay.com/itm/10pcs-Laird-T-FLEX-74 … FoAAOSwzaJX5VGu

The artifacts the card exhibits might be the cause of over heating as the paste wasn't providing the same level of contact as the pads would in which case only BGA rework would fix that.

Good luck,
G1nX

TheAbandonwareGuy wrote:

Jade Falcon wrote:

I'd start by cleaning the card, if thes silver past then it could be shorting out something
Also post of photo of it artifacting

snip photo

OK so the photo was not loading on my end until now. im 90% sure that's a vram problem.
could be a bad cap, short from the paste or a bad vram chip. Missing cap or resistor can do it too.

TheAbandonwareGuy wrote:

Edit: Could a mod renamed this to "BFG Technologies FX5900 Ultra Artifacting"? […]
Show full quote

Edit: Could a mod renamed this to "BFG Technologies FX5900 Ultra Artifacting"?

High-res pictures of the board: https://m.imgur.com/gallery/NmScv

I've finally got around to working on my BFG FX5900 I bought a while back. The AGP contacts were dirty but I was able to clean them with a pencil eraser. The card starts artifacting as soon as it's powered on. I disassembled the heatsink (which was already missing screws so I'm guessing this isn't this cards first rodeo) and there's some aftermarket thermal paste that I've never seen before (extremely sticky, it looks like silver powder mixed with a light adhesive) and the area around the memory chips (I'm guessing it's a memory issue) had some on the traces so I'm thinking either the board is damaged or the traces are being messed up by the paste. I cleaned up the paste and there's no obvious damage but if it's artifacting I must be missing something.

Could you guys help me out? I really need a second pair of eyes.

As was suggested clean all of the TIM off the chips and board. Even UNDER the MEM chips with lots of Denatured Alcohol. That was the wrong kind of TIM to use. Use either Arctic Silver Ceranique or silicon pads.

Jade Falcon wrote:

TheAbandonwareGuy wrote:

Jade Falcon wrote:

I'd start by cleaning the card, if thes silver past then it could be shorting out something
Also post of photo of it artifacting

snip photo

OK so the photo was not loading on my end until now. im 90% sure that's a vram problem.
could be a bad cap, short from the paste or a bad vram chip. Missing cap or resistor can do it too.

Yeah I'm pretty sure the issue lies in the memory subsystem as well. I'm just hoping its the paste messing up a trace and not a ram chip completely blown out. I think I need to put this on hold until I get some rubbing alchohol and some Arcticlean. I'm afraid without it I'll end up damaging the board in my attempts to fix it.

Never ever ever ever got a card working again when it had such failures unfortunately.... Very common problem with some memory chip failing.

Never had any problem with AS shorting anything and i can't see how it is even possible as this stuff is NOT conductive.

meljor wrote:

Never had any problem with AS shorting anything and i can't see how it is even possible as this stuff is NOT conductive.

Of course it is. Why do you think silver paste is not conductive..?

derSammler wrote:

meljor wrote:

Never had any problem with AS shorting anything and i can't see how it is even possible as this stuff is NOT conductive.

Of course it is. Why do you think silver paste is not conductive..?

They say so on their website. Also never had a problem and back in my day i spilled a lot of that stuff and had it on cpu pins etc. I'm a fair bit better with it now 🤣

It's definitely conductive.

Yes, AS silver or copper TIM's can be conductive. However its been shown that only happens in rare cases where high pressure loads are used between the surfaces. I can't remember exact lbs mentioned back in the early 2000's, but it was more than 50lbs of pressure between conductive surfaces before conductivity was measurable. Other than industrial equipment, why would anybody use that much mounting pressure on a heatsink for PC parts?

Now due to the mix of oils used in AS solve and copper TIM's, they can act as a multi layer dielectric and become capacitive. This can be a problem with sloppy application or using it in extremely tight spaces.

Either way, why does anybody even use AS "grease type" TIM's in this day and age? There are so many superior and cheaper solutions on the market that I can't understand this. AS makes fantastic thermal epoxy for electronics, their ceramique is ideal for extreme cooling using LN2 or DI. Their ArctiClean is truly fantastic and a must for prep work when using liquid metal TIM's. The old AS silver and copper TIM's are out of date, require a curing time of quite a few thermal cycles, more difficult to clean, and offer minimum cooling benefits compared to contemporary brands. All this and at premium price for a minimum amount per unit.

My opinion, but a 20g tube of MX4 for about $20.00 or less is well worth the price compared to AS5 3.5g at an average $9.00.

Lets not all forget that this might not be AC5, but some other silver paste. It does seem kind if thin and wet if its old AC5. AC5 tends to dry up and crack as it ages in my experience. Infact it looks more like Mx-2, but darker, however that could just be the lighting.

Jade Falcon wrote:

Lets not all forget that this might not be AC5, but some other silver paste. It does seem kind if thin and wet if its old AC5. AC5 tends to dry up and crack as it ages in my experience. Infact it looks more like Mx-2, but darker, however that could just be the lighting.

When I saw the pics for the first time, my thought was "What thick layers of TIM on the memory chips!".
It actually reminded me a bit of the silvery stuff Zalman used to use, which was in a tiny jar with a veeery small brush. I never liked how thick it made the layers after application. MX-2 is definitely white, I've used MX-2 for years. It doesn't seem to spread over the edges in that way, seeing there was too much applied on those memory chips.
Btw, I didn't really take heed to the one memory chip which seems to be partially uncovered by the TIM. Perhaps that's where the problem area lies??

If it's just a single memory chip that's at fault, would flashing the card into using only half the memory perhaps fix the issue? If the card is dead as it is, imo it's better to at least have a halved functional card.

Yeah Mx-2 is more of a grey/white. This looks more like ether that super zalman stuff or shin etsu, but it could be just about anything.
Also flashing the card to a 128 or 64mb version may work if the problems is with VRAM.