My GPU's cooling is losing efficiency

[Kelesis]

Greetings.

It's my first post here.
I'm not sure whether my issue best fits this forums rather than the GPUs forum. Anyway...

My current graphics card is exactly a Zotac GeForce 9800 GTX. I know there's much better HW out there - I'm just attached to this one and want to preserve it.

Here are some pictures of the product:
http://www.google.it/search?hl=it&s...0.183.2553.15j9.24.0...0.0...1c.1.uHNY-0GL1O8

After cooking the card *prior* to this one, I developed a small obsession for temperatures. So I've been monitoring the temperatures of my 9800 since day zero (Feb 19th 2009).

Once a year I'll open the case and make a complete removal of dust and bunnies. While at it I also clean the graphics card. And everytime I do that I notice a small decrease in GPU cooling efficiency. Now I estimate that given the same Game, O.S., display drivers, period of the year and climatic conditions, my GPU is 7 to 12 C° hotter than it was when brand new. Both in idle and duty conditions.

Notice that I never removed the heatsinker nor disassembled the fan. It's all original. I limited myself to blow air into the heatsinker and gently brushed the fan blades with a soft brush. I'm unable to tell whether dust is stuck deep into the heatsinker (blocking the airflow) or it's the fan slowly wearing off, even though I never had it spin past 60% power (it is 35% power when in idle).

I'm thinking to substitute this fan with a new one. And of course I'll have to remove the heatsinker (and will clean it for good while at it).
This scares me. I've never disassembled one such heatsinker and am unaware of pitfalls and problems. As for the fan, I don't know *if* I can find a suitable replacement for it (I don't even know the current fan's specs!)...

Any guidance or links you can give me is appreciated.

 

[EarthDog]

Well, the heatsink itself doesnt lose efficiency. Its still the same chunk of metal. That said, its possible that the TIM is past its useful life. I would try putting on some MX-2 or MX-4 (remove the old of course) and see if that helps.

That said, You have to be sure that EVERYTHING is the same... ambient temperatures, clean the intake fans on the case, etc. For every degree C difference in ambient it translates to a 1:1 on your temps.

As far as the fan, who knows... did you monitor the RPMs? It shouldnt get less efficient either as its still the same fan spinning at the same speed (assumed, % doesnt matter really).

 

[Anonaru]

Also consider securing the fan (So you don't damage it) and pushing a moderate amount of air pressure through your card-- It very well could be some well-stuck dust and gunk, you've had the card for quite a time now.

 

[doyll]

I like the idea of taking it all apart, doing a good cleaning of everything including every fin in the heatsink and re-assemble with new TIM. Isopropyl alcohol is safe to use on most everything and will dissolve oily film if there is any on fan or cooler fins. It's amazing what new TIM with clean fins and fan blades can do.

 

[S_I_N]

yup sounds like retimming time and as suggested make sure the blades are free of dust and by all means hold it from spinning when you blow air on it as that can assist in killing a fan.

 

[EarthDog]

How can blowing on a fan kill it?

 

[Bobnova]

If you can clean it without taking the cooler off the card, that's what I'd do.

Unless you have an air compressor you're unlikely to be able to move enough air through the fan to kill it. Canned air and trying very hard to spin the fan maybe.
To kill a fan you need to turn it (much) fast than it's rated for, given the GPU fans are typically in the 3000-4000rpm range.
Now if you are using an air compressor, don't overspin the fan!

 

[S_I_N]

I was more referring to the age and yeah to fast or in oppisite direction of intended can slowly (or quickly) ruin the bearings.

 

[wagex]

bearings arent directional.

 

[S_I_N]

no they are not but they are used to rolling one way for so long (in this case 3-4yrs) rolling them to long or to fast in the opposite direction MAY start the degrading process. I'm not new to PC's in any way shape or form I also have common sense and common sense should tell one not to over spin a fan in the opposite direction for to long is all I was warning about. Not trying to debate life span/cycles of bearings.

 

[wagex]

im not trying to debate either just didnt want some one worrying about something that isnt going to hurt his pc.

 

[doyll]

wagex is correct.

Over speeding with air and/or to much front/back/front thrust on keeper/clip whatever yes. but not spinning opposite direction on bearing surface.

 

[S_I_N]

tell ya what y'all keep spinning those bearings anyway you want I will always hold the blade from spinning when I blow out my fans. And we'll see who posts about fans making funny noises first I'm just suggesting its not a good idea to spin them opposite from intended or faster than intended and that the better practice is to try to not allow them to spin as mentioned. Will it kill the fan and or bearing most likely not but it MAY cause issues with said fan.

 

[Bobnova]

no they are not but they are used to rolling one way for so long (in this case 3-4yrs) rolling them to long or to fast in the opposite direction MAY start the degrading process. I'm not new to PC's in any way shape or form I also have common sense and common sense should tell one not to over spin a fan in the opposite direction for to long is all I was warning about. Not trying to debate life span/cycles of bearings.

Do you have any papers or links or anything to back that up? Logically it doesn't make much sense to me.

 

[doyll]

Please don't try and confuse us with facts. Our minds are already made up.

I think the bottom line is be careful about spinning fans with compressed air or you can damage them.

 

[S_I_N]

not really but I'm sure if one was inclined to they could google something up. Its purely as I said common sense. Here an example albeit not apples to apples/ Let say your vehicles tires would you run them on one side of the car for 3-4years (roughly 36,000-48,000 miles) then swap them to the other side??? no you wouldnt cause that would create a different friction pattern on them there by causing the tread to tear off. Now if you rotated said tires every 6months as recommended then that friction pattern is negated. Bearings on said wheels can not and should never be swapped side to side as that wear pattern is burned in so to speak. Whilst fan bearing are vastly different the same rules should apply. And while they are not designed to be ran one way only when that process starts they create a wear pattern due to the friction created as they spin. Yes they are lubricated but that does not stop wear patterns from forming. Now running them the opposite direction CAN or MAY cause a new pattern to evolve thereby when rotated back the origional direction tearing CAN or MAY occur. I'm not saying it will and by default given there isnt an overly heavy load on fan bearings its still a possibility I'd rather avoid. Seeing as I paid 18$ a pop for my fans 13 (actually only paid for 9 the other 4 were a FB prize from Bitfenix)of them is a decent investment that I'd rather not have to re-spend if not necessary.


sorry doyll was typing this prior to your post

 

[Bobnova]

I talked to a tire shop about the tire thing actually, I was curious what their rotation pattern was as I'd heard that don't swap sides thing quite often but never seen any backup (24 year auto mechanic here). It doesn't matter, their standard pattern for rotating the tires changes the direction of rotation. Never had any issues. Ever.

By the CAN OR MAY logic, we shouldn't clean the fans, as the fan CAN OR MAY have built its wear patterns based on dust weight on the blades.
Would you be curious to know how many long-been-in-production fans I've spun every which way with compressed air? Dozens. Guess how many have died, or had bearing issues shortly thereafter? None.

 

[doyll]

Here in UK nobody rotates tires. The consensus is doing so reduces traction because tires wear into tracking and braking pattern.

USA is rotate, rotate, rotate. Some front to back in same rotation, some go diagonal and reverse rotation.

I know on my 4x4 I rotate and don't worry about direction they turn to keep them all wearing even.

Racing is each tire has it's one position, pressure too.

 

[doyll]

I know Bobnova, but I've killed a couple with air.

Spun them just to see how fast they would go before they blew apart. It was amazing how fast they went!

 

[S_I_N]

well this is what makes our country what it is. We all have opinions and they are not all wrong. But also being a mechanic since the early 80's I do know a little about bearings and as I said before its not really apples to apples with those compared to case fans.

but as we have said before to the OP retim that sucker and see what happens Clean the fan how you see fit.