Guest @ 3.135.217.228
Register

Donation

 

Donate Towards BK`s Servers

Please Donate to Help Keep BKs game server up and running Donate

Server

 
Recent Topics Prev 7   Next 7
Forum
Author
Replies
Last Post
4
Sun Apr 21, 2024 9:49 pm
Snap71 View latest post
0
Fri Apr 19, 2024 12:54 pm
MaxwellSmart View latest post
1
Sat Apr 13, 2024 2:43 pm
Natb1 View latest post
3
Thu Apr 11, 2024 1:10 am
Snap71 View latest post
3
Mon Apr 08, 2024 12:51 am
Luke View latest post
6
Wed Apr 03, 2024 4:42 pm
Natb1 View latest post
5
Sat Mar 30, 2024 7:33 pm
Luke View latest post

GPU over heating, shutting the whole system down

Joined: Apr 20, 2009
Posts: 1384
Location: Columbia, MD USA
Posted: Thu Jul 21, 2011 7:44 pm
Hey guys,

Before I left for my trip, my system was really starting to lag and then the day I left the system would just shut down after 10 mins or so.

When I got back, figured it out was the GPU overheating.  It is an ATI 4890.  I opened up the case, put a box fan right next to it and turned it on high.  I can run the computer fine with temps in the low 70s (C), but when I turn on BlackOps with all graphic settings on low, it is running at 100-115 (C).  I saw an extra cooling fan on amazon that takes up one of your slots right next to the GPU for $15 I can get.

What do you think?  Bad card?  Will the fan do the trick?  Any ideas, other solutions are welcome.  Thanks!

PastorDan




Joined: Nov 17, 2010
Posts: 1914
Location: PA
Posted: Thu Jul 21, 2011 8:25 pm
PastorDan,

The 4980 was built to run on the hot side, but I think the 100-115C is over that limit!  I know AMD (ATI) rates the card under full load to normally run between 80-95C, and that is what they warranty.  1st thing is have you gotten the dust bunnies out of the Graphics Card.   Built up dust on the fan will slow it down so that will increase heat.  Second a lot of the ATI and nVidia Cards you can buy third hand replacement fans for the graphics cards which work better than the factory stock fans.  You are not that much hotter than recommended temps, so cleaning your existing fan, putting on a replacement fan, or getting a fan card to pop in next to the graphics card, should do the trick for you.

AFTER MARKET FAN for 4890: http://www.zalman.com/eng/product/Product_Read.asp?idx=338





Joined: Apr 03, 2010
Posts: 2971
Location: Loveland, Colorado
Posted: Thu Jul 21, 2011 8:27 pm
before you start spending big money, open up you gfxs card.  Remove the heat sync, clean it all up.  Remove all the thermal paste from the card and the GPU and put some new thermal paste on.  You could only need about half the size of a pea either on the GPU or the heat sync.  It may sound weird but the thermal paste is VERY important and if it gets to hot it can get hard and stop working.






Joined: Jul 26, 2007
Posts: 9890
Location: Grand Rapids, MI
Posted: Thu Jul 21, 2011 8:30 pm
I had that happen to me back when I had a desktop. I just pulled the heat sync and cleaned it. Then I went to best buy and got some new heat sync past and put some on.

Hope that helps.
GOA.Luke*BK*



Joined: May 31, 2011
Posts: 2631
Location: Fort McMurray, AB
Posted: Thu Jul 21, 2011 10:56 pm
Yepp. A good cleaning and some thermal paste will do it well. My case stays quite clean, but my past few video cards  have always been the first components to get clogged up, as they used the impeller-style fan, as the  4890's do, If memory serves.

Of course, make sure the fan isn't dead. If so, then you'll have to get another cooler. If your case tends to get toasty, I'd suggest one that pushes the warm air out the back. I had an Arctic Cooling VGA cooler on my 9800 a while ago. It was fairly cheap, and performed very well.


-- Of all the soles I've ever tasted, his had the most   ....cumin. --
Joined: Jun 21, 2009
Posts: 343
Location: Warshington (state)
Posted: Fri Jul 22, 2011 1:43 am
Also check the fan speed under normal and heavy load. If you don't hear it noticeably increase in speed going to heavy load, then you may have a bad fan. One way I found to test the fan is to touch it just enough to stop it and then let it go. It should take right off quickly and be back up to normal speed almost instantly. If the bearings are going bad, it will slowly speed back up and you can usually hear a small amount of grinding (friction) as it spins back up.

Joined: Sep 01, 2009
Posts: 2881
Posted: Fri Jul 22, 2011 12:37 pm

Silent_Wolf wrote (View Post):

before you start spending big money, open up you gfxs card.  Remove the heat sync, clean it all up.  Remove all the thermal paste from the card and the GPU and put some new thermal paste on.  You could only need about half the size of a pea either on the GPU or the heat sync.  It may sound weird but the thermal paste is VERY important and if it gets to hot it can get hard and stop working.


Thats where I would start, and since your in there do the same for the CPU.  Make sure to get all the paste off, q-tips and rubbing alcohol works great and drys fast.  If you can get it use Arctic Silver 5, IMO its the best stuff out there.






Joined: Apr 20, 2009
Posts: 1384
Location: Columbia, MD USA
Posted: Fri Jul 22, 2011 8:12 pm
I took the GPU out and it was a mess with dust and gunk!  No wonder it didn't work.  I went head and took it apart and cleaned each piece.  Bought thermal compound and after cleaning the old paste off and cleaned the surfaces off.  Applied the new paste, put it all back together, so keeping my fingers crossed!




Joined: Apr 03, 2010
Posts: 2971
Location: Loveland, Colorado
Posted: Fri Jul 22, 2011 9:01 pm
Come on now, dont leave us hanging.........Did it work?






Joined: Apr 20, 2009
Posts: 1384
Location: Columbia, MD USA
Posted: Sat Jul 23, 2011 3:51 am
Nope.  Still overheating...

I can play, but I have to open up the side panel and put a box fan full blast.




Joined: Jul 12, 2007
Posts: 13654
Posted: Sat Jul 23, 2011 12:06 pm
now this may seem like a stupid question but does the fan still work for one and also doe the fan speed up when you start the game ect if you fan is not spinning fast enough it will not be pulling the heat way



It is Going To Hurt
http://taylormadehosting.co.uk
Joined: Jun 21, 2009
Posts: 343
Location: Warshington (state)
Posted: Sat Jul 23, 2011 3:41 pm
Pastor Dan, I re-read your symptoms and sorry for missing it in the first posting. If your system is completely shutting down and not doing a video lock up, then it is most likely not the GPU. When a GPU overheats, you get anomalies like artifacts first and then your whole video will lock up. I have not yet seen a video card shut a system down as it does not have the ability to tell the motherboard it is overheating. The MB has sensors in various spots on the primary chips like CPU, north-bridge, etc that tell it what the temps are. When those temps hit a threshold, it shuts down to protect itself.

First step is to remove the heat sink from the CPU and do the same thing you did with your GPU. Clean fan, replace old paste, etc. You can also remove the heat-sinks from the other chips and do the same thing. Basically a complete cleaning and paste replacement on all heat-sinks.

If this doesn't work to resolve the issue, then quite honestly the only thing remaining that I can think of that would make a system do a shutdown is the power supply. If your power supply is not functioning properly under load, then it just won't have the power to run your system and the power supply will shutdown.

So since you already have the paste, I would suggest the cleaning of the main chips and see what happens. If it still happens, then power supply replacement may be necessary.

Also, just a suggestion. Get into the habit of blowing the PC out every 2-3 months to get rid of dust bunnies and at least once a year do a complete tear down and cleaning. Replacing the old paste and any worn out fans. This will help keep it in optimal gaming condition.

(which reminds me... I need to clean mine... hope this helps)

Joined: Apr 20, 2009
Posts: 1384
Location: Columbia, MD USA
Posted: Sat Jul 23, 2011 8:42 pm
Bolerro...thanks for the suggestions, I will give them a try today...It can't hurt me!




Joined: Jul 13, 2009
Posts: 445
Location: Medicine Hat, Alberta
Posted: Sat Jul 23, 2011 10:43 pm

Bolerro wrote (View Post):

Pastor Dan, I re-read your symptoms and sorry for missing it in the first posting. If your system is completely shutting down and not doing a video lock up, then it is most likely not the GPU. When a GPU overheats, you get anomalies like artifacts first and then your whole video will lock up. I have not yet seen a video card shut a system down as it does not have the ability to tell the motherboard it is overheating. The MB has sensors in various spots on the primary chips like CPU, north-bridge, etc that tell it what the temps are. When those temps hit a threshold, it shuts down to protect itself.

First step is to remove the heat sink from the CPU and do the same thing you did with your GPU. Clean fan, replace old paste, etc. You can also remove the heat-sinks from the other chips and do the same thing. Basically a complete cleaning and paste replacement on all heat-sinks.

If this doesn't work to resolve the issue, then quite honestly the only thing remaining that I can think of that would make a system do a shutdown is the power supply. If your power supply is not functioning properly under load, then it just won't have the power to run your system and the power supply will shutdown.

So since you already have the paste, I would suggest the cleaning of the main chips and see what happens. If it still happens, then power supply replacement may be necessary.

Also, just a suggestion. Get into the habit of blowing the PC out every 2-3 months to get rid of dust bunnies and at least once a year do a complete tear down and cleaning. Replacing the old paste and any worn out fans. This will help keep it in optimal gaming condition.

(which reminds me... I need to clean mine... hope this helps)


Actually a GPU can shut down a system, just depends on the errors.  I had a GPU that would always crash my system, but powersupply sounds more like it.






Joined: Mar 29, 2009
Posts: 54
Location: Massachusetts, USA
Posted: Sun Jul 24, 2011 2:16 pm
Yup, PS was what I was thinking as well.  I've had several go bad in just that way.  Normal usage, web browsing OK.  Gaming, shutdown after an ever decreasing number of minutes.  It's the current draw on the PS when you have all the bells and whistles going.  I try to start out with a PS that's 25-40% bigger than what I need as components degrade over time.  That way the supply lasts a bit longer for not too much more money.

All times are UTC [DST enabled]
Forums ©