My nvidia 8600GT (PCI-E x16) keeps overheating and quitting the PC game (Portal 2) on me. The hardware looks good (good PSU, memory, no blown caps., etc) and software is simple and clean. The game crash happens randomly but I can always hear the fan inside the case started spinning faster (louder) right before the game quits.

The case has one 80mm fan at the front panel sucking in air into the case. The PSU has one fan blow the air out of the case. There is also a 120mm rear fan right under the PSU pushing out air.

The case photo: http://img507.imageshack.us/img507/1821/insidedi4.jpg
(Dynex case/power I got from the bestbuy for $20 a couple of years ago)

I think the problem is that all the cool air from the front panel is directed though the upper region of the board because of the PSU and the rear fan. The extension slot area (the lower part of the board) doesn't get any air flow so once the video card's temp rises from gaming it keeps building up the temperature and eventually quits working (the little GPU fan doesn't do much except moving hot air around).

I thought about getting one of those case slot fan
http://i.ebayimg.com/00/$(KGrHqUOKj8E19L36O,NBNpWwPHrK!~~0_3.JPG

but from several forums people don't really notice significant improvement from those fans.

Lower ambient temperature.

Yeah, my sister has the same case it feels pretty hot, even at idle. Granted, she likes to keep her house at uncomfortably hot temp.

(gigabyte 785G, unlocked amd 5000+, Seasonic PSU, 7200 RPM Samsung HDD)

That card has a fan on it, yes? Does it operate properly?
With the side off the case, aim a fan directly at the card and see if the problem persists. If so the problem probably is heat related.
The problem you'll run into with adding another fan is three fans will be pushing air out, and only one will be pulling air in.
I think the ideal thing to do (short of buying a new case), would be to cut a hole in the side cover near the video card and mount a fan blowing in.

Is the power supply hot? Touch it after it shuts down. The top of the case above the power supply should not be much hotter than ambient under load. If it's hot, it's struggling either because it's fan is failing or it's overloaded.

How are the temps with the side panel off and a fan blowing in?

You would really benefit from some low RPM fresh air blowing in through the side of the case or getting sucked into a side-vent below the video card AND/OR opening up 1 or 2 PCI slot covers at the bottom-rear and see if enough fresh air gets sucked in.

Noway to tell what kind of air flow you have without part #'s from the fans you have now.

Your cable management is excellent.
In the Nvidia control panel you can regulate the fan. Turn it up to 100% all the time.

make sure that you have an even air flow through the case and not a negative one. a negative air flow would be like a higher number of fans pulling in air and a lower number of fans pushing out air.

Video cards run much hotter than other components. 77ºC should be an acceptable temperature for your card.
Please tell us your temperatures, your resolution and and at what level your settings are for the game.

That PSU is junk. Find out the temps of all the components. If the CPU and chipset is below 60C and the video card is below 90C, then you should replace the PSU.

That PSU is junk. Find out the temps of all the components. If the CPU and chipset is below 60C and the video card is below 90C, then you should replace the PSU.

Seconded.

You don't need more fans, but those three fans in the upper left may be fighting one another. Try reversing the upper case fan and maybe the lower right fan as well. Moving the lower hard drive up or down one bay may also help.

You don't need more fans, but those three fans in the upper left may be fighting one another. Try reversing the upper case fan and maybe the lower right fan as well. Moving the lower hard drive up or down one bay may also help.
The configuration of the fans is perfect. You do not want 2 fans right next to each other feeding each other. That will prevent them from doing their job. Believe me, I tested this. If a fan blowing cool are in is right next to an exhaust, the exhaust just sucks the cool air back out instead of pulling warm air out.

The configuration of the fans is perfect. You do not want 2 fans right next to each other feeding each other. That will prevent them from doing their job. Believe me, I tested this. If a fan blowing cool are in is right next to an exhaust, the exhaust just sucks the cool air back out instead of pulling warm air out.
There is no such thing as a perfect configuration of fans, even if a fan blowing cool are in is right next to an exhaust. :lol:

If fans working in series is always bad, why are jet engines built that way, and why have auto makers sometimes recommended adding such a fan when converting old cars from R-12 Freon to R-134a?

Please post your test results and methodology or cite references to Mike Chin's work. :lol:

There is no such thing as a perfect configuration of fans, even if a fan blowing cool are in is right next to an exhaust. :lol:

If fans working in series is always bad, why are jet engines built that way, and why have auto makers sometimes recommended adding such a fan when converting old cars from R-12 Freon to R-134a?

Please post your test results and methodology or cite references to Mike Chin's work. :lol:


You've just made a bad mistake, if you are going to bring up automotive engineering at least expand on it to a point where someone who is not a mind reader can get a picture of what you are talking about. Fan location is everything. A incorrectly placed fan will harm more than it helps.
You're barking up the wrong tree son.

AND you're not adding anything useful to the thread.

Same with my dads, he has a 8600 gt in his comp and it idles around 55C and under load it hits 80~100 C in an antec 900 case. Changing thermal paste didn't help and Rivatuner doesn't get the fan to spin 100%. But the thing is, my gpu doesn't exit the game even when it hits 90 C. Are you sure you aren't overclocking the card or changing the voltage.

Also I'm with them about the psu, those psus that come with cases sucks. You have 1 intake fan 80 mm, but two exhaust? That may be a problem as well.

Same with my dads, he has a 8600 gt in his comp and it idles around 55C and under load it hits 80~100 C in an antec 900 case. Changing thermal paste didn't help and Rivatuner doesn't get the fan to spin 100%. But the thing is, my gpu doesn't exit the game even when it hits 90 C. Are you sure you aren't overclocking the card or changing the voltage.

Also I'm with them about the psu, those psus that come with cases sucks. You have 1 intake fan 80 mm, but two exhaust? That may be a problem as well.

Try MSI Afterburner then use the "Unofficial" mode to adjust fan speed.
The steps are here: http://forums.guru3d.com/showthread.php?t=338906.
XFX said I could not adjust the fan on my card because it had 2 wires. Stock Afterburner didnt work either, but with unofficial mode it worked flawlessly. :nod:

i am the OP and I appreciate all the comments. Further test suggested that the GPU Temp never exceeds 77C before the game quits. I am ordering a new PS ( #N82E16817341017 on newegg) and see if it solves the problem. If PS is not the problem then I will have a backup PS.

The PS came with my Dynex case has the following label:
http://i963.photobucket.com/albums/ae118/lancey2010/Computers%20-%20Hardware/PICT0007-1.jpg

and I was wondering if the Amp rating on the 12v rail is just to low for 8600GT.

I will try MSI afterburner and see if it changes anything in terms of heat issue but I've tried using a desktop fan blowing directly to the board with the case open and the game still quits, which was another reason that I think it's not overheating problem here.

There is no such thing as a perfect configuration of fans, even if a fan blowing cool are in is right next to an exhaust.

If fans working in series is always bad, why are jet engines built that way, and why have auto makers sometimes recommended adding such a fan when converting old cars from R-12 Freon to R-134a?

Please post your test results and methodology or cite references to Mike Chin's work. :lol:
You've just made a bad mistake, if you are going to bring up automotive engineering at least expand on it to a point where someone who is not a mind reader can get a picture of what you are talking about. Fan location is everything. A incorrectly placed fan will harm more than it helps.
You're barking up the wrong tree son.

AND you're not adding anything useful to the thread.

I never said fan location wasn't important, but sometimes reversing the air flow of one fan can help.

I didn't realize anything more had to be said about the automotive A/C modification because most people have seen at least one fan behind their car's radiator and probably don't need a diagram to get an idea of what it would be like to add another fan to the front of the radiator. But to elaborate, when R-134a refrigerant was introduced in the 1990s, auto makers put out service bulletins about modifying existing A/C systems that were based on Freon R-12 so they'd work well with R-134a (the EPA's website used to have summaries of those bulletins), and many modifications involved improving the air flow through the condenser, including by adding a fan to the front of the condenser to help the existing fan behind the radiator:

Honda A/C pusher fan mod (http://www.honda-tech.com/showthread.php?t=2334672)

Mazda RX7 A/C pusher fan mod (http://www.rx7club.com/showthread.php?t=663162)

Ford truck A/C pusher fan mod (http://www.ford-trucks.com/forums/1065301-what-is-the-coldest-ac-temp-acheiveable-3.html)

Why did professional engineers think that adding a fan to feed the existing fan, that is, as you previously stated, "fan blowing cool are in is right next to an exhaust," was a good idea for the automotive A/C if it was actually a bad idea?

You seem to miss 24 and want to reenact one of the bureaucratic showdowns at CTU headquarters. :lol:

i am the OP and I appreciate all the comments. Further test suggested that the GPU Temp never exceeds 77C before the game quits. I am ordering a new PS ( #N82E16817341017 on newegg) and see if it solves the problem. If PS is not the problem then I will have a backup PS.

The PS came with my Dynex case has the following label:
http://i963.photobucket.com/albums/ae118/lancey2010/Computers%20-%20Hardware/PICT0007-1.jpg

and I was wondering if the Amp rating on the 12v rail is just to low for 8600GT.

I will try MSI afterburner and see if it changes anything in terms of heat issue but I've tried using a desktop fan blowing directly to the board with the case open and the game still quits, which was another reason that I think it's not overheating problem here.

Thats a little overkill for your system. Are you sure you don't want to upgrade the graphics card too since your willing to spend on a new psu.

Thats a little overkill for your system. Are you sure you don't want to upgrade the graphics card too since your willing to spend on a new psu.

Well it turned out that I can't use the promo code for the new PS I'd like to purchase off newegg so I guess I will remove the old video card and just use the onboard vga card and lowest settings for gaming.

The PS came with my Dynex case has the following label:
http://i963.photobucket.com/albums/ae118/lancey2010/Computers%20-%20Hardware/PICT0007-1.jpg

and I was wondering if the Amp rating on the 12v rail is just to low for 8600GT.

Apparently the 8600GT consumes no more than 43W, so 18A @ +12V should be fine for it, and I doubt that it, your CPU, and disks combined need anything close to to the limit. Also it seems that many graphics chips run hot.

OTOH Dynex and RocketFish are made by Huntkey, a company whose products seem to be inconsistent in quality. For example, HardwareSecrets.com found that a 400W model (link (http://www.hardwaresecrets.com/printpage/Dynex-400-W-Power-Supply-Review/557)), internally labelled "200-300W", was able to put out 438W, but a 450W Huntkey could do only 360W and exploded at 450W, despite its circuit board being marked "350-450W" (link (http://www.hardwaresecrets.com/printpage/Huntkey-Green-Star-450-W-LW-6450SG-Power-Supply-Review/527)).

The only ways to tell if your problem is the PSU are to substitute a known high quality unit or measure the voltages at both idle and while the game is running to see if they droop more than about 3-5% at full load.

The common test for overheating is, remove the case side panel, and blast a desk fan into the computer. If performance improves, then the problem was heat related, but it may not necessarily be due to an hot graphics card because cooling also increases power supply capacity, and PC Power & Cooling showed that an old 550W Enermax lost 1/3 of its capacity when the temperature rose from 25C to 40C, while the PCP&C unit lost only 10% capacity. Also low quality capacitors can lose a lot of capacitance in just a few years, and Huntkey seems to use a lot of cheapo Teapo brand. One more thing to consider is the PSU's fan -- is it binding? Don't check this except with the AC power cord unplugged (just turning off the surge protector or power strip isn't enough), and use only a non-metal stick to rotate the blades (chopstick, plastic straw). I've seen some fans get really tight.

you deserve the facepalm of the year award. You are comparing cooling systems that have absolutely nothing in common in their design, function, and enclosure. And the fact that you are babbling on about things you clearly have no idea about besides what you are told is increasingly apparent.
All I said was that fans in series sometimes improve cooling, and I gave some examples. Don't be so upset just because somebody disagreed with you.

And you still haven't shown any test results for fan position & orientation vs. temperatures.

All I said was that fans in series sometimes improve cooling, and I gave some examples. Don't be so upset just because somebody disagreed with you.

And you still haven't shown any test results for fan position & orientation vs. temperatures.
It's not scientific proof so take from it what you will, but it's true(I would know, that's me).

http://www.nzxt.com/forum/showthread.php?8689-HOW-TO-proper-cooling-with-Alpha-Case

It's not scientific proof so take from it what you will, but it's true(I would know, that's me).

http://www.nzxt.com/forum/showthread.php?8689-HOW-TO-proper-cooling-with-Alpha-Case

What sucks about those results:

1) Not all fans perform the same. What fans shipped with his case may not be what shipped from a different vendor, batch, etc.

2) Size, RPM, pressure, CFM and control voltage/signal are the factors, not case model # and installed hardware. You can control the temps in any case with the 'right' fans.

Other than that, he figured out how to get the best temps with what he 'has', by process of elimination.

What sucks about those results:

1) Not all fans perform the same. What fans shipped with his case may not be what shipped from a different vendor, batch, etc.

2) Size, RPM, pressure, CFM and control voltage/signal are the factors, not case model # and installed hardware. You can control the temps in any case with the 'right' fans.

Other than that, he figured out how to get the best temps with what he 'has', by process of elimination.


Here is the photo of my setup (the photo in my original post was from someone else's that I found online):
http://i963.photobucket.com/albums/ae118/lancey2010/Computers%20-%20Hardware/PICT0005-1.jpg

Here are some other photos about the fan and case of my build:
http://s963.photobucket.com/albums/ae118/lancey2010/Computers%20-%20Hardware/

Again I am not sure this is a heat-related issue any more but don't have a way to accurately measure power supply's quality.

The tech. support from EVGA suggested me to put the video card into another system and see if it's indeed the card that causes the game crash. By isolating the problem it may give a better clue. I don't have another system w/ PCI-E x16 slot so that suggestion was not very useful.

Leave the cover off (works great on desktop style units).

Here is the photo of my setup (the photo in my original post was from someone else's that I found online):
http://i963.photobucket.com/albums/ae118/lancey2010/Computers%20-%20Hardware/PICT0005-1.jpg

Here are some other photos about the fan and case of my build:
http://s963.photobucket.com/albums/ae118/lancey2010/Computers%20-%20Hardware/

Again I am not sure this is a heat-related issue any more but don't have a way to accurately measure power supply's quality.

The tech. support from EVGA suggested me to put the video card into another system and see if it's indeed the card that causes the game crash. By isolating the problem it may give a better clue. I don't have another system w/ PCI-E x16 slot so that suggestion was not very useful.
make sure your CPU fan is blowing up, not down on the CPU, and fix those cables in the middle of the case -- they're interfering with the intake at the front.

And learn how to light your subject when you're taking a picture man :tongue:

make sure your CPU fan is blowing up, not down on the CPU, and fix those cables in the middle of the case -- they're interfering with the intake at the front.

And learn how to light your subject when you're taking a picture man :tongue:

Since both PSU and rear panel fan are sucking air out of the case, I would prefer the airflow for the CPU is downward (to cpu) instead of trying to suck air out of the case.

I hope this photo gives better visual:
http://s963.photobucket.com/albums/ae118/lancey2010/Computers%20-%20Hardware/?action=view&current=PICT0004.jpg

It's the fact that the measured temperature of GPU/CPU/HDs never exceeds 77C makes me think it could be the PSU that's not giving enough current however from the kill-a-watt the whole computer never draws more than 130W. Again that might already be the limit of the PSU I have (though ATX12V v2.2, according to wikipedia, has standards high enough for my setup) so I guess it's either PSU quality issue or something mysterious - have to get another PSU or test the video card in another computer to really tell where the problem is.

Since both PSU and rear panel fan are sucking air out of the case, I would prefer the airflow for the CPU is downward (to cpu) instead of trying to suck air out of the case.

I hope this photo gives better visual:
http://s963.photobucket.com/albums/ae118/lancey2010/Computers%20-%20Hardware/?action=view&current=PICT0004.jpg

It's the fact that the measured temperature of GPU/CPU/HDs never exceeds 77C makes me think it could be the PSU that's not giving enough current however from the kill-a-watt the whole computer never draws more than 130W. Again that might already be the limit of the PSU I have (though ATX12V v2.2, according to wikipedia, has standards high enough for my setup) so I guess it's either PSU quality issue or something mysterious - have to get another PSU or test the video card in another computer to really tell where the problem is.
The CPU fan is trying to draw air from the same "reservoir" as exhaust fans. Meaning that it's trying to draw air that's already at a low pressure and push it over the CPU -- meaning that not as much air will move and not as much cooling will take place. By blowing up, air will be drawn from around the CPU, especially around your cable jungle. As the air passes over your CPU, it'll cool it.

The CPU fan is trying to draw air from the same "reservoir" as exhaust fans. Meaning that it's trying to draw air that's already at a low pressure and push it over the CPU -- meaning that not as much air will move and not as much cooling will take place. By blowing up, air will be drawn from around the CPU, especially around your cable jungle. As the air passes over your CPU, it'll cool it.
i guarantee if anyone tries this, temps will go up

i guarantee if anyone tries this, temps will go up
thanks for your opinion?

Or do thermodynamics and fluid dynamics not apply to a computer case?

The CPU fan is trying to draw air from the same "reservoir" as exhaust fans. Meaning that it's trying to draw air that's already at a low pressure and push it over the CPU -- meaning that not as much air will move and not as much cooling will take place. By blowing up, air will be drawn from around the CPU, especially around your cable jungle. As the air passes over your CPU, it'll cool it.

There is a plastic duct right above the CPU fan which direct the air from the side panel of the chassis (side panel with grill) as seen in the photo below:
http://s963.photobucket.com/albums/ae118/lancey2010/Computers%20-%20Hardware/?action=view&current=PICT0002.jpg

Since both the PSU fan and the rear panel fan are creating a negative pressure inside the case while there is one front panel fan directing the air from the outside into the case, i think directin some more air from outside of the case onto the CPU heatsink through the side grill should lower the CPU temp.

If I get motivated enough I might do some experiment (switching air blowing direction) inside the case and report back to this thread.

thanks for your opinion?

Or do thermodynamics and fluid dynamics not apply to a computer case?
of course they apply, so why dont you go and test this and report back with real results.

of course they apply, so why dont you go and test this and report back with real results.
Did it when I built my HTPC. Not exact same box as OPs though. 4C cooler the way I have it, with CPU fan blowing up towards exhaust.

That's not going to hold for all boxes, by any means.

The CPU fan is at an odd position in the pics -- or is it just me? Looks like it's turned counterclockwise a little bit and not attached properly? In the pic that OP posted in #32 ( #6 of 11), it looks like it's turned and not attached in the screw holes correctly. I "think" I can see a screw hole in the upper right about an inch to the right of where the upper right screw would go through the CPU fan to attach it to the heatsink. The whole fan "appears" to be turned to the left a little bit and not seated correctly over the heatsink and processor.

i guarantee if anyone tries this, temps will go up
have you tried out heatsink temps that "pull" the air off the CPU?

Also ever wonder why people use push/pull settings as well?

The CPU fan is at an odd position in the pics -- or is it just me? Looks like it's turned counterclockwise a little bit and not attached properly? In the pic that OP posted in #32 ( #6 of 11), it looks like it's turned and not attached in the screw holes correctly. I "think" I can see a screw hole in the upper right about an inch to the right of where the upper right screw would go through the CPU fan to attach it to the heatsink. The whole fan "appears" to be turned to the left a little bit and not seated correctly over the heatsink and processor.

I got the heatsink/fan on ebay(Lenovo LGA775 desktop CPU heatsink/fan). Here is another auction listing that shows the photo of the thing.
http://cgi.ebay.com.my/IBM-Lenovo-INTEL-LGA775-Cpu-Fan-Heatsink-31033965-/180561164606?pt=LH_DefaultDomain_0&hash=item2a0a48b93e

I like the fan because it can be really powerful (and noisy) if I keep it running at max. rpm by disabling the BIOS auto fan speed management.

Because the way the fan is mounted on the heatsink, I don't think it's possible to change the air flow direction. I will have to find the original intel CPU fan/heatsink or other fan/heatsink to do that experiment.

does your 8600GT have a fan? I suspect that maybe the GPU fan is spinning higher and higher instead of your CPU. it would help if you could monitor your temps real time; using MSI Afterburner to find your in-game temps or a synthetic benchmark like Furmark could tell you whether your GPU is the problem or not. your build looks AMAZINGLY clean dust-wise for something that's a few years old, but maybe check the dust buildup on the 8600GT. check the vents on the GPU heatsink where dust likes to build up.

i suspect the GPU much more than the CPU because I used to have a case with very crappy air flow...really just a single 120mm fan pushing air out the back and my dual core Athlon 64 X2 4200+ CPU didn't have problems. eventually my video card did die though but it could've been any number of causes.

I got the heatsink/fan on ebay(Lenovo LGA775 desktop CPU heatsink/fan). Here is another auction listing that shows the photo of the thing.
http://cgi.ebay.com.my/IBM-Lenovo-INTEL-LGA775-Cpu-Fan-Heatsink-31033965-/180561164606?pt=LH_DefaultDomain_0&hash=item2a0a48b93e

I like the fan because it can be really powerful (and noisy) if I keep it running at max. rpm by disabling the BIOS auto fan speed management.

Because the way the fan is mounted on the heatsink, I don't think it's possible to change the air flow direction. I will have to find the original intel CPU fan/heatsink or other fan/heatsink to do that experiment.

Take the 4 screws out that hold the fan guard in place. If the fan doesn't come off, check for a locking lever holding the fan onto the mount. Flip fan over. Reassemble. There will be 2 arrows stamped somewhere on the fan. One for airflow direction and the other for fan rotation.


The 80mm front intake fan is probably 1300rpm and ~34cfm. The rear 120mm is 1000rpm and 35cfm.

Plus you have the PSU fan moving ~30cfm. This is a standard configuation

If you want better airflow and cooler temps, I'd bump up the intake to 60-80CFM. This would be good for the front intake:http://www.newegg.com/Product/Product.aspx?Item=N82E16835132020

Then you can take your existing front fan and move it somewhere else like under or near the video card or cool the items in the upper drive bays, etc.

I like the fan because it can be really powerful (and noisy) if I keep it running at max. rpm by disabling the BIOS auto fan speed management.



Loud doesn't mean it's moving air better than another fan.
In fact sometimes you don't even need a fan, just a really big heatsink.

You could probably rig up a 120mm fan on the video card and get a case fan for the grill for intake? Look for CFM (air flow) values. Higher the better, if you don't care about sound.

thanks for your opinion?

Or do thermodynamics and fluid dynamics not apply to a computer case?

I finally got around to switching the fan around on my stock Phenom II x4 heatsink. Idles temps were the same at 38C. Load temps went up 1-2C. The real killer is that the fan was about 50% louder. Then I used SpeedFan and saw the CPU fan was spinning 2200RPM faster. Max RPM in stock configuration is 4500rpm..with the fan reversed it was going up to 6700RPM. It was unbearably loud and surely would damage the fan prematurely.

Needles to say I put the fan back to the original direction ASAP

I finally got around to switching the fan around on my stock Phenom II x4 heatsink. Idles temps were the same at 38C. Load temps went up 1-2C. The real killer is that the fan was about 50% louder. Then I used SpeedFan and saw the CPU fan was spinning 2200RPM faster. Max RPM in stock configuration is 4500rpm..with the fan reversed it was going up to 6700RPM. It was unbearably loud and surely would damage the fan prematurely.

Needles to say I put the fan back to the original direction ASAP
Thanks for posting the info!

As I said, sometimes it works, sometimes it doesn't. Depends on the setup.

What you posted is interesting, though -- it sounds like your temp control system (either Speedfan or BIOS) is trying to maintain that 38. Let it go a little warmer and it'll be quieter. You won't hurt the system if it's running at 45.

Speedfan drops the fan speed to 60%(my setting) when temps lower to 38C. Any higher and it speeds up.
The problems is that the 60% and 100% speeds were different with the fan reversed. 100% blowin at the CPU is 4500rpm. 100% blowing away from the CPU was 6700RPM! There is little sensor towards the middle of the fan and Im guessing it has to do with that.

That card has a fan on it, yes? Does it operate properly?
With the side off the case, aim a fan directly at the card and see if the problem persists. If so the problem probably is heat related.

:iagree: