Overheating GPU / very hot TMPIN0

[awkward_hippo]

Hello,

Been struggling to get a new game to play on my system.

I think GPU has been overheating. It seems to be idling at about 92c, and when I play starcraft 2 it shoots up to 128+ and system crashes. ( I didn't know all this until I started doing some research).

There was an issue with this game with nvidia graphics cards with the framerate freaking out but I changed the log.

It seems to be more of system thing than a game specific thing as about a 6 months ago another game was overheating as well but not enough to crash the computer (comp would beep at me).

Running HWMonitor, CPU temps stay around 35-40 deg.

TMPIN0 is constantly at 128 deg though which seems ridiculously hot.

Yesterday I opened the case up and cleaned around as best I could for about 30 minutes, and it seems to be running a bit cooler and taking longer to overheat playing game but not great.

Thoughts? Also is this the right way to post?

System specs:
Intel Core 2 Quad Q9400 2.66GHz, 1333FSB (Quad Core) 6000K * Spire Socket 775 Intel fan * Biostar P43D2-A7 P35, 1600FSB, PCI Express, Sound, 1Gb LAN * 8GB (4x2GB) PC6400 DDR2 800 Dual Channel * GeForce 9600GT 1GB PCI EXpress 16X dual head, tv out * 500.0GB Hitachi/ IBM 7200RPM SATA2, UDMA 300 8m cache * Samsung LightScribe 20x SATA DVD Recorder Dual Layer +R/RW -R/RW * AC 97 3D Full Duplex sound card (onboard) * Nikao Black Neon ATX Case w/ Temperature control, front USB & iEEE * Okia 550W ATX Power Supply w/ 6pin PCI-E * Microsoft Windows Vista Home Premium 64bit DVD

 

[Bill Brasky]

Is your 9600GT still under warranty? If not, you have two choices: 1) remount the heatsink on your graphics card w/ new thermal paste or 2) buy a new card.

Remounting the heatsink can be a little intimidating at first if you've never done it before, but it's really not that hard. You pull out some screws from the back, clean the die with 90% isopropyl alcohol (or higher), reapply some ceramic thermal paste, and then screw the heatsink back on.

HWmonitor is probably fine, but a nice program call gpu-z can really give you a lot of info about your card. I'd recommend looking at it. gpu-z

Just for reference, 50C would be a normal idle temp for a gpu. 92C is wayyyy too hot.

 

[Seero]

Read this

 

[awkward_hippo]

I will take a look at adding a heat sink to my video card, sort of sounds like a fun little project.

I downloaded GPU-Z, its got lots of info but not sure what exactly is too pertinent.

One question, obviously running GPU at high temps is bad and can burn out your card, has mine just not burned out yet because it still works? Or can it "burn out" and just sort of start sucking but still work.


Also, any idea what TMPIN0 would be? Looked around online but seem to find vague answers like "part of your computer case" or some part of the motherboard maybe? Also why would that be running so hot on my machine?


Thanks
Julian

 

[awkward_hippo]

also would underclocking my card using rivatuner or something like that help to keep it cooler?

Julian

 

[3DVagabond]

From what I can gather googling it, TMPIN# is various temps. I've seen listings for system, NB, CPU, SB, MOSFETs. Nothing saying it's the GPU though. Seems though to be related more to mobo components?

 

[Bill Brasky]

I will take a look at adding a heat sink to my video card, sort of sounds like a fun little project.

I downloaded GPU-Z, its got lots of info but not sure what exactly is too pertinent.
Thanks
Julian

Open Gpu-z and click on the sensors tab. It looks like this. It will tell you the temperature of your graphics cord. Also, you won't be adding a heatsink to your card. You would take off the existing heatsink and replace the thermal compound that transfers heat from the graphics core to the heatsink. The factory paste originally installed on your card could have dried up, which would cause the high temps your seeing. Please check your temps with gpu-z.

 

[awkward_hippo]

very hot 94 degrees just sitting here on internets....computer is 1.5 years old, not sure about warranty I got it through magic micro online so might give them a call tomorrow.

Julian

 

[Rifter]

reseat/clean the heatsink, and apply good TIM.

 

[betasub]

1. Clean out dust & lint from inside case & on graphics card heatsink, ensuring good airflow.

2. Check warranty.

 

[Seero]

I will take a look at adding a heat sink to my video card, sort of sounds like a fun little project.

I downloaded GPU-Z, its got lots of info but not sure what exactly is too pertinent.

One question, obviously running GPU at high temps is bad and can burn out your card, has mine just not burned out yet because it still works? Or can it "burn out" and just sort of start sucking but still work.


Also, any idea what TMPIN0 would be? Looked around online but seem to find vague answers like "part of your computer case" or some part of the motherboard maybe? Also why would that be running so hot on my machine?


Thanks
Julian

To answer your question, TMPIN0 is a channel of data that came from your motherboard. The data should represents the reading of a digital sensor that detects temperature. However, it isn't a thermometer and the accuracy of the reading can be way off. In your case it is very likely to be the GPU temp because your fan speed is at 100%. If you haven't manually force the fan speed, then your video card itself believes that it is too hard and cranked its fan up to max speed, indicating that it is too hot.

The idle temp should not be 30c higher than room temp. That means, there is something wrong if the idle temp is above 60c. There are two things, the temp and the fan speed. Your GPU temp must be really hard for the fan to run at 100%, especially when you are not playing games.

Is the fan even turning? Are there hot air coming from the back of the v-card? At 100% fan speed at 90c, it should sound like a hair dryer and feels like one. If it does sound and feel like one, then I suggest you to bring it to future shop (or shops that provide repair services) and have someone to take a look at it.

My guess is poor vent. Do you by chance have pets with fur and smoke near the PC?

 

[Powermoloch]

Holy cow buddy ! it's pretty hot in there. The best thing you can do is to follow what's been typed out above me.

1) Make sure you have good cable management.
2) Free of very large dusts.
3) Ensure there is ventilation (fans working as well)
4) Reseat GPU's heatsink. It's not bad at all.

Good luck !

 

[awkward_hippo]

So after I cleaned out my computer etc, rebooted and found it hadn't cooled much. So after a day or so I thought to check the fan while it was actually on and low and behold, it wasn't actually spinning. Dammit

Called Magic Mirco where I bought my computer from, 2nd year of warranty so 55$ deductable on the video card.....card itself is 65-70 bucks......talked to the tech guy who was awesome, he explained its common for fan to break and I can either try to rig a new fan or buy a new card.

So took out the video card, ripped off the nvidia fan attached to the heat sink....took a fan from the side of my case closest to where the video card would be (my case has 2 external fans plus the fan on the power supply) unscrewed it and then jerry-rigged (spelling?) it to the heat sink using thin copper wire.

Fired it back up, and my GPU is now cruising at 54 deg C, and topping out about 61 degrees while playing starcraft 2. Down from 90 deg and topping out at 128 (not sure if i believe this reading but hey) and ultimately computer crashing.

I am actually amazed that the card still works because I assume the fan hasn't been working for quite some time.

I feel pretty darn good about myself.

Also found out that I thoroughly enjoyed learning about the physical/hardware makeup of my computer and how it all fits together etc.

Plus excited to play games now!

 

[Seero]

So after I cleaned out my computer etc, rebooted and found it hadn't cooled much. So after a day or so I thought to check the fan while it was actually on and low and behold, it wasn't actually spinning. Dammit

Called Magic Mirco where I bought my computer from, 2nd year of warranty so 55$ deductable on the video card.....card itself is 65-70 bucks......talked to the tech guy who was awesome, he explained its common for fan to break and I can either try to rig a new fan or buy a new card.

So took out the video card, ripped off the nvidia fan attached to the heat sink....took a fan from the side of my case closest to where the video card would be (my case has 2 external fans plus the fan on the power supply) unscrewed it and then jerry-rigged (spelling?) it to the heat sink using thin copper wire.

Fired it back up, and my GPU is now cruising at 54 deg C, and topping out about 61 degrees while playing starcraft 2. Down from 90 deg and topping out at 128 (not sure if i believe this reading but hey) and ultimately computer crashing.

I am actually amazed that the card still works because I assume the fan hasn't been working for quite some time.

I feel pretty darn good about myself.

Also found out that I thoroughly enjoyed learning about the physical/hardware makeup of my computer and how it all fits together etc.

Plus excited to play games now!

Say no more brother. GAME ON!