Sign in with
Sign up | Sign in
Your question

Weird cpu overheating during games

Last response: in Video Games
Share
January 2, 2013 10:42:33 PM

Hello,Hello, I've been having a recent issue with my year old, custom build gaming pc that involves strange cpu overheating. Only recently, my cpu has started to hit around 60-65 degrees C. The kicker is that if i simply alt tab out of the game, it goes straight back down to 40 C, like if it were idling. Is it usual for the cpu to get hot only if the game is actually running, but not if it is alt tabbed out yet still on. Also, only my cpu is overheating, my gpu is fine. Don't know if this will help but this also doesn't happen for games that actually overheat my pc, specificly Far Cry 3. It gets that hot even if i'm alt tabbed out, so I will be getting a new heatsink soon. Any suggestions about what the problem may be, I would love to keep playing games. My specs are windows 7, 64 bit, an intel core i5 2400 @ 3.10GHz cpu, 8 GB of ram, a biostar group TZ68K+ mobo, GeForce 9800 GTX+ gpu, 699 GB Hitachi HDS721075DLE630 hard drive. Any help would be appreciated.
January 2, 2013 10:48:59 PM

Really, because at about 70 it turns off my computer entirly, not just a restart. When I turn it on, i don't get an error message. This has only happened on Far Cry 3, and on a couple of other games that aren't that graphic intensive, like left 4 dead 2 and saints row the third. I still plan on getting a new heatsink, which will hopefully fix te far cry 3 problem. So what is "Hot" for an i5? Also, if my computer isn't getting too hot, why is it shutting down?
m
0
l
January 2, 2013 10:54:09 PM

What kind of power supply do you run, and cooler for the i5? Unless overclocked that should definitely not be a issue, and even if so what is the case, also try to run all fans at 100% and see if it does not shut off, and post temps
m
0
l
Related resources
January 2, 2013 10:56:08 PM

I'm leaning towards the psu though
m
0
l
January 2, 2013 10:58:53 PM

My power supply is a zt series 650 watt modular atx power supply and it's not overclocked. I was wondering if maybe my power supply was more of the issue than my heatsink. Which would be more likely to fix my problem if replaced?
m
0
l
January 2, 2013 11:17:53 PM

Quote:
First pull your video card and remove the heat sink shield. If dusty clean it out and see if that fixes your problem.

Ok, I'll do that and let you know if it works. I have cleaned out the dust fine but I haven't done the heat sink shield yet.
m
0
l
January 2, 2013 11:46:03 PM

Quote:
First pull your video card and remove the heat sink shield. If dusty clean it out and see if that fixes your problem.

Kinda embarrased but could you link me to a good tutorial on how to remove and clean the heat sink shield. Never done it before.
m
0
l
a c 112 à CPUs
January 3, 2013 12:20:19 AM

the max temp of the intel 2400 is 72'c so its working correctly if its shutting off the pc at 70'c. things you can do to increase efficiency of your fans. if you bios has a fan control where you can increase the fans rpm at a given temp is lower the minimum temp at which the fan reaches 50% this will broaden the thermal scale. it may still get up to 65'c but should take a bit longer to get there but the chances are because your running the fan faster at lower temps you should be able to reduce the over all max temp by 5-10'c

getting a good 3rd party cooler is the better option. there are some cheap but very good options out there. i have a thermolabs barham which i used when i first got my i7 and it worked well. very well in fact for a £20 cooler. i managed to run my i7 920 @4.2 ghz wile maxing out at 73'c not many coolers at the time could match that and none for the money.

the 212 evo is probably the current best comparison to the thermolab cooler although it could use a better fan. swapping the original for a fluid dynamic bearing fan will make this noisy chunk into an effective and quiet cooler.
well worth the all in asking price of 35 fo the cooler and new fan.
m
0
l
January 3, 2013 12:42:06 AM

Quote:
Its not really the shield I was talking about but all the slots in the actual heat sink that you can't get to with the shield on. There should be four screws on the back of the card that hold it on. Remove those and it should come off. You'll want to clean the thermal paste off the gpu and heat sink and re-apply when done.

You could always get a can of compressed air from wal-mart and blow it out.

Oh, I've already done that
m
0
l
a c 112 à CPUs
January 3, 2013 1:08:37 AM

how to remove the heat sink and clean it...
first you will need some bits and bobs.
some good quality thermal paste zalman tuniq arcticcool all make very good pastes that are within 1'c of each other for perfomance...
acrtic silver number 5 or mx 2 or mx 4
tuniq tx-4
zalman zm stg1 are all good and will do the job well...

stay well clear of akasas or similar ultra budget brands they do work but not very well.

a size 1 posi drive or cross cut screw driver.
a small 1 inch brush
some lint free natural fiber cloth
a bottle of rubbing alcohol 90% or better isopropyl alcohol if you cant find any TIM remover (thermal interface material/thermal paste).
a clean and dry work surface (i prefer working on the floor)
some paper towel.
an optional earthing lead.

turn off the pc completely.
let all the residual current drain from the board and psu by leaving the plug in the wall after you power off the pc for about 5 mins or till you see all the motherboard led's go out.
while its pluged in you will want to earth yourself. you can do this by touching the case as the earth from the wall will ground you.
if you have a carpeted room i recommend you dont work in it, if its nylon as 1 swipe of a foot could put enough electricity into the board to kill your cpu or something else.

1s your satisfied your earthed properly by attaching your earthing strip to a radiator or some other ground if you have 1.
remove the plug from the wall and open the case.

your ready to start cleaning.

after you open the case have a look at what goes were your looking for the lead that comes off the cpu fan to the cpu fan header.
unplug this first.
if your fan is a clip on then prise open the clips from either side and remove the fan and place it to 1 side.

find the locking mechanism for your heatsync some will be
finger screws, posi screws, or quarter turn grip locking clip.
zalman do a simple snap catch which you just lift the loose catch and with a little pressure it should just pop off.
which ever they are gently remove it/them
if there the press down type from intel then give them a quarter turn and pull upwards they should come away with a small pop if it doesn't dont force it just turn it back the quater turn and try pulling gently till you hear the pop..
the other kinds will just unscrew or simple clip.

remove the heat sync then if you want you can do 1 of 2 things for the cpu.
you can leave it where it is and use a cotton swap with a little cleaning solution (tim remover) to remove any paste residue.
or
you can unclip and slide out the cpu and clean it off then.

clean both heat conducting surfaces properly by wiping the bulk off with a paper towel then use the natural fiber cloth with a little tim on it to clean any remaining residue of the old paste.
clean and dry both contact surfaces...

take the brush and use it to clean between the fins of the cooler and the blades of the fan 1s your happy that everything is dust free you can start putting the heatsync back together.

put the cpu back in its housing if you removed it, making sure the locking clip is lifted and fully extended, place the cpu in the socket making sure its the correct way round. it will just sit on the surface... now gently pull down on the lever and the cpu should drop and slide into the socket. 1s the lever is in the fully down position lock it in place.

make sure you havent gotten any finger prints or dust on the heat spreader of the cpu. if you have give it a quick wipe with the cloth with tim on it. making sure its clean and tim free after...

now take your tim paste (likely in a syringe) put a small amount in the middle of the cpu a flat spot about the size of a garden pea or a few grains of rice.
next place the heat-sync over the cpu and press down gently giving a few gentle twists left and right to help spread the paste... lift off the heat-sync and remove anything that has spilled over. if you did it properly there should be a fingerprint type pattern on both contact surfaces with no overspill. again place the heat-sync on the cpu this time when you press it down and twist it. keep twisting left and right a small amount till you feel the resistance change from slippy to stiff. this means the air has been forced out and its time to lock the heat-sync down...
tighten the bolts finger tight in an X fashion corner to corner top to bottom, top to bottom. place the fan back on and clip it down, plug in the header , close the case and turn the pc on...
if all is well the pc will boot and temps will be down...
good luck.
m
0
l
January 3, 2013 1:47:48 AM

Nice post HEXiT this should help him solve the issues!
m
0
l
January 3, 2013 3:02:44 AM

Yeh, it helped lower it by about 5 degrees, but it is still getting too hot. I'm mostly like just gonna save money and get a new gpu(mine is too old anyhow), cpu/heatsink(also could do with some upgrades), and powersupply(cause why not).
m
0
l
January 3, 2013 3:56:16 AM

FC3 is by no means a easy game, is more heavy than bf3 and the newest games. But no it is normal, but if you are using the stock cooler, than get a better cooler like the 212 evo.
m
0
l
January 3, 2013 5:38:48 PM

Try underclocking a bit and adding a GPU heatsink with a water cooler.
m
0
l
January 4, 2013 8:30:12 PM

Ok, i bought a 212, and it should be in a few days but the problem is getting worse. It is starting to get to 60 if i have too many windows open now. Please hurry, newegg delivery!
m
0
l
January 4, 2013 8:45:04 PM

Well, not sure if this helped but i messed around in my Nvidia control panel and turned off v-sync. It seemed to lower the temps by about 5 degrees(playable without crashing). Would this actually help the temps or am I just stupid and hopeful?
m
0
l
a b à CPUs
January 4, 2013 10:12:36 PM

Hexit, your information isn't wrong. Sandy bridge CPUs start to throttle at 95°C and shut down at 100°C.

@OP: What are your graphics card temperatures?
m
0
l
January 5, 2013 1:13:44 AM

Sunius said:
Hexit, your information isn't wrong. Sandy bridge CPUs start to throttle at 95°C and shut down at 100°C.

@OP: What are your graphics card temperatures?


My gpu remains at a pretty constant 57 or so. It doesn't get any hotter in games. The only other temp that gets really hot is core 1, and i'm not sure what that is/how to reduce it.
m
0
l
January 5, 2013 1:14:14 AM

Update: It's getting much worse, hope the new fan comes soon
m
0
l
a b à CPUs
January 5, 2013 7:21:19 AM

It's not overheating. Definitely not fan or voltage problem.

@OP: I'd suggest you test your memory. Use memtest for that. Wait for at least 3 passes.

www.memtest.org
m
0
l
January 5, 2013 5:13:03 PM

Sunius said:
It's not overheating. Definitely not fan or voltage problem.

@OP: I'd suggest you test your memory. Use memtest for that. Wait for at least 3 passes.

www.memtest.org

I can't do memtest because the only disk burning software i found to work with it won't download on my computer. Can you give me some more information about what could be the memory problem/how I could possibly fix it?
m
0
l
a b à CPUs
January 5, 2013 5:20:42 PM

You don't need any disc burning software as Windows 7 can do that natively. If you want to write it to DVD, right click the .iso file and select "Burn disc image". Or, if you want to use a USB drive, just download the auto installer.
m
0
l
January 5, 2013 5:26:12 PM

Sunius said:
You don't need any disc burning software as Windows 7 can do that natively. If you want to write it to DVD, right click the .iso file and select "Burn disc image". Or, if you want to use a USB drive, just download the auto installer.

Ok, and what memory would have the problem, my ram or hardrive?
m
0
l
a b à CPUs
January 5, 2013 6:17:35 PM

RAM. When you hear memory, it's more than likely talking about RAM :) .
m
0
l
January 5, 2013 9:44:33 PM

Sunius said:
RAM. When you hear memory, it's more than likely talking about RAM :) .

Figured. I still wasn't able to get the memtest working but I did go ahead and do windows memory scan. Got 0 errors. Is there any way to fix ram problems other than getting more and if I were to get more, how much would it cost? Also, why would a lack of ram cause these random shutdowns? Also, would it explain my rising cpu temps?
m
0
l
a b à CPUs
January 5, 2013 10:19:53 PM

I never said lack of it. I meant it could be faulty. Windows memory scan is bad. Don't use it.

And rising temperatures are normal. CPUs perform calculations, and their product is heat. That's how they work.

Anyway, on which step of memtest have you stuck?
m
0
l
January 5, 2013 10:25:00 PM

Sunius said:
I never said lack of it. I meant it could be faulty. Windows memory scan is bad. Don't use it.

And rising temperatures are normal. CPUs perform calculations, and their product is heat. That's how they work.

Anyway, on which step of memtest have you stuck?

Well, over the past week the temps have gotten hotter by over 20c while playing games, even nonintensive games like minecraft are unplayable with my cpu getting so hot it shuts off, a problem I have never had before. My problem with memtest is that I can't get either the normal download or the usb download to work. I can download them but then I can't do anything with them.
m
0
l
a b à CPUs
January 5, 2013 10:42:33 PM

What do you mean you cannot do anything with them?

Here's a link to the .iso download:

https://www.dropbox.com/s/0ju8oghcjqz4c3o/mt420.iso

It should work even if for some reason www.memtest.org doesn't. Just right click the downloaded file and press "Burn disc image".

P. S. Last time I'm telling you: 60-65°C is not overheating. 90°C would be overheating. It's the temperature that's causing the issue.
m
0
l
January 5, 2013 10:47:21 PM

Sunius said:
What do you mean you cannot do anything with them?

Here's a link to the .iso download:

https://www.dropbox.com/s/0ju8oghcjqz4c3o/mt420.iso

It should work even if for some reason www.memtest.org doesn't. Just right click the downloaded file and press "Burn disc image".

P. S. Last time I'm telling you: 60-65°C is not overheating. 90°C would be overheating. It's the temperature that's causing the issue.

For my processor, it autoshuts off at 73C or so because of heat.
m
0
l
a b à CPUs
January 5, 2013 10:48:30 PM

Where did you read that? That's false information. i5-2400 shuts down at 100°C.
m
0
l
January 5, 2013 10:50:32 PM

Sunius said:
What do you mean you cannot do anything with them?

Here's a link to the .iso download:

https://www.dropbox.com/s/0ju8oghcjqz4c3o/mt420.iso

It should work even if for some reason www.memtest.org doesn't. Just right click the downloaded file and press "Burn disc image".

P. S. Last time I'm telling you: 60-65°C is not overheating. 90°C would be overheating. It's the temperature that's causing the issue.

I'm also not getting a "burn disk image" option
m
0
l
a b à CPUs
January 5, 2013 11:30:02 PM

Aren't you using Windows 7? What is the top option when you right click the file?
m
0
l
a c 112 à CPUs
January 6, 2013 12:09:58 AM

the tcas of the i5 2400 is 73'c (source intel ark) anything over this depending on the quality of the sample can cause throttling or a shut down.
some will quite happily go on to the tjmax of 104'c but there is always a potential for damage to the cpu 1s you go over tcas.
it could well be that theres a thermal setting in bios that shuts down the cpu 1s it gets to a certain temp. some boards, have this some dont. i would have to look at the bios to see.
what ever the reason for the sut down 70+ is to warm 60 or less would be better.
maybe the op shgould go into bios and load optomized defaults. making sure he sets stuff like enhanced precision timer and pci-e priority b4 he saves and quits.
my guess is he didnt load the optimum defaults when he got the motherboard and cpu so it may well be over volting if its incorrectly set up.
if you can get windows up and running maybe you can run a simple test. prime 95 blend and then memeory also run superpi 32m. it will tell you very quickly if you have a memory issue. on the 32m test super pi uses a lot of ram and will cause the pc to lock up if the ram is unstable.
prime 95 blend will tell you if your cpu and or memory is unstable.

m
0
l
January 6, 2013 12:53:51 AM

Sunius said:
Aren't you using Windows 7? What is the top option when you right click the file?

Add files to archive
m
0
l
January 6, 2013 1:15:33 AM

HEXiT said:
the tcas of the i5 2400 is 73'c (source intel ark) anything over this depending on the quality of the sample can cause throttling or a shut down.
some will quite happily go on to the tjmax of 104'c but there is always a potential for damage to the cpu 1s you go over tcas.
it could well be that theres a thermal setting in bios that shuts down the cpu 1s it gets to a certain temp. some boards, have this some dont. i would have to look at the bios to see.
what ever the reason for the sut down 70+ is to warm 60 or less would be better.
maybe the op shgould go into bios and load optomized defaults. making sure he sets stuff like enhanced precision timer and pci-e priority b4 he saves and quits.
my guess is he didnt load the optimum defaults when he got the motherboard and cpu so it may well be over volting if its incorrectly set up.
if you can get windows up and running maybe you can run a simple test. prime 95 blend and then memeory also run superpi 32m. it will tell you very quickly if you have a memory issue. on the 32m test super pi uses a lot of ram and will cause the pc to lock up if the ram is unstable.
prime 95 blend will tell you if your cpu and or memory is unstable.

Would hitting the "restore defaults" set my bios to optimum defaults because I am having an issue where the same shutdown occurs around 15 seconds after I open bios, making it a bit hard to mess around with stuff. It took several speed tries to even turn off my autospeed fan so that it would run at full all the time.
m
0
l
January 6, 2013 1:23:20 AM

Quote:
....This is hopeless. Take your pc to a shop before you perma break it.

I'm gonna do that if the new heatsink fan that I'm getting in a few days doesn't work
m
0
l
a b à CPUs
January 6, 2013 7:40:19 AM

HEXiT said:
the tcas of the i5 2400 is 73'c (source intel ark) anything over this depending on the quality of the sample can cause throttling or a shut down.


T_case doesn't have anything to do with monitored core temps man. It's usually 15-20°C lower than the core temperature but you cannot measure it unless you have a thermometer placed between heatsink and the CPU.

Anyway, OP: are you sure you're right clicking the .ISO file and not the .ZIP/.GZ file?

P. S. I can guarantee that new heatsink will not help.
m
0
l
January 6, 2013 8:56:41 PM

Sunius said:
T_case doesn't have anything to do with monitored core temps man. It's usually 15-20°C lower than the core temperature but you cannot measure it unless you have a thermometer placed between heatsink and the CPU.

Anyway, OP: are you sure you're right clicking the .ISO file and not the .ZIP/.GZ file?

P. S. I can guarantee that new heatsink will not help.

Yeh, i'm clicking on the file called mt420.iso
m
0
l
January 6, 2013 9:00:14 PM

Sunius said:
T_case doesn't have anything to do with monitored core temps man. It's usually 15-20°C lower than the core temperature but you cannot measure it unless you have a thermometer placed between heatsink and the CPU.

Anyway, OP: are you sure you're right clicking the .ISO file and not the .ZIP/.GZ file?

P. S. I can guarantee that new heatsink will not help.

Also, what would happen to my computer if I were to straight up by more ram. Would it alter any of my memory or just possibly fix the problem?
m
0
l
January 8, 2013 9:50:11 PM

Sunius said:
T_case doesn't have anything to do with monitored core temps man. It's usually 15-20°C lower than the core temperature but you cannot measure it unless you have a thermometer placed between heatsink and the CPU.

Anyway, OP: are you sure you're right clicking the .ISO file and not the .ZIP/.GZ file?

P. S. I can guarantee that new heatsink will not help.


Thanks for the help but you were wrong about the heatsink. I got the heatsink and it lowered temps by 10-20 degrees. Everything works fine
m
0
l
a b à CPUs
January 9, 2013 3:45:51 AM

I hope it doesn't come back.

However, if it was indeed overheating, you must have looked and wrong temperatures, because 70°C is not overheating...
m
0
l
a c 112 à CPUs
January 10, 2013 2:31:01 AM

it is entirley possible he got a bad sample that isnt stable up to the tjmax. it often happens if you pick a board that auto over-volts past recommended... my gigiabyte for instance runs my cpu at 1.18125v at stock while a comparable evga runs its cpu voltage at 1.2500v both are running the cpu at stock 2.6-2.8turbo.
thing is i can run my cpu at 3.6 on stock volts but on the evga the same cpu was unstable and throttled the turbo due to heat.
100 is my tjmax but its not recommended you go over 74. mine tops out at 65'c @4.2 while my mates was doing 76 at the same speed.
so yeah not all boards or cpu samples are made equal.
m
0
l
!