cancel
Showing results for 
Search instead for 
Did you mean: 

Heeelp

Wardope
Level 7
Hello. i got a asus maximus 4 gene-z Mobo
crosair 8 GB ram 2x4
corsair forse GT 120 GB SSD
WD 2Tb storage HDD
intel i5 2500k.
Nvidia Gtx 465 (will be changed SOON) with Gtx 580 or 590.

First i just used XMP. to get my ram speed correct at 1866 Mhz

and MP set to x37 to archive 3700 Mhz.
and Core voltage set at 1.200
At prime 95 (64 bit) version my core temp reads 64C. under full load.

i dont know if its just me bieng silly bit i think the temp is a bit high. for that small oc.
My i got plenty Case cooling its like a tonado inside my InWin dragoen slayer case.
and my Cpu cooler is the Artic freezer 13 (Not pro)
im 100% positive that my Thermal compound is applayed correct. ( just a littel Rise sized pice in the middle of the CPU)

any Help apriciated.

Kind regards. Wardope.
The only reason people get lost in thought is because it's unfamiliar territory!
5,542 Views
10 REPLIES 10

skubany
Level 7
just a littel Rise sized pice in the middle of the CPU


Have you always been applying this much thermal paste?

Arctic Silver 5 website says to apply "5 cubic millimeters (or 1 uncooked grain of white rice)". If you were following those instructions to the letter, in my recent testing of various AS5 dot application amounts, I've found that this amount was too little and it's best to actually put a dot that is 3-4 times the "5 cubic millimeters" amount. Although this may be specific to my heatsink.

If there is too little thermal paste the temp will be noticeably too high. If there is too much thermal paste then the temp will be 4-8C higher from the optimal amount.

If you can, remove the heatsink and see if the thermal paste covers the majority of the CPU. The CPU should be covered at least ~80% of the top surface.

If this cooler is new, check its base to see if it's even. I had a heatsink whose base was very uneven.

skubany wrote:
Have you always been applying this much thermal paste?

Arctic Silver 5 website says to apply "5 cubic millimeters (or 1 uncooked grain of white rice)". If you were following those instructions to the letter, in my recent testing of various AS5 dot application amounts, I've found that this amount was too little and it's best to actually put a dot that is 3-4 times the "5 cubic millimeters" amount. Although this may be specific to my heatsink.

If there is too little thermal paste the temp will be noticeably too high. If there is too much thermal paste then the temp will be 4-8C higher from the optimal amount.

If you can, remove the heatsink and see if the thermal paste covers the majority of the CPU. The CPU should be covered at least ~80% of the top surface.

If this cooler is new, check its base to see if it's even. I had a heatsink whose base was very uneven.


Sorry for bad english I ment one pice of rise as in grain of white rice. hope you understand.
The only reason people get lost in thought is because it's unfamiliar territory!

skubany
Level 7
I do understand 🙂

The only way to know if it's the paste is to remove the heatsink and see how your original paste application spread out on the CPU and that will tell you if the high temperatures are paste related or heatsink related (as in the heatsink is too small for the CPU).

My old heatsink was unable to cope with my newest Phenom II x6, so I got a better one which works great.

Based on what you wrote the temp might be too high for a 1.2 vCore, as this influences the CPU temperature the most, based on my recent testing. For instance, I have Corsair H80 on a Phenom II x6, which uses larger transistors and therefore I would assume heats up even more compared to your i5, and yet at vCore of 1.35-1.4 and OC of 3,6Ghz with ambient temp of 24-25C, my CPU does not heat up above 50C socket/40C core after 40min of SP2004 (which is essentially prime95).

If your ambient temp is in the 24-25C range and you have adequate case cooling with fast enough CPU cooler fan, then the only thing left is insufficient paste application, too small a cooler or the temps are correct - either one of the three. For a 90mm fan it should be at least 1600RPM, I think, for good cooling, possibly more. I'm more familiar with 120mm fans since I've done some testing on those.

btw its strange because CPU idle with The freezer 13 Pro is around 29C. but once i crak the Prime 95 up i get a ridiculously 70% heat increasement. to 64C 😞
The only reason people get lost in thought is because it's unfamiliar territory!

Wardope wrote:
btw its strange because CPU idle with The freezer 13 Pro is around 29C. but once i crak the Prime 95 up i get a ridiculously 70% heat increasement. to 64C 😞


Good idle temps paired with unusually high* load temps usually point to an undersized cooler. The cooler gets saturated with heat and can't evacuate it any faster so once you get to that point the temps just climb.

*note that I don't have a 2500K so I'm not confirming that your temps are high. I'll leave that to others who have tested.
A bus station is where a bus stops. A train station is where a train stops. On my desk, I have a work station…

Wardope
Level 7
just checked again and used new thermal compound. it was applayed the way it shuld.
My Cooler is the actic freezer 13 PRO. i think its the cooler that is the problem because i get almost the same temps with stock cooler
so im going to replace it with swiftech Water cooling and then i can ofc add a waterbock to my GFX when i get the new Gfx.
The only reason people get lost in thought is because it's unfamiliar territory!

skubany
Level 7
xeromist
I concur. "low idle temps...very high load temps" is exactly what I had with my heatsink which ended up being not enough for the newer processors.

Another good test for the heatsink is to touch it's base and compare the temperature to the remainder of the heatsink (fins, point further away from the base). On my 'too small' heatsink the base was burning hot on one side and much cooler on another which showed that the heat transfer through the heatsink was not fast/efficient enough.

Wardope
Level 7
Find it strange because the artic freezer 13 PRO shuld be up to the job
here is a ling to the cooler http://www.arctic.ac/en/p/cooling/cpu/365/freezer-13-pro.html
The only reason people get lost in thought is because it's unfamiliar territory!

skubany
Level 7
I was looking at the cooler and did see heat dissipation of 200W for non-PRO and 300W for PRO, which would make this cooler sufficient for your processor. I say that, because when testing my cooler on an Phenom II x6 @ 3,6Ghz it was drawing about 160W when approaching critical temps, 70C. I expect i5 to run cooler than an Phenom II x6. That's why your course of action should be:
1) ensure proper amount of thermal paste by making sure that a) heatsink base is flat, b) after removing the heatsink thermal paste covers at least 80% of the CPU
2) during load testing, touch the heatsink in various locations starting with the base and determine how well it transfers heat. If the base of the heatsink is much cooler than the CPU then CPU-heatsink contact is bad, if the base is extremely hot and the pipes and fins are not, then inter-heatsink heat transfer is bad.

Only after these steps would I consider getting another heatsink.