2500k temperature woes

Goose · 12 Dec 2011 at 11:47

In the past few weeks i've had a couple of threads where i've been struggling to get reasonable temperatures out of my CPU with both the Noctua NH-D14 and the Thermalright Silver Arrow. I ended up returning both as i got completely fed up with it all. I bought a Gelid Tranquillo as a cheaper alternative deciding i'll just run a gentle OC on it. After installing it i got the following results;

Below shows the temps of a short IBT run on high on stock settings.

Using my motherboard's OC Bios settings for 4.4ghz i got the following temps on another IBT run on high.

To me these temps (80c on the hottest core!) again seem really high so i'm begging to wondering if i have a really awful 2500k and its not just me being terrible at installing aftermarket coolers.

If anyone has any thoughts i'd be glad to hear them.

Thanks

Lunawolfy · 12 Dec 2011 at 12:04

Spread the TIM around the chip with your finger wrapped in cling film (the condom method)

I don't trust blob and push

Goose · 12 Dec 2011 at 12:09

Lunawolfy said:
Spread the TIM around the chip with your finger wrapped in cling film (the condom method)

I don't trust blob and push

I tried both methods with previous coolers with no noticeable difference. I did the blob method with the Gelid.

Lunawolfy · 12 Dec 2011 at 12:11

Have you got other software to report temps? Real Temp, HWMonitor?

Threepwood · 12 Dec 2011 at 12:18

What case are you using? How many case fans are in it. Where and how is the case located at home? How warm is the room the pc is in?

Always used the credit card method myself, I cant see how any other method would be better at getting a perfectly thin and flat coating of paste.

Goose · 12 Dec 2011 at 12:24

shadesofgrey said:
What case are you using? How many case fans are in it. Where and how is the case located at home? How warm is the room the pc is in?

Always used the credit card method myself, I cant see how any other method would be better at getting a perfectly thin and flat coating of paste.

Case in sig, 2 intake fans, 2 exhaust. The case is about 2 feet from my radiator but the room is still cold enough that i need to wear socks

I've always shied away from the credit card method because i dont want to scratch the chip or contaminate the paste with whatever might be on the card.

HaggisUK · 12 Dec 2011 at 12:27

Less is more http://www.hardwaresecrets.com/article/What-is-the-Best-Way-to-Apply-Thermal-Grease-Part-1/1303/1

Plec · 12 Dec 2011 at 12:31

shadesofgrey said:
Always used the credit card method myself, I cant see how any other method would be better at getting a perfectly thin and flat coating of paste.

Nothing wrong with the credit card or cling film spread method (i use to use this method in the past - still do with exposed heat pipes) - but allowing the HS to spread the TIM due to the pressure of the fitting is easy and effective (although a different approach is usually required with exposed heatpipes at the base of the HS.).

Take a look at this film using a clear plastic strip to show how different application methods spread when put under pressure - link.

Goose · 12 Dec 2011 at 12:37

HaggisUK said:
Less is more http://www.hardwaresecrets.com/article/What-is-the-Best-Way-to-Apply-Thermal-Grease-Part-1/1303/1

Yeah with Gelid i used a grain of rice sized blog in the centre so i dont think thats the issue.

Karateyid · 12 Dec 2011 at 12:47

I'd download Realtemp and see what the temps are saying there... I've seen big differences in temps reported...

How fast are your case fans going ?? can you feel toasty warm air out of the exhausts??

Threepwood · 12 Dec 2011 at 12:54

I appreciate the links, but on the first one there is way too much paste spread on there, 3 times the ammount that I would be able to spread using a 'credit card'.

The second link has him placing the 'heatsink' parallel onto the chip, its better to have a slight angle that ends up parallel.

You never scratch the chip with the CC meathod, the card never touches the chip... you use the card at a very flat angle pushing the paste down so it spreads like the line meathod (but with the line from one side, not in the middle) while also pulling it along the chip... Then you scrape paste off, if needed, and give it a final sweep with the card at a more vertical angle to ensure it is perfectly flat.

Anywhoooo

I'm not sure what is wrong then, vcore looks a little high for 4.4 maybe, try to reduce it a little and see if she is stable/cooler.

Goose · 12 Dec 2011 at 15:18

Removed the cooler to reapply the paste and reseat the cooler itself. My temps are exactly the same.

The photo's below show the thermal paste spread from my first installation. Obviously it didn't spread very well so i applied extra pressure to the screws when reseating the cooler using a pair of pliers when my humble screwdriver would no longer do it. Although as you can see from the above picture it has no helped my temps.

Goose · 12 Dec 2011 at 15:40

I decided the document my re-install of the cooler to see if anyone can point out an error.

Backplate with screws in place.

Another angle.

Topside. Paste applied with washers and spacers attached to the 4 mounting screws.

Cooler put in place.

All 4 spring loaded screws tightened as much as i could.

Tekgun · 12 Dec 2011 at 16:39

I would say you need just a little more TIM from those previous photos.

Those temps are not massively high for air are they?
I would not expect to see temps like that when gaming or anything else for that matter, IBT pretty intense. I have a custom loop with single 120 rad and at 4.5 see about 74 on the hottest core with IBT, but only 52 max when gaming.

lollyhayes · 12 Dec 2011 at 16:50

Hmm, there is a small degree of disparity between the cores 67 to 79 (12 degrees), it could just be the way it came out of the factory, or t could be to do with the distribution of the paste on the cpu.
Try using a tiny bit more TIM, make sure you tighten the screws diagonally e.g in a square (with a screw at each corner) go top left, then bottom right, top right then bottom left.
Make sure you sequentially tighten as well dont go max tightness on the first screw, go throughall tightening them a bit at a time.

Also, don't worry about the temps, IBT takes the temps waaayyyyy higher than any normal day to day app or game would.