Applied AS5 in Wrong Direction...

[metallbizkit]

Ok, so I just built a new PC from scratch (my first time woohoo). I have a Q9550 with Zalman 9700LED and Arctic Silver 5 in between. Being a total noob, I somehow was using the directions for application of the AS5 for a dual core and not the quad core instructions. Basically the difference was that I was supposed to apply a horizontal line for the quad, but I applied a vertical line as you would do for a dual. As this was my first time mounting a HSF, I happened to slide the HSF around a bit on top of the CPU (so maybe the AS5 got nicely spread out?).

My temps are 30-35C on idle and with Prime95 it is pretty much locked at 45-50C (stress test with small TFTs testing a lot of RAM -- as it says). I'm not sure of the exact ambient/room temp, but it is pretty close to around 72F. So the temps on the CPU seem fine, but I plan to overclock the CPU from 2.83ghz to 3.4ghz and I want to make sure I don't need to reapply the AS5 first. The AS5 directions state in caps to make sure you apply the paste in the proper direction (DOH!).

Being that I'm new to this, I figured I would ask the experts. So basically: reapply in the proper direction or stay with what seems to be running fine.

 

[badge]

You can reapply the AS5 if you feel the need. Be sure to clean off the old paste thoroughly. Isoprophyl alcohol and a soft cloth work well. If you notice when you take off the HS, the AS5 you applied will have 'melted' and spread itself from the specific areas where you deployed it! So, no matter where you place the paste on the CPU protective cover, the paste will 'spread itself'. Be sure not to put to much paste when you reapply the AS5. A couple of rice sized drops should be enough. HTH.

 

[Grimmy]

First off, please specify what sensor your using. The Tcase sensor would be the IHS sensor, which would give you a single reading. In windows your MB utility app that shows CPU/System temps, or if you use speedfan and determined correctly the temps when they are labeled temp1/temp2/temp3.

The Tjuction would be for the cores. Those readings are off temp apps that normally labeled Core Tempatures, or Core0/Core1/Core2/Core3. On idle, temps should not be trusted, since the programs need a tjmax in order to give a reading. The reason why different apps may give a different readings is because their tjmax are not the same. Some apps allow you to change the tjmax if you calibrate the sensors to give a more accurate temp when your CPU are under a load.

The stress test that would give you the most heat right when you start it, is the small TFT maximum FPU stress test that doesn't test much ram. Usually when I see the highest temps is when the stress program uses the 800000 iterations. Keep in mind these stress tests heat the CPU more beyond normal programs. Only Folding@home programs can come somewhat close depending on the WU your system is working on.

As far as the thermal grease, I've always find it best to put a bb size in the center, and spread it out on the IHS evenly. Using plastic wrap on your finger to spread it out help keeping foreign particles out of the thermal grease, or using something like a credit card will also work.

The last thought I have is, regardless of what HS your using, the best thing you can do to have lower temps, is to provide your system with cool air. So the 2 key things to also remember is, airflow, and cool ambient temps.

Edit:

On a side note, I rather go by the Tcase temp, then cores, especially on idle temps.

 

[metallbizkit]

Well I have been using CoreTemp to get my readings (I have given the readings the program reads--I haven't adjusted them). I have no idea what sensor I'm using. I hadn't tried speedfan, but just installed and tried it out... 41/44/46/46.


So would you recommend I reapply the thermal grease or does it seem like I'm fine? Keep in mind I want to OC to 3.4ghz.

 

[geofelt]

I think you are good to go. Your temps look fine. There are a number of opinions on the best way to apply the paste, but they all come down to getting the proper amount(not too much, and not too little) and not having air bubbles. AS5 will spread around and cure over time. I would not bother redoing it.

 

[Grimmy]

Here's a screen cap of all the apps that I tend to use, but I depend on speedfan the most since I've configured it myself:

You will noticed the MB utility app, which is the LAN Party DK app. That was how I configured speed fan to show the CPU/System/Chipset temps.

This is how I determined my ambient temp that the front PC case takes in:

Now for your stock settings I'd agree with geofelt, they look okay. But try doing the FPU max stress test. That should give you the most CPU heat.

Also, here is a link to my OC test on my DFI MB:

Q6600 - DFI P35 DK T2RS Swap

If you look at my stress test on that, that was when my bedroom was around 85F on the first test, about 80F on the second. And you should notice that the first test is the FPU, while the blend test is second.

I hope you understand that the ambient temp is the main factor in getting cooler temps from the first pic on this thread when my ambient temps are 75F.