• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

TIM testing procedure

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
Here is an updated graph, if you think thee graph could be formatted better or would like to see different info included let me know.

image (5).png
 
I might would recommend using a benchmark maybe that isnt so hard on temps. might get to see more variation in temps between pastes. When using the hottest test I would think that it might push temps rather close. Just a guess on my part though.
But nice work so far. And it is appreciated

Just the opposite.

And the cooler doesn't matter. What does matter is driving up the max temps as high as you safely can so you get as much separation as possible between the different products tested. I think Locekey's explanation of percentage of efficiency differential between products explains it well. A 1% difference at 50c amounts to only .5c differential. At 100c it's a full 1c. Easier to measure with confidence.
 
Last edited:
So not to hi-jack this thread but if I were to say I applied my thermal paste much closer to a pea sized glob than a damn BB, should I consider re-applying?
 
Does your performance say you need to reapply? You could run some benches with your Tim the way it is and then reapply and compare results. That said, unless you are short on Tim there really isn't any reason not to shoot for a perfect mount (IMO)
 
@ Lochekey,

I did the same thing at Toms Hardware back in 2011, and we have a lot of new products on the market today.

If you contact vendors and let them know what you are intending to do most will gladly submit a testing sample to you that will save you some expense.

If you go through with this not only will you have valuable information to share but you won't be supported or influenced by any company trying to sway the results, which has been done in many reviews.

Being able to control and regulate ambient room temperature is very important as many testers have not done that and consistent ambient does make a difference.

The most valuable data you're going to get out of this is how each mounting footprint can vary and though mounting pressure has been discussed as a concern consistency is more important.

You'll also need a confirmed way of cleaning off the old to get all traces of it removed so one TIMs residue is not affecting the next application.

You'll also be learning the characteristics of each TIMs viscosity where a pea size may be necessary to the thicker TIMs and less than a BB for others.

You will mount and unmount so many times you could do it consistently in your sleep, good luck to you! Ryan
 
Interesting test Loche! I personally just stuck with the pre-applied TIM when I was using intel stock coolers and when I DID buy "aftermarket" TIM, it was just cheap $5 stuff off the shelf at Best Buy. Do you have any liquid metal TIMs that you can put up against more traditional types, like silicon or ceramic?
 
Won't know til I "waste" it trying. Lol It's a pretty fast procedure, and if I am going to keep building PC's for myself and helping friends with theirs I better be able to save the 2 degrees Celsius the first time around.
 
Won't know til I "waste" it trying. Lol It's a pretty fast procedure, and if I am going to keep building PC's for myself and helping friends with theirs I better be able to save the 2 degrees Celsius the first time around.

Yeah, true. I was just poking fun at ya. Maybe you and Lochekey can compare notes too ;)
 
Wow there has been a lot of activity today. Let me take this one by one.

@Silver - thanks for the advice on your experiences with testing. I am keeping ambient between 22-23c, I am basing results of of the delta between core temp and water temp to help alleviate ambient temp variations. I have thought about contacting companies for samples but figured I should get through my pile of stuff first. I did have one company ask me to test some new TIM and I will touch on that in a little.

@Vishera- No I do not currently have any of the liquid metal products. I may pick some up in the future for comparison.


@Nebulous- This is serious scientific research man, it is not a waste of TIM:D. I had a hard time saying that with a straight face. Seriously though I go through so much TIM nowadays that I buy it by the pound.

So next up I have some Thermal Grizzly aeronaut and some mx-4 to test. I also have some Cooler Master ice fusion and some paste that came with my Scythe cooler. I only had enough AS5 left for one mount so I just put that on order as well.


As I alluded to earlier I will be doing some testing for a new startup called Vostok Thermal Compounds. You can find a link to their facebook page here. I will share my initial results and thoughts on these compounds but will reserve an official inclusion in my results until a retail version is available.
 
The cooler will make a difference...and agree to use the worst cooler you can.

It's like this (thermodynamics 101):

1. The CPU is the source of heat (power...units of watts)
2. There is a thermal impedance between the CPU die and the lid...let's call this Rc (units of deg C / watt)
3. There is a thermal impedance between the CPU die and the heatsink through the TIM...let's call this Rt
4. There is a thermal impedance between the heatsink and the air...let's call this Rh.

The 3 thermal impedance are in series. The temperature of the CPU is the product of the CPU power and the sum of the 3 thermal impedance values...plus ambient.

A better heat sink or TIM will have a lower impedance...the lower the impedance, the lower the temperature. You want the heatsink to be as "poor" as possible to give the highest CPU temps. This will make it easier to measure the difference in the TIMs.

Intel publishes the Rc for a given socket, but you will have a hard time finding the values for the heat sink and TIM.

This is a simple model, but you can research it yourself.

Irregardless, if you switch coolers now, you will need to repeat all TIMs for an apples-to-apples comparison.


 
I look at it this way. I am pushing the core temp to about 85-90c. Going to a crappier heatsink will just force me to drop the overclock to keep the cpu in the 85-90c range as I am not willing to exceed that.

I feel using the better cooler is actually stressing the TIM more than a crappier cooler as it allows me to clock the cpu up resulting in a higher wattage draw. This will then make the TIM have to transfer a higher heat load.
 
Wow there has been a lot of activity today. Let me take this one by one.

@Nebulous- This is serious scientific research man, it is not a waste of TIM:D. I had a hard time saying that with a straight face. Seriously though I go through so much TIM nowadays that I buy it by the pound.

I never doubted for a second it wasn't. I just thought since Rain was going to do the same research, you two can compare notes and share the knowledge. Nothing funny about that.

I've had my share of testing different TIMs over the years, trust me.
 
Sorry Nebulous, my humor tends to be a little dry. I was trying to be funny there, I think I failed.
 
I look at it this way. I am pushing the core temp to about 85-90c. Going to a crappier heatsink will just force me to drop the overclock to keep the cpu in the 85-90c range as I am not willing to exceed that.

I feel using the better cooler is actually stressing the TIM more than a crappier cooler as it allows me to clock the cpu up resulting in a higher wattage draw. This will then make the TIM have to transfer a higher heat load.
Agreed...

I could be missing some fundamental things here too... but, I'd imagine it best to get as much heat as one can get from the source (cpu) THROUGH the TIM to the heatsinks. While there are other mediums to go through die->tim->ihs->TIM->heatsink, I don't believe those to be relevant here since they are a constant. The only variable is what is being tested, the TIM between the IHS and heatsink.

I would much rather see the TIM tested with a more appropriate heat load for the situations we use it...instead of forcing a higher temp with less load.

Also in the general temperature range we use it, are the properties of any of these pastes going to be different? Meaning, at 80C and 40W or 80C at 100W the TIM is still moving the same amount of heat, regardless of the source...It's thermal conductivity (W/m-k) wouldnt have changed at the same temperature. Assuming that is correct (likely my undoing here, lol), it really wouldn't matter either way, would it? In other words... I'm not testing a lighter (low watt) with a yellow flame (max temp), we want to see a bonfire (more watts) with the yellow flame (max temps)....but does it matter in the end??
 
Last edited:
If you test the Liquid Metal TIMs, test those last they are a pain to get off.

Believe it or not after all the testing I did, I returned to using Old Faithful, "Artic Silver 5".

Because of it's long curing time, it did not perform as well as the TIMs that require zero cure in time, but actually tested within 1c to 2c of the best of the best.

However once the Artic Silver 5 was fully cured, it's cooling performance improved as much as 3c ~ 5c actually beating the best of the best that required zero cure in time.

Since TIM is not something I change daily it just made more sense to use the AS5 as it is a fantastic product, and one of the only TIMs that actually contain Silver, (Ag).

Silver is one of the highest conductive elements when it comes to thermal conductivity.

http://environmentalchemistry.com/yogi/periodic/thermal.html

Just to share this information with you I had an interesting experience with the owner of Innovation Cooling regarding their submitting a test sample of IC Diamond.
He refused to submit a test sample unless I agreed to guarantee testing his product using 50lbs of clamping pressure, he told me he would supply test pressure paper that I would use, and then send back to him to be laboratory tested to see if I was using enough clamping pressure. This was already becoming an inconvenience as submitting to his demands was going to extend the testing time waiting on the pressure verification. After much consideration I decided I would agree to his terms to assure a 50lb clamping pressure on one condition.

That condition was that if his product was required to be tested at 50lbs of clamping pressure, I was also going to test the other submitted samples at the same 50lbs of pressure. He argued the other TIMs did not need 50lbs of pressure and that much pressure would thin them to literally a micro layer, I responded, "That's exactly right!, but it is only fair to the other products to test everything exactly the same.".

He refused to submit a testing sample and even though I already had a tube of IC Diamond 24, he would not give me permission to include it in the Thermal Roundup.

http://www.innovationcooling.com/application.html
 
Last edited by a moderator:
Back