Thermal Paste Comparison, Part One: Applying Grease And More

[master9716]

This is the type of article you find on Hardware.info , Thanks to Toms now every single company will re-design their coolers .

 

[Calculatron]

Or at least allow optional mounting directions as a standard.

 

[pandawatch]

This is great cause I just flushed my custom hosed H220 and had to reapply AS5 to my 4670k and instinctivly went with a generous line down the rectangular die with plent of mounting pressure. When is it to much pressure?

 

[jbird2383]

what about tinting??? Artic Silver application method suggest you tint both the heatsink and cpu before applying Artic Silver 5 for the microscopic valley ways. I've stuck with Artic Silver 5 and always used a pea sized amount

 

[InvalidError]

When is it to much pressure?

When you crack your PCB, the socket or the solder balls under the socket due to excessive motherboard bending and your computer stops working because of that. With soldered IHS, there is also a possibility of cracking the CPU die.

Unless you use nearly solid paste, moderate yet uniform clamping pressure should make most of the excess paste ooze out over time. Some clamping mechanisms like the Hyper 212 have spring-loaded screws with the screw threads just long enough to prevent the springs from bottoming out, which limits pressure to whatever the springs can provide and make it effectively impossible to over-tighten.

 

[masmotors]

IC Diamond gives me good temps witha 620 h2o and 4670k overclocked onprime does not go over 56c prime for 9 hrs and with my 212+ witha fx 6100 stock it never goes over 36c

 

[glitch20]

Knew it spread method sucks it It looks crazy to cover the surface w/ thermal paste when stock coolers cover round the center lol - gee weeeez the stock cooler that everybody hates is actually doing it right

 

[FormatC]

It says that results are in Kelvin on your actual benchmark results. 30 degrees kelvin would be roughly -405 degrees fahrenheit. I think that is a mistake, or I'm buying that thermal paste.

At first - It is common to specify the difference between two temperatures in Kelvin. (Tcase1 °C - Tcase2 °C = Tdifference K). This is scientifically accurate

The IC Diamond 24 was also tested (but later), so you get the results in our charts and in a follow-up (currently only in German, sorry for this):
http://www.tomshardware.de/Innovation-Cooling-Diamond-24-Review,testberichte-241361.html

 

[smeezekitty]

Who uses Kelvin in the real world? Celsius would be more understandable.

 

[FormatC]

@smeezekitty:
Why °K? It is a question of education and accuracy. But please understand my problem. I can use °C instead of °K because the result is in each case the same (only the unit is different). The difference between two temperatures in °C or K is (not only in the real world) absolutely identical! It is only a difference and not an absolute temperature value.

For you:
Don't merge K (Kelvin) and °K because it is simply not the same! Only K (Kelvin) with out the magic ° stands for an absolute temperature, the °K is the difference between two temperatures (and in our case absolutely correct). But what is more accurate and what is more compatible for the masses? I'll never be able to please everyone.

I've learned it in school and study that I have to use the right unit for each case. Ok, this was 30 years ago and it is possible that the habits have changed. May be, I'm simply too old or living on the wrong continent. In Germany no one has complained about this °Kelvin thing, because it is really clear that this is only a difference and not a temperature (thermodynamics, class 8). On the contrary, because °C would be very ambiguous.

 

[InvalidError]

Why °K? It is a question of education and accuracy.

Accuracy is affected by the precision of your measurement instruments and unit conversion coefficients. The actual units used are of no importance.

Since Celcius and Kelvins are exactly the same thing apart from having different definitions of their zero point, mixing the two simply confuses people for no good reason. IIRC, avoiding mixing up different units is also part of scientific conventions (how many accidents have been caused by people confusing inches for centimeters, kilos for pounds, etc. due to working in environments that mix multiple unit systems? Too many.) so if you want to specify temperature differences in Kelvins, you should write temperatures in Kelvins too.

 

[mamsa]

Something I did not see mentioned was Burn-In-Time for the pastes. I read somewhere this is 200 hours (for Arctic silver). Obviosly I don't expect you to run the machines for 200 hours but still, would be nice if that was mentioned in the article.
"...and its Arctic Silver 5, and does say that temps will drop several degrees over the break in period (~200 hours for AS5)."
Source: http://forums.hexus.net/pc-hardware-components/192609-take-thermal-paste-burn-into-account-when-overclocking.html

 

[FormatC]

@InvalidError:
Again - K and °K is not the same. The old unit for temperature difference was grd, the current is °K, a difference unit, not K. K and °C are absolute temperature units and I know that it is not a good idea to merge this both units in any content. But I had to show the Delta.

@mamsa:
I know the problem with the burn-in-time for this products, but a modern paste is able to work at the begin. I've tested it with three products over a longer time and I got an interesting result. The viscosity of the older procuct increases after lot of hours because a part of the added silicone oil has gone. But is this really better? The Arctic Silver is not bad, but outdated. This is like an old car. If you take a look at the charts you'll see that the difference between all this products is smaller as expected. The most important factor is not the what but the how

 

[InvalidError]

@InvalidError:
Again - K and °K is not the same. The old unit for temperature difference was grd, the current is °K, not K. K and °C are absolute temperature values.

Which is yet more unnecessary confusion: everyone knows what a delta in degrees Celcius is. Most normal people have no idea nor care about what Kelvins are nor that there are two different ways of referring to it just to make things more confusing and nitpicky..

 

[FormatC]

everyone knows what a delta in degrees Celcius is

Really? I'm sure that a lot of people have a problem to understand the term Delta...

May be you're right, but I compared it for fun times with my school books. We've learned °K and not °C. Ok, but I understand now, that the usage of units is in different countries very different too

If we are honest:
If I had written ° C, another would have complained. This is my real world and I'll never be able to please everyone.

To find a compromise: I'll use Δ°C in the future (for our translated content)

 

[nukemaster]

I think many people will like Delta Degrees Celsius for sure. Not that it actually makes any difference. People just seem to be used to seeing it that way.

 

[FormatC]

@nukemaster:
Hmm, may be you're right. I've noticed that the German language is slowly dying too. Too many influence from emigrants, anglicisms, superficialities and something other. It is really hard for me to see, how the world changes May be, I'm too old?

 

[InvalidError]

To find a compromise: I'll use Δ°C in the future (for our translated content)

Good pick. This is a fairly common form.

Kelvins just aren't seen much outside thermodynamics and fundamental physics - even the SI refers to it as the "thermodynamic temperature" instead of just "temperature."

Another reason why I thought Kelvin would be a bad pick: how many people in the USA have a good grasp of what Celsius are beyond being "that other fancy unit some countries use for weather temperature"? Kelvins, although identical apart from its zero-point (the Kelvin's definition story required some mathematical creativity to maintain compatibility with Celsius), would likely be even less familiar.

 

[FormatC]

The most of my articles were written in German and we have here in Germany a lot of very "academic" nitpickers in our forum. It is very difficult to find the best way to make it right for everyone

This is the detriment of all translations. And my own English is too crappy to translate all this stuff myself. I've learned English 30 years ago and forgot it 29 years ago.

 

[Calculatron]

I think the current format is fine; perhaps just add some footnotes (or maybe a link to a page) explaining methodology. We honestly shouldn't be so adverse learning something new, but sometimes when we don't see the method to the madness we get a little grumpy with all the madness.

I suppose that, with you being in Germany, it would be out of the question to offer to send you my tube of Dow Corning TC-5022 to test?

 

[smeezekitty]

how many people in the USA have a good grasp of what Celsius are beyond being "that other fancy unit some countries use for weather temperature"?

I wish the USA would just get out the dark ages and switch to metric already

 

[drwho1]

very good article, looking forward for part 2.

 

[planes]

I dance to my own demons. I take my Hyper 212 and put the solder on the bottom of it. I used arctic silver on my last build, took a credit card and spread a nice amount over the pipes and let excess seep in between the pipes until I get an even layer across what looked like the size of the chip.. Have my 6800k OC's to 4.7 and passed an hour of Prime 95 with low temps. Sure could go higher but I'm fine at this speed.

 

[06yfz450ridr]

Still spread my paste by hand with a plastic bag etc covering my finger. I dont trust it ever to spread evenly but spreading arctic silver sucks its like glue

 

[ihog]

Would a very thin copper shim between the heat spreader and the cpu die fix it up?

The copper would most likely scratch the die. And, even with the shim, you'd still need thermal paste.

I would not be surprised if part of Intel's decision to go paste for mainstream CPUs had to do with soldered CPUs dies getting damaged by shipping rough-handling.

True, but I'd venture to say cost was an even larger factor.