[HardOCP] Thermal Paste Shootout - Q209

[IKIKUINTHENUTZ]

Thermal Paste Shootout - Q209

We take 10 high profile thermal pastes used for mating heatsink surfaces to CPUs and other components and subject those to over 1200 hours of testing to find out which one reigns supreme. Drop or spread? And we tried American cheese too, just in case you find yourself in a pinch.

Source

[W1zzard]

completely wrong way to measure this. they just looked at how intel suggests coolers should be qualified (qualify == keeps the system stable != scientific measurement)

only one application per thermal paste? or did i miss something?

a $60 multimeter to measure temperature? wtflol. that meter has an accuracy of +-1°C

"Ambient temperature will be kept at 25C for the duration of the tests and measured with a MicroTemp EXP non-contact infrared thermometer." - that meter has something like +-2° accuracy...

total that's +-3°C .. as a result those charts are pretty worthless since all pastes are within 3°C of each other.

great review. pick 6 pastes, roll dice, less work. if you have no idea how to review something, stay away from it.

[autobot]

Once again...another wonderful review from the fine folks of [H].

They really should just stick to reviewing simple stuff like cases and case accessories.

[roofsniper]

after the phenom II review i deemed everything there useless. its been awhile since i have read anything from there but i decided to check this review out. and yea im not missing anything. its a waste of time just to read it.

[demonkevy666]

I think they need to use more heat sinks.
just using one heat sink seems kinda pointless anymore.

[belmontrich]

i'll check out the link later.. thanks..

[lkiller123]

completely wrong way to measure this. they just looked at how intel suggests coolers should be qualified (qualify == keeps the system stable != scientific measurement)

only one application per thermal paste? or did i miss something?

a $60 multimeter to measure temperature? wtflol. that meter has an accuracy of +-1°C

"Ambient temperature will be kept at 25C for the duration of the tests and measured with a MicroTemp EXP non-contact infrared thermometer." - that meter has something like +-2° accuracy...

total that's +-3°C .. as a result those charts are pretty worthless since all pastes are within 3°C of each other.

great review. pick 6 pastes, roll dice, less work. if you have no idea how to review something, stay away from it.

Well, for you, how are you going to review it? Measure the temperature using "Realtemp"? Don't criticize something that you can't do..

[zanzabar]

Well, for you, how are you going to review it? Measure the temperature using "Realtemp"? Don't criticize something that you can't do..

do u know who w1zzard is, he makes ati tray tools. and all that i have to do to get that test proper is the proper thermal probes/DMM and a climate controlled room + alot of time.

[lkiller123]

LOL that's true for alot of time, but I guess that's something that [H]OCP lacks D:

[BDW88]

do u know who w1zzard is, he makes ati tray tools. and all that i have to do to get that test proper is the proper thermal probes/DMM and a climate controlled room + alot of time.

He's also the guy who did the EVGA Precision app..

[lee143]

Well, for you, how are you going to review it? Measure the temperature using "Realtemp"? Don't criticize something that you can't do..

The stupidity of this post just made me laugh

[mep916]

He's also the guy who did the EVGA Precision app..

No, AFAIK, EVGA received help from the creator of Rivatuner for the Precision app. Wizzard is still the man, though, and for someone to flame him here at XS is unfortunate, to say the least.

[W1zzard]

Well, for you, how are you going to review it? Measure the temperature using "Realtemp"? Don't criticize something that you can't do..

i cant tell you how i would do it, otherwise i would have done such reviews. and yes i can criticize even if i have no better suggestion.

let's see.
- get a thermometer that measures down to 0.1°C accuracy (not resolution) at least. better is 0.01°C. yes that will be expensive.
- get a proper controllable environment. temperature kept within 0.1°C - solutions exist for that but pricey again. or get a good probe and normalize the temps (not that good)
- have one cpu to measure tcase with a probe (like hardocp did)
- use realtemp on a different cpu, calibrate it using proper equipment (no not an IR thermometer)
- alternatively figure out a way to hook up a thermometer to the on-die diode inside the cpu. possible, i've done it for athlon64
- ensure that the ihs' are evenly flat on both cpus
- test dual core, quad core, i7 cpus (could be 3 reviews)
- use at least 2 heatsinks to qualify if there is a difference
- ensure to always use the same mounting pressure (spring mounting is good for that)
- ensure proper cleaning of the heatsink base, maybe sand it down for every measurement so you have no residue from other pastes?
- repeat application, measurement and cure in 3 to 5 times
- devise a method how to apply the same thickness/spread of thermal paste every single time
- possibly look at several different batches of the same paste.

expect to spend at least a few thousand $ for equipment and several hundred hours to get everything figured out and get results.

one question for you to think about: what is the cpu temperature? its case? where on the case? on the die? where on the die? the package maybe? but where? which thermal gradients exist between those three?
is it better to favour low temperatures or better overclocking potential in thermal pastes? maybe add some oc tests.

[BDW88]

Ummm yep I think that would be the proper way to do it lol

Oh and W1zzard thanks for all the apps you have made couldn't live without them.

[lkiller123]

No, AFAIK, EVGA received help from the creator of Rivatuner for the Precision app. Wizzard is still the man, though, and for someone to flame him here at XS is unfortunate, to say the least.

I didn't mean to flame him, I was just baffled by the way he was criticizing [H]ardOCP.. I respect his knowledge and contribution to the enthusiasts, but hey, everyone can make mistakes and not everything is perfect. I agree with W1zzard that the review contains errors and it is inaccurate, still, [H]OCP is a great thing

i cant tell you how i would do it, otherwise i would have done such reviews. and yes i can criticize even if i have no better suggestion.

let's see.
- get a thermometer that measures down to 0.1°C accuracy (not resolution) at least. better is 0.01°C. yes that will be expensive.
- get a proper controllable environment. temperature kept within 0.1°C - solutions exist for that but pricey again. or get a good probe and normalize the temps (not that good)
- have one cpu to measure tcase with a probe (like hardocp did)
- use realtemp on a different cpu, calibrate it using proper equipment (no not an IR thermometer)
- alternatively figure out a way to hook up a thermometer to the on-die diode inside the cpu. possible, i've done it for athlon64
- ensure that the ihs' are evenly flat on both cpus
- test dual core, quad core, i7 cpus (could be 3 reviews)
- use at least 2 heatsinks to qualify if there is a difference
- ensure to always use the same mounting pressure (spring mounting is good for that)
- ensure proper cleaning of the heatsink base, maybe sand it down for every measurement so you have no residue from other pastes?
- repeat application, measurement and cure in 3 to 5 times
- devise a method how to apply the same thickness/spread of thermal paste every single time
- possibly look at several different batches of the same paste.

expect to spend at least a few thousand $ for equipment and several hundred hours to get everything figured out and get results.

one question for you to think about: what is the cpu temperature? its case? where on the case? on the die? where on the die? the package maybe? but where? which thermal gradients exist between those three?
is it better to favour low temperatures or better overclocking potential in thermal pastes? maybe add some oc tests.

Yes, that's the way, and maybe you can test it a few more times to ensure the average temperature is correct. But consider that the OCP is a company, it has a limited budget and a limited amount time. I am sure that anyone can think of that idea if they have loads of money and time to spend on. I respect you, and I did not mean to flame you in the first place.

[W1zzard]

But consider that the OCP is a company, it has a limited budget and a limited amount time. I am sure that anyone can think of that idea if they have loads of money and time to spend on.

then they should review something different that they can do with their skills, time and money budget.

what would you say if i reviewed a vga card and each other card in the comparison group was tested at a different cpu frequency?

[guess2098]

I like this :P

American cheese is the name given to cheese that goes through a specific type of pasteurization process. Often yellow or orange in color, this cheese has also been known to take on a white color. A versatile type of cheese, tracing its origins back to the very start of this great nation, American cheese has found uses in barbeques and kitchens all over the country. Finding detailed specs has been no easy feat so without measurements in specific gravity or thermal conductivity we will be testing purely on its performance.

[Retro]

The stupidity of this post just made me laugh

Yep, me too

[Warboy]

Cheese FTW.

They didn't test coolermaster thermal fusion 400?

[gr8golf]

I didn't mean to flame him, I was just baffled by the way he was criticizing [H]ardOCP.. I respect his knowledge and contribution to the enthusiasts, but hey, everyone can make mistakes and not everything is perfect. I agree with W1zzard that the review contains errors and it is inaccurate, still, [H]OCP is a great thing

Since you are insisting on such, please provide a quantified, scientific explanation of how [H]OCP is a great thing. Your explanation should certainly detail how they are able to post so f'n much advertising on their site and remain *objective*.

[jmke]

pick 6 pastes, roll dice, less work. if you have no idea how to review something, stay away from it.

main reason why I've not touched thermal paste testing for time being, test setup has be overhauled to actually be able to measure the differences and be accurate enough to tell. not very easy when on a budget