[Skinnee Labs] The TIM Thread...

[Vapor]

Do you have any Thermalright Chillfactor 3 to test? It came with my TR Venomous X and I don't feel like spending $10 for some new paste

Thermalright CF2 and CF3 (and maybe the original CF) are scheduled for testing right after AS5/Ceramique/Ceramique 2, which are currently in testing.

Vapor: if manufacturer's suggested method of application is clearly inferior, why use it instead of your time proven bead in center method? Imho this roundup is more about TIM paste relative performance, not how bad one might perform applied by manufacturer instructions?
(imho one might do few test mounts of manufacturer suggested methods just to check if some particular paste really isn't better off with them, but as soon as it shows on average worth results - revert and perform all the tests with correct application)

I follow the instructions--manufacturers should know their product better than I do and should be able to recommend the best way to use it. I'm not going to make a blanket statement and say "don't manually spread paste." While I can't think of a scenario where spreading it would be superior, but there may be one.

Indigo Xtreme has instructions that need to be followed, for Arctic Silver pastes I use the instructed line method (which may be better than the bead method), and the nail polish bottle pastes (e.g., Zalman STG1) I'll paint on as instructed. If there are no instructions provided (including a mini-spatula is not the same as providing written instructions), I just do my normal bead application. Basically, the MO is to follow instructions but if there are no instructions, use the bead method and if the instructions can't be followed, just make it work (Z9 was an example of not being able to follow instructions).

As for testing both instructed method and bead method, each TIM already takes 7.5 days and I have roughly 30 TIMs left to test. That means the testbed needs to survive into 2012, which is a lot to ask (this testbed has already outlived all my previous testbeds in terms of at-load hours). Considering the durability concerns and the amount of time each TIM already takes, I'm not looking to slow things down by double testing some pastes.

At some point I may 'borrow' the CPU block testbed and test TIM application techniques, but that depends on the CPU block testbed freeing up.

[little_scrapper]

Does anyone else spread their TIM over entire surface with card/plastic bag+finger?

I did in last build. But wondering results in spreading vs drop in middle. Or the X.

Which is best way in your opinion?

The answer seems to depend on the viscosity of the TIM

http://www.youtube.com/watch?v=EyXLu1Ms-q4&NR=1
http://www.youtube.com/watch?v=ffK7L...720&feature=iv

[Vapor]

The answer seems to depend on the viscosity of the TIM

http://www.youtube.com/watch?v=EyXLu1Ms-q4&NR=1
http://www.youtube.com/watch?v=ffK7L...720&feature=iv

Only things I can tell from the first video are that manually spreading introduces air pockets and that there isn't enough paste being used for AS3

From that, and I wouldn't put a huge stock in it, it looks like pea/line/rice seem to be pretty good, no potential air traps with those examples.

What you want from a TIM application is to ensure there are no air pockets and to use as little as possible but still get near-full coverage. Air pockets depends on pea vs. X vs. spread, et al. and lift-off and tilting during mounting. Near-full coverage depends on viscosity and how much is applied.

[Vapor]

Updating the first post now with Arctic Silver 5, Ceramique, Ceramique 2, Thermalright Chill Factor, Chill Factor 2, and Chill Factor 3

[defect9]

It's interesting to see the difference (in this weighted comparison) between the best performer to the non-fail performer is still less than 10c.

[Captain_Harlock]

Thanks again for another great round of testing Vapor! Looking over your failed TMP-1000 test, it reminds me of the original Arctic Silver (1) line. Looks exactly the same from your pics, and I would use the same description for it when it comes to removing it. But it did perform very well for its time. Curious as to just what made the TMP-1000 suck so much?

[Vapor]

Not sure what about TMP-1000 made it so bad, but AquaTuning Silver Grease is a similar-looking paste and performs even worse. TIMs have come a loooong way since AS1, too, wouldn't be surprised if it failed miserably.

I have a hunch EVGA Frostbite is based on or related to another silvery paste. No idea how the Homight paste performs...nor EVGA Frostbite. Point is, the silvery pastes have me wary.

[Captain_Harlock]

I'm still lovin my Liquid Ultra. I love my IX setups in the past, but limited applications and price simply isnt worth it if you change blocks to much. I could care less about the color leeching from Liquid Ultra (as I tend to never sell/trade my chips anyway), but 4-5 apps per tube and quick to clean and apply makes it well worth the price to performance ratio to me. Cant wait until you add it to you charts. Maybe next year. ^_^

[Martinm210]

It's interesting to see the difference (in this weighted comparison) between the best performer to the non-fail performer is still less than 10c.

Yeah, but that's more than the difference between best to worst water cooling block....

Awesome work as always!

[Vapor]

I'm still lovin my Liquid Ultra. I love my IX setups in the past, but limited applications and price simply isnt worth it if you change blocks to much. I could care less about the color leeching from Liquid Ultra (as I tend to never sell/trade my chips anyway), but 4-5 apps per tube and quick to clean and apply makes it well worth the price to performance ratio to me. Cant wait until you add it to you charts. Maybe next year. ^_^

I can't wait till CLU/CLP are added too, means I'll be done

Changing the test schedule I think....the Tuniq trio can wait, going to test EVGA Frostbite, Phobya HeGrease, and TC Quantum next (after IC Perihelion, OCZ Freeze, and Noctua NT-H1), I think.

[fgw]

I can't wait till CLU/CLP are added too, means I'll be done

Changing the test schedule I think....the Tuniq trio can wait, going to test EVGA Frostbite, Phobya HeGrease, and TC Quantum next (after IC Perihelion, OCZ Freeze, and Noctua NT-H1), I think.

your going to test OCZ Freeze? used it and was happy with it, but cant find it here (europe) anymore and was told its already discontinued. is it still sold in the us?

[Vapor]

your going to test OCZ Freeze? used it and was happy with it, but cant find it here (europe) anymore and was told its already discontinued. is it still sold in the us?

Freeze has been EOLd but got enough requests that I figured I might as well Never know if the paste comes back to the market under a different brand and people who do currently use it could use some reference. Testing's done, updating OP now

Freeze and NT-H1 were very good, IC Perihelion was very bad overall and especially bad on CPUs where contact is usually pretty good. Also of note: IC Perihelion had a sickening smell of fish

Average Temperature Weighted for Overall Performance Impression:


Average Temperature Weighted for Typical CPU Mounting Quality:


“Overall Performance Impression” takes into account all three contact settings, “Typical CPU Mount” takes into account Moderate Contact, Great Contact, and Best Mount temperatures. ”Overall Performance Impression” is an average designed to describe the overall performance of a TIM where “Typical CPU Mount” is an average designed to describe expected performance with CPU blocks and heatsinks, where contact quality tends to be good.

[Hondacity]

no phobia yet i see, i think its same like the gelid xtreme. same packaging, same specs, same color and consistency.

[Vapor]

no phobia yet i see, i think its same like the gelid xtreme. same packaging, same specs, same color and consistency.

Have both HeGrease and GC-3/Extreme, haven't tested either yet....just did one of my quick tests to tell pastes apart and I can't notice any definitive differences. So they may well be the same, good call there.

Of note, I also can't tell apart Nexus TMP-1000 and Homight (seems to be the generic of EVGA Frostbite), which may not be a good thing for Frostbite

[Vapor]

First post updated with Shin-Etsu G751, X23-7762, and X23-7783D results added They didn't fare that well, surprisingly.

Also detailed the queue more

For the lazy who don't want to go back to the first post, here's the two most pertinent charts (IMO)...

Average temperature weighted for overall performance impression:


And average temperature weighted for typical CPU mounting:

[Hondacity]

oh noh vince won't be happy, lol

how did you apply the frostbite? blob ? line? peanutbutter spread? i tried looking for your method, can't seem to find it.

[Vapor]

IIRC, application method is on the 2nd page of all the individual reviews, below the test spec. But basically, if there are instructions included/readily available for a TIM, I follow those. If there are no instructions (including a mini-spatula is not the same as instructions), I do the smallest blob possible that spreads over all/nearly-all of the IHS.

Frostbite/HeGrease/Quantum is already in progress and Frostbite has already officially failed with Poor contact

[Hondacity]

yeah frostbite is hard to use, if you mount it a little sideways you won't have good contact and it blows. and too much of it sucks soooo bad. but its good if the mount and amount is perfect

[Vapor]

Well, failing means temperatures over 90C and, like TMP-1000, these were ~95C (and quickly). Shin-Etsu X23-7783D managed 74C and Prolimatech PK-1 managed 75C with the same test conditions, so Frostbite is off to a very bad start.

A couple weeks ago I did some quick (2.5hr, not 10hr) 3 mount tests with an Apogee XT (somewhere between Moderate and Great contact settings) of Frostbite vs. Nexus TMP-1000 and there was no appreciable difference between them. Considering they also have indistinguishable physical features, I wouldn't be surprised if Frostbite and TMP-1000 were the same (or very closely related/copied). No amount of 'getting it right' was able to fix TMP-1000 so Frostbite is really not looking too good for air/water use.

[Church]

Wierd. I somehow expected 7783D to fare a bit better. Not overtaking PK1, but still better. But it loosing to MX2 in both moderate/great contact, i wonder why it did better then that in old TIM roundups? Maybe old testbed had closer to poor contact? Or older paste batches were better?

[Vapor]

The old AMD testbed might have had closer to Poor contact (AMD IHSs don't tend to be contact-friendly)...but the change between IX vs. 7783D (got worse with the new tests) compared to IX vs. MX-2 (fairly constant across tests, IIRC) indicates 7783D seems to have changed, so could be a batch variance. Also, MX-2 beat IC Diamond, a good indicator that contact was pretty good on the Intel testbed (I know from pre-testing that ICD excels with Poor contact)

In addition to batch variance, a few other possible explanations:
1) There's actually two 'breeds' of Moderate contact and I only test one and the Apogee GTZ from the first Intel is pretty typical of the other. Type1: flat base, moderate pressure + bowed IHS. Type2 (Apogee GTZ): bowed base, moderate pressure + bowed IHS. They're both "Moderate" contact in that they tend to be a couple degrees off Great contact but different pastes will respond differently depending on their bulk conductivity and how well they thin under pressure. This doesn't really explain why 7783D performed so well on AMD, though.
2) Maybe some difference in adhesion to nickel (CPU-360) vs. copper (Apogee GTZ) causing a thicker bondline? This doesn't really explain why 7783D performed so well on AMD, though.
3) Some change of unknown origin.

Batch variance has the best fit though and has a precedent with MX-3 having two very distinct batches.

[vhaarr]

Thank you for the amazing work here, Vapor. Some of the worst contenders in your tests so far are the most popular choices here in Norway and having this post to link to when other people recommend Arctic Silver Ceramique or MX-3 is invaluable.

However, this post by paulbagz has me worried about the PK-1 results.

Is it possible that Prolimatech changed their PK-1 paste? Could it be that later batches are of bad quality?

Thank you!

[paulbagz]

Vhaarr, I have wondering myself about "batch" like performance however I have used two seperate tubes and yeilded the same results.

I surely hope I am doing something wrong but alas I cannot see how I am after 8+ mounts using different methods as well as different water cooling gear.

-PB

[Vapor]

Thank you for the amazing work here, Vapor. Some of the worst contenders in your tests so far are the most popular choices here in Norway and having this post to link to when other people recommend Arctic Silver Ceramique or MX-3 is invaluable.

However, this post by paulbagz has me worried about the PK-1 results.

Is it possible that Prolimatech changed their PK-1 paste? Could it be that later batches are of bad quality?

Thank you!

The MX-3 results are due to a bad batch, hence the asterisk on the name (had multiple tubes from a bad batch and a partial from a good batch--in the MX-3 review there's some Good Batch MX-3 numbers, but I didn't have enough to do a complete set with the good batch, just the bad batch). So it is possible PK-1 has a bad batch but extraordinarily unlikely a manufacturer simply changed the paste. No Ceramique batch is very good though

I read over paulbagz's thread and what I'm seeing doesn't look like it has anything to do with TIM. The temps are 45-50C worse than they should be--the worst paste I've ever tested shouldn't do that. I'll post in his thread with more info/questions.

[Vapor]

First post updated with EVGA Frostbite, Phobya HeGrease, and TIM Consultants Quantum results added New 'Best Paste' in HeGrease and Quantum ended up being a solid competitor to MX-2 (in performance and pricing). Frostbite seems to be identical to Nexus TMP-1000; I cannot tell them apart based on physical characteristics and all test results were within .06C of each other (all but one metric were actually within .02C).

Also detailed the queue more

For the lazy who don't want to go back to the first post, here's the two most pertinent charts (IMO)...

Average temperature weighted for overall performance impression:


And average temperature weighted for typical CPU mounting:



For those that want HeGrease but have trouble finding it (not at Newegg, for example), I am very confident that HeGrease and Gelid GC-Extreme are the same and strongly suspect Scythe Thermal Elixer is the same as well.