My second thermal compound comparison.

[WarCon]

I was contacted by Scott Gallmeyer from Nanotherm and he asked if I would like some of his new Nanotherm to test. I said ok, and I have already started. He sent along a tube of the original formula and a tube of the new formula. I am going to test these two and a selection of Arctic Silver products (Arctic Alumina, Arctic Silver II and Arctic Silver 3), sorry to you Radio Shack fans(not going to risk my processor this time with such an inferior product). This test is going to take a while since I give each product a minimum of three days running. I take measurements after I first put it on and then after 72 hours to allow for the break-in period on all of the compounds. Mind you this is in no way a longevity test for these compounds.

The test procedure is as follows:
I lightly lap my waterblock between each compound, in order to make sure that I am testing only the compound I just applied. I thoroughly clean both surfaces with isopropyl alcohol and let dry. I apply the compound to the waterblock and then remove most of it with my spreading tool and apply it from there to the core. I then wipe the excess off of the waterblock, but not with any solvent, just a clean "as lintfree cloth as I have" and then mount it with a lexan holddown that uses the motherboard mounts. I am measuring the temperatures with the socket thermistor and a side core mounted thermistor. I measure the ambient in front of the watercooler's radiator as the air is being drawn in. I use the thermistor on the motherboard for system temps. I measure after 1/2 hour of the system being in bios, 1/2 hour at idle with minimum processes running, and 1/2 hour full load (using Prime95 and CPUBurn 4). I measure right after I apply the compound and then later after the compound has been on my system at least 72 hours. After I finish the last set of measurements, I then start pushing my overclock to measure the highest stability for that compound.

I removed all of the extraneous data, so that the results wouldn't be confusing........

Here is a link to an HTML page that has the data in a Table format........(Nevin was kind enough to host the link, thank you Nevin......)

CPU temps adjusted to a nominal 20C and again listed as initial/final.
*Note: Socket temps are rounded to whole numbers as to not change accuracy.
...........Nanotherm (old)........Nanotherm(new).........Arctic Alumina............Arctic Silver II...........Arctic Silver3
BIOS
Socket.....37C/36C....................36C/
Sidecore..33.0C/33.4C..............33.2C/

IDLE
Socket.....26C/25C....................26C/
Sidecore..24.6C/23.4C..............24.6C/

FULL LOAD
Socket.....43C/43C....................43C/
Sidecore..38.2C/38.6C..............38.1C/

O/C(stable)..1430.......................


So far the two Nanotherm's are testing nearly identical.

 

[FlowerMan]

I always look forward to reading your comparisons, but for the life of me I could not understand your numbers because of the weird forum formatting... got any Excel tables and graphs? I'm sure Nevin or Scott can help you out with hosting jpg's of the graphs if you need a host, because Nevin hosted pics for my comparison over at the HardForums

 

[WarCon]

I will go ahead and make an Excel chart and see if someone will host it for me. If anyone wants to volunteer to host it just let me know. I will need to update it every 3 days or so.

Anyway the numbers are showing that the old nanotherm and the new are performing the same so far.

 

[FlowerMan]

Like I said, just email Nevin or Scott and they will provide you with pic hosting

 

[MrThompson]

More good reading. Thanks WarCon. Looking forward to the rest of your results.

Regards,

Colin

 

[Lizardman]

so what compounds have you tested so far? I just can't tell from your numbers.

 

[FlowerMan]

 

[WarCon]

I have tested the old Nanotherm both initial temps (when I first put the compound on) and final temps. And I have the initial temps from the new Nanotherm. The rest will be coming in every 3 days or so each. Maybe I should of finished test before posting.................nah

Is the chart that far off for you guys? It displays pretty linear for me. I will try to get some webspace to post an Excel chart as this goes along.

 

[FlowerMan]

oic, you havent finished testing... no wonder why the chart looked like it was missing something n/m, looks fine to me

 

[WarCon]

Bump - added link to table.............

 

[WarCon]

Link to table is now working.......

 

[FlowerMan]

bump

Nice table. Looks like the new formula doesn't really outperform the old one by much, eh? I didn't expect much of a difference, since Nanotherm was pretty good the first time around Now I'm just looking foward to your AS3 results... please test that one next!

 

[gdawson6]

Make sure you let the Artic Silver III have like 3 days after its applied till you take the readings, The people who make it says it takes like 72 hours to fill in all the spaces until it works the way you expect.

 

[FlowerMan]

<< I take measurements after I first put it on and then after 72 hours to allow for the break-in period on all of the compounds. Mind you this is in no way a longevity test for these compounds. >>



Dude, this is WARCON you're talking to You know, the "Ultimate Thermal Compound Reviewer"!

 

[WarCon]

Bad News for anyone thats concerned. I was having problems with my install of Arctic Silver 3 and was reapplying the compound today and I busted the wire on my sidecore thermal probe. Anyway after soldering it back together it is no longer calibrated nor is it linear anymore. Since it was a major portion of this test, I am stopping my testing for the foreseeable future.

Anyway, I was so discouraged from that, that I put on ASII and don't plan on changing it for awhile. Heck who knows what I will break next time........

If I can scrounge the money for a new compunurse, then I may rerun the tests at a later date.

Anyway it was looking like the new Nanotherm wasn't much (any) different than its original formula. It actually tested at only 1420, where the original ran 1430.

Sorry,

Warren

 

[gogeeta13]

Shoot man, and I know how much time you had invested in this=\

Well, kudos to you for doing this for a bunch of forum people, you deserve more credit then people give you. If I was still with some tech sites, I would get you a job there!

Thanks man!

 

[FlowerMan]

Sorry to hear that. Maybe you should continue testing with simply the in-socket thermistor? Perhaps not all of your time and effort would be lost

 

[WarCon]

For giggles, I may run highest overclock tests on the remaining compounds, but I still want to give the current compound at least 72 hours before I do that.

I will most likely scrounge to make a test bed, instead of using my computer all the time. It won't give temps that are related to a processor, but should show in a much less compressed manner the difference between compounds (And the way I am planning on making the test bed, it will let me test heatsinks too.)

That is at least a month or so down the road though.

gogetta13, thanks for the kind words....