[OFFICIAL] Free ICD7 thermal compound for OCF members Results Thread

[JoeC]

Where was the majority of these "samples" taken from? All 3 forums listed on the front page?

? Just the results from this thread of those that reported before and after Temps. There are 3 zero or no change results that show up as spaces on the graph

Raw data can be found here http://spreadsheets.google.com/ccc?key=pYFDF5y415rhMLI2m-uK4Zw&hl=en

 

[omniskillz]

sorry for the late report.
but i used ic7 on my new setup.

idle temps i can only give what my bios shows me after i do a nice load run cause for whatever reason the new intel e8400 has cpu temp stuck at 36 using coretemp and using my motherboard software it tells me my idle temp is 47.... way off.
i used a raytek infrared temp guage on cpu block and added 5 degrees, seems to be pretty accurate from my testing with it.

any way
d-tek fuzion waterblock blackice xtreme 3 rad, d5 pump.
e8400 running with 1.5v @ 468fsb 4.2ghz

room ambient temp 27
ic7 as5

idle 34 35
load 48 50

i used ic7 first on the cpu and block. using 91% isopropyl it was hard to get off. block and cpu still had residew on them.... what kind of cleaning salution do i need for ic7? usually as5 is no problem.

plus ic7 material seems pretty dry , thick and pasty.. haha not complaining.

discovered i have a leak in my radiator nozzle so going to disassemble and try ic7 again.

 

[JoeC]

Troubleshooting Tip #1

I came across one non obvious issue the other day while while testing a new sink where my application of ICD7 was performing less than expected. I replaced it with a thinner "wetter" grease and got the result of both pastes being essentially equal.

The only difference than my normal setup was the mounting hardware where I usually use heavy duty hardware to ensure that I reach the optimum BLT on every mount with every test.

I scavenged some higher tension springs from another heat sink and retested and the problem was corrected and I was back to the desired result.

Highly bulk loaded compounds Like ICD7 need higher pressure and a good solid mount to reach optimum performance versus wetter or lightly loaded (under 90%) compounds.

A wetter grease will reach an optimum at much lighter lighter loads But will be more subject to bake out and pump out affecting long term stability.

shimming when applicable or higher tension springs may be required to reach optimum performance.

 

[Liquid3D]

Joe your preaching to the pasty faced choir now

Via long-term testing running my Q6600 at 9x334FSB (1.28V) = 128W ~ 9x420FSB (1.450V) = 205W I have emprically verifed your statement (at least I hope to show this in my upcoming addendum to my Spreading it Thin TIM Round-up.

I discovered thermal compounds such as IC-7, with higher conductivity / higher viscosity as compared to the recent rapid curing pastes, can be summarized in those Lite Beer commercials of 1970's. The exception being the panegyrical salutations today claim: "More fillers - less particles" (conductive particles).

Products like Tuniq TX-2 seem to rely more heavily on the catalyst which essentially transports the thermal particles to the gaps and striations. This aspect of the compound is also known as the vehicle or filler. These pastes seem great because they show their potential immediately as they are less viscuous with rapidly "evaporating" fillers (to use a poor description for what is essentially curing).

I tested TX-2 for a month and for the last month IC-7 on a D-Tek WB with a Q6600, basically overclocking this chip as high as 3.78GHz. I have been trying to reach the 4GHz plateau. In order to expidite cure time I have been doing two things, first I've been checking and tensioning the mounting springs and increased tension to the point the springs are almost maxed out. The second method which I have found improves the curing process needs a little introduction.

It was some years ago I befriended Nevin House of Arctic Silver whom taught me alot about thermodynamics. When AS5 came out he suggested after a week or so of testing, grabbing hold of your HSF and rotating the block on the processor just a few mm's would expedite the curing process. Obviously this can only be done very carefully and your only going to be able to "rotate" the heatsink on the paste and CPU IHS so much.

I tried this with AS3 and AS5 and the results were that AS3 not only out-performed AS5 it lasted much longer. Nevin claimed AS5 improved with time, I found this was true after follwing his instructions but only for about 60-days. With AS3 however, the effects of rotating or rather shifting the heatsink improved the curing time and that AS3 did perform better over a longer period of time.

I am now finding this to be true with IC-7. I find IC-7 is a paste to which the expression: "Less fillers - more particles" applies. And for these reasons, just as you stated pressure and an even increased pressure as the paste cures in the first, say 48-hour break-in period, is even more important.

Since I didn't have IC-7 in time for my Round-up the addendum or extended Round-up will dicuss viscosity and application. I will reiterate how important these factors are, especially the all important element of Time...

 

[HawaiiSuperman]

Hello everyone-

Having temperature issues I switched from a as5 to IC7 yesterday. These are my idle results.

3 hours idle in xp with as5 9 month old application
cores-1-4: 52,47,47,47
10 minutes idle after new application of as5
cores 1-4: 54,47,47,47
as5 removed, IC7 slapped on, 1 hours idle
cores 1-4 47,42,42,42

Qx6700, 2.6ghz, 1.240 Vcore

The temperature difference with IC7 is great but I must have screwed the application because these temps are still way too high for my low Vcore. Ran out of IC7 so I'll have to re-order some and wait FOREBBER for it to get to Hawaii. My temp difference is absolutely shocking between as5 and IC7. Hopefully it will last long term and I'd have to say from personal experience that the advertised reults arent bull***.

?JoeC, I'm guessing you are the creator? If so, good work, will the temperatures drop as time goes on or should I just go ahead and re-order more IC7 for another application?

 

[Audioaficionado]

to the forums HawaiiSuperman

Before you blame the TIM, how well is your "San Denki San Ace Sealed Liquid Cooling" system working and how good is your case ventilation and ambient temps?

Also what are your full load temps?

 

[JoeC]

Pasty Faced Choir

Liquid3D

Very clever to have worked that one out on a desktop ,takes some persistence and dedication.

IC Diamond thermal resistance from 65lbs down 40 lbs per square inch doubles.

With the larger sinks in use these days it did not occur to me that some small % would be lightly loaded. I will be interested in your test results when they become available.

 

[Omsion]

as well.

Those temps...are pretty high for idle, and defenitly seems to indicate that something is amiss. What are they like at load?

 

[HawaiiSuperman]

to the forums HawaiiSuperman

Before you blame the TIM, how well is your "San Denki San Ace Sealed Liquid Cooling" system working and how good is your case ventilation and ambient temps?

Also what are your full load temps?

9 months ago and on as5 with the sanyo denki, my temps were 36C idle and about 47 under load at 3.2ghz with the 1.35Vcore, just about daily. About two months ago the temps started creeping higher to what they were before the intitial TIM replacement. The sanyo denki liquid cooler did a great job and is rated for 40,000 hours. Case ventalation is optimal.

 

[jivetrky]

I finally did my testing a couple of days ago, I've been reluctant to post them because I think I HAVE to have some kind of an issue with my HSF or something. When I first OCed this system, my temps maxed at around 62c. But that was at 3.56Ghz and 1.55v vcore.
And after reading JoeC and Liquid3D's posts I think that my problem is quite likely mounting pressure. I may try a thinner TIM and see if that helps me, but I want to get better springs or something for my Tuniq. I think my issue is that my springs on my Tuniq are not nearly strong enough (ones that came with it). I have all four of my bolts bottomed out. My top bracket is bowed, so I assumed there was enough pressure.
But I think that it's likely my high temps are because of this. I'll go ahead and post my results, but you can certainly expect they aren't true to what they should be. When I first applied AS5 a year or so ago (or whenever it was that I built this machine) my temps were better. I think the springs have just worn out. I am going to take a run to the HW store tomorrow and see if I can find some replacements.

my results so far:

25c ambient in all tests (Measured with a close range average of all four temp probes on an Aerogate 3)
3.4Ghz @ 1.456v on IDLE ~1.438-1.440v on LOAD (Actual voltage read in CPUz)
Tuniq Tower 120 - Tuniq Fan on maximum
Temps measured with Coretemp v0.96.1

AS5 ~6 month cure:
IDLE: Core0: 46c / Core1: 45c
LOAD: Core0: 64c / Core1: 63c

AS5 fresh application (no cure):
IDLE: Core0: 46c / Core1: 46c
LOAD: Core0 65c / Core1: 64c


ICD7 1^st and 2^nd application identical in results:
IDLE: Core0: 46c / Core1: 45c
LOAD: Core0: 65c / Core1: 65c

Only thing I can figure now is that perhaps I wore out the springs even more after the AS5 testing and the HSF really isn't on there that tight. Most people seem to be getting a few c drop in temps, I didn't

But I would NOT recommend including these findings in any database as I think my HSF is to blame. I'll give it another try when I get some better springs or just a better method of holding the thing down.

 

[maxfly]

i havent had much time to give this stuff a try since getting my sample.until yesterday that is.i replaced some four year old as3 and these are the results from a p4 2.8 533 1.5v @stock.

as3 idle-44c
as3 load-58c

ic7 idle-34c
ic7 load-42c

despite the age of the as3 i think thats a pretty solid drop in temps.ill definately be using the ic7 in future builds.

 

[Liquid3D]

Liquid3D

Very clever to have worked that one out on a desktop ,takes some persistence and dedication.

IC Diamond thermal resistance from 65lbs down 40 lbs per square inch doubles.

With the larger sinks in use these days it did not occur to me that some small % would be lightly loaded. I will be interested in your test results when they become available.

Thank you. It's your constructive criitism which has been most helpful.

Ideally I would like to have a simulated die which is both variable in wattage as well as re-producing the same heat "signature" a Quadcore might, even though it's through a IHS. Of course as far as the heat-signature, I have no doubt it would be disputed, although I would still insist the sim-die itself actually place the heat source below a real IHS taken off a Intel Q-core, concave shape and all....

The second element would be the ability to adjust pressure via lbs per square inch...

It's interesting with a TIM which requires more or less pressure to work it's best that there is so much variance when we think of the springs included with water-blocks. I've never seen two sets from different vendors which were exactly alike (or even close).

Not to mention the hardware supplied with 600g ~ 900g Air Cooling Towers some using push-pin mounting hardware. Once that Tower is mounted onto the mobo, now horizontal in the case, there's the distinct possibility the leverage exerted on the "lower" half of the IHS is substantially greater then the "upper" half.

In fact I can simulate this myself purposely as you had pointed out in my photos taken of "TIM spread patterns" after speciffic application methods.

 

[JoeC]

Ideally I would like to have a simulated die which is both variable in wattage as well as re-producing the same heat "signature" a Quadcore might, even though it's through a IHS. Of course as far as the heat-signature, I have no doubt it would be disputed, although I would still insist the sim-die itself actually place the heat source below a real IHS taken off a Intel Q-core, concave shape and all....

The second element would be the ability to adjust pressure via lbs per square inch...

It's interesting with a TIM which requires more or less pressure to work it's best that there is so much variance when we think of the springs included with water-blocks. I've never seen two sets from different vendors which were exactly alike (or even close).

Not to mention the hardware supplied with 600g ~ 900g Air Cooling Towers some using push-pin mounting hardware. Once that Tower is mounted onto the mobo, now horizontal in the case, there's the distinct possibility the leverage exerted on the "lower" half of the IHS is substantially greater then the "upper" half.

In fact I can simulate this myself purposely as you had pointed out in my photos taken of "TIM spread patterns" after speciffic application methods.

My view is that a simulated die is just that. It is a simulated die with simulated shortcomings(makes sense?). My 1" die gives me a C/W pretty close to what you see in system but it does factor out a lot of the real world variables, Pressure, case flow, horizontal mountings, heat sink airflow etc.

A simulated die works best against a real world reference point, a kind of check against system variables. No one set up gives you the total picture IMO What is non obvious in one set up becomes noticeable in another.

Optimally the Intel method of filing a groove in the IHS for a thermocouple with .1 of a degree resolution on center of the CPU provides a third reference point. Intel has recommended this in the past in any testing that requires a higher level of confidence such as qualifying a heat sink or thermal paste. Internal diodes have their own issues so a thermocouple IHS mount is a valuable addition in the arsenal as well as good compromise if no simulated die is handy.

Averaging results as we have been doing in this promotional exercise has been valuable as a kind of 4th overall view combined with a 5th set with the ASTM lab tests Provide enough information to confirm results and where anomalies emerge identify or enhance troubleshooting the problem.

All combined we have a pretty big hammer to hit the nail on the head

 

[Liquid3D]

Agreed

 

[JoeC]

I finally did my testing a couple of days ago, I've been reluctant to post them because I think I HAVE to have some kind of an issue with my HSF or something. When I first OCed this system, my temps maxed at around 62c. But that was at 3.56Ghz and 1.55v vcore.
And after reading JoeC and Liquid3D's posts I think that my problem is quite likely mounting pressure. I may try a thinner TIM and see if that helps me, but I want to get better springs or something for my Tuniq. I think my issue is that my springs on my Tuniq are not nearly strong enough (ones that came with it). I have all four of my bolts bottomed out. My top bracket is bowed, so I assumed there was enough pressure.
But I think that it's likely my high temps are because of this. I'll go ahead and post my results, but you can certainly expect they aren't true to what they should be. When I first applied AS5 a year or so ago (or whenever it was that I built this machine) my temps were better. I think the springs have just worn out. I am going to take a run to the HW store tomorrow and see if I can find some replacements.

Thread on Anandtech

About Tuniq mounting issue's that might help

 

[JoeC]

Troubleshooting Guide

We added a Troubleshooting Guide to our site with a Pressure vs C/W chart as a guideline as a result of some discussions on this thread.


Troubleshooting Guide

1. Clean the CPU thoroughly to remove all previous paste applications completely. Any residue paste, lint, dust, etc and your results will be adversely affected.

2. Follow IC Diamond 7 Application Instructions. Do Not Use other manufactures instructions to apply as a grain of rice, in a line, etc.

3. IC Diamond 7 is highly loaded with diamond particles (94%) and as a consequence its consistency is very thick. It requires maximum allowable high pressure to achieve a proper Bond Line Thickness (BLT). A lightly pressure loaded heatsink will produce a thicker joint which increases resistance, compromising performance. The heatsink may have to be shimmed with washers or some other method to increase pressure; this will increase performance and shorten burn in/cure times. (See Chart below.)

4. Some large, tall heatsinks which extend some distance from the CPU, when mounted horizontally, may have a tendency - if lightly pressure loaded - to have an increased gap at the top end. This may require shimming to correct this problem which will increase thermal compound performance.

 

[Liquid3D]

Excellent idea, and a definite step above much of the info out there.

Best advice I can give anyone, mounting your heatsink is something which should be repeated once prior to case installation and running.

Apply, mount and remove to inspect to see how the well the paste has spread along baseplate/IHS surfaces. Of course this is just a very crude visual inspection given what's occuring microscopically, however; it may save you a lot of headaches once you install that motherboard.

I know it seems like a waste of paste, but the small amount you use is really not that much given the quality of this paste as a long-term product.

With "Tower Coolers" look for products which come with secure through-board mounting hardware and always mount the heatsink on your Mobo outside the case on the anti-static bag. When I reviewed the original Thermalright HR-01 I remeber trying it with their hardware, then obtained better results (which couldn't be used in the review) using a WB back plate and spring-bolts.

If you have a Intel Stock cooler, you can do what I did in my TIM round-up. I removed the plastic push-pins (which can be done carefully and you won't even break them) and used my own spring/bolt combo which is why I was able to mount and remount almost 30x without having a nervous breakdown.

Cheers

 

[Roofles]

Just wanted to post some results here as I did a basic test with my 8800GTS as taking off the Tuniq Tower is a freaking pain in the rear, so I'll do that at a later date.

So my testing suit was basically a core temp reading on my 8800GTS, I was running loops of FUR at 1680x1050 with 8xAA to load test it, with fan set to 100%. I applied Arctic Silver Ceramique first burned in for about an hour then removed it and applied ICD7 and did the same thing. Arguably I may or may not have hit cure times on the ceramique and I didn't think there was one for the ICD7 but with fairly unimpressive results there are larger underlying factors here than my TIM. So without further adieu...

Ceramique: 60C Load at 100% Fan
ICD7: 60C Load at 100% Fan

My initial error was that I didn't check the stock TIM at 100% fan but at 75%, but once I realized my error it was too late to repeat its test. So in general I have no idea if I had any benefit at all from switching thermal pastes on the GTS.

I'll test on my Q6600 at a later date once I finally hit 3.6GHz and stop messing with that! I was just hoping to toss a comparison out there for you Joe so that I would be off the hook as it took me a month or so to actually keep my end of the bargin! But again, thanks for the free tube.

 

[JoeC]

Hardwarelogic is starting to post the results of their ICD7 giveaway for those that have an interest

http://forums.hardwarelogic.com/f14/ic-diamond-7-carat-thermal-compound-12332.html

 

[JoeC]

AMDZONE giveaway results

AMDZONE is starting to Post their results for those that have an interest

http://www.amdzone.com/phpbb3/posting.php?mode=reply&f=1&t=135166