Official IC Diamond Test Results

[PhillipM]

Right, after it being mentioned that this stuff is superb for laptops (due to longetivity and the conveluted cooling arrangements at a guess) I tested it on an aspire 7100 I had lying around.
The heatsink between the CPU and chipset is linked on this, even thought the chipset half has it's own fins, the link can be seen easy by the influence of the CPU on the chipset temperature under load.

Anyhow, intial results on the standard paste/heatpad combo on the CPU/Chipset respectively, over a week of monitoring, averaged out at:

CPU
Ambient: 18*c
Idle: 62*c
Load: 71*c (at which point it starts to downclock to reduce heat)

Chipset
Ambient: 18*c
Idle: 41*c
Load: 64*c

I then decided to try it with AS5 to give me a 'usual' baseline, again, averaged over a week, I applied it using my normal method with a plastic spreader, as over the years that's always given the best results for me personally.

CPU
Ambient: 19*c
Idle: 59*c
Load: 66*c

Chipset
Idle: 39*c
Load: 60*c

Quite a nice drop really, especially as under load the CPU was no longer having to downclock (otherwise i think the stock setup would have gone another few degrees further.)

Anyway, on to the IC, I tried the plastic spreader method but it was immediately obvious that the compound is so dry that there was no way to 'wet' out the surface without trapping air from multiple passes, I could see it wasn't working, so reverted to the suggested 'bloody big blob' method.
The acer heatsinks are mounted on relatively soft springs and don't provide much in the way of pressure - in fact it barely squashed the paste down, never mind spread it - so I used a couple of plastic modelling clamps to apply 40lb of pressure directly above each die (no heatspreader here!), and left it for a few hours so the compound could creep.

I've literally just finished and put it back together, so I don't have average results at the minute, but the readings I've got over the past 30 minutes are:

CPU
Ambient: 21*c (oven's on)
Idle: 49*c
Load: 60*c

Chipset
Idle: 36*c
Load: 42*c

It's interesting to note that with the linked heatsinks there's a huge drop in the influence of the CPU temperature on the chipset (probably due to the inverse square nature of heat transfer across the block), so I can see why they claim big benefits for laptops, that's probably a lot more life in the chipset given that kind of temperature drop.

Anyway, the AS5 looks fairly poor at the side of it - but I'm guessing the reason for that is the usual laptop problem of quite a large gap, especially on the chipset side, and generally thin/wavy patches in the heatsink for it to fill - it's fairly thin after all, especially compared to the thermal pad that was on there, and IC is thick enough to stick bricks together with, so probably did a better job of filling.

Anyway, enough rambling, verdict?
For the size of the tube, and the performance, I think I've found my new standard.

Next up, 4 core phenom desktop...

[Brainiac361]

Thought i would test this on my Laptop, as im getting annoyed at the fan on it.

First results are bloody impressive. Although i think the old TIM had failed!

All results are recorded at a 22C Ambient.


Original TIM:



The processor started to throttle back as it got too hot! (Over 100C)


IC24:



Results:

Original:
Delta Idle T = 17C
Delta Load T = 79C


IC24:
Delta Idle T = 11C
Delta Load T = 32C

Difference:
Delta Idle T = 6C
Delta Load T = 47C

47C Improvement!!!! And now the Noisy Fan has gone!!

This stuff is impressive. May use it on my 5850 Soon!

[El Wayneo]

Finally got mine, will change paste over the weekend and report back.

Will do Main Rig, HTPC, Netbook, 2 x GPU.

[Chastity]

Greets from NBR!

I just wanted to pop in and say a couple of things about doing paste jobs on notebooks.

Unfortunately not every line or model of notebook gets the kind of love they deserve in regards to cooling protection. I have an Asus G73JH-A1, and the temps on it were not as good as I would have preferred. I was idling in the low to mid 60C, and hitting 96C during gaming, and 102C+ on Furmark testing ( I stopped the tests before reaching shutdown temps ). After discussing this with Asus' R&D and with AMD, they assured me that the temps are within spec (114C max), and that everything should be ok.

However feedback on the NBR user forums proved otherwise. After a month, temps would creep up from initial temps, both idle and otherwise. What is happening is the TIM applied was thinning out and drying under load, and air was settling in. Also, as I found with mine, one of the heatsink screws wasn't 100% secure.

So I did my repaste with ICD7, and initially had Furmark temps topping out at 85C, and about a week to 2 weeks later, now peak at 77C. So overall, I had a temp drop of -23C. This was possible due to an inefficient TIM application from the manufacturer.

So my advice to my fellow laptop users, is please, by all means, monitor your temps from when it was new to going forward in the future. If you note a steady rise in temps, you should start to consider a repaste. No one should have to live with triple-digit temps.

- Chastity

[IC Diamond]

Chasity, By way of introduction started a lively little thread on notebookreview by repasting her laptop with IC Diamond and was in part the genesis of my recommendation here. The IC Diamond application to laptops seems to yield exceptionally good results for whatever reason.

[Chastity]

"for whatever reason" is the difference between assembly-line application versus personal.

[PhillipM]

[Zaim]

Got mine today, tried to apply it using a card and i couldn't, no matter what i did the paste just wouldn't apply properly, it was very thick and when i try to apply it just comes off and sticks to the card. , thinking of buying Coollaboratory Liquid Ultra Thermal Compound + Cleaning Set now.

[Marv]

Should help -> http://forums.aria.co.uk/showpost.ph...9&postcount=57

Using that method worked really well for me.

[Mr. Grapes]

when i upgraded the CPU in mu laptop, it got a re-paste... i thing it'll be getting some diamond now too

[El Wayneo]

I don't think this is good for water cooling.

Everything else i've tried it on has seen a drop in temps, especially on load, but over MX3 i have a 3C rise in idle and about 8C on load.

Have tried numerous remounts as well.

[IC Diamond]

There is a kind of an art to getting a great mount sometimes. It's easier to self train if you can see what you are doing.

I have been collecting Contact/ pressure samples specifically on water blocks, % wise they are a little worse than the retail heat sinks although a couple of the best images I have are water blocks

This is a Fuzion block that went through 3 iterations. The Fuzion blocks run about 50/50 where only 1/2 the time I see a positive result vs retail FHS which are 80/20

Pressure improves contact



Not very good on the first shot second looks better

Third try looks pretty good, notice the change in the histogram

[IC Diamond]

I am putting together an intro for a C/P competition at overclockerstech.com and generated the attached charts from some forum user tests done awhile ago. I think they make a nice connection to the preceding C/P lab Report.


Technical/Troubleshooting Note:

As most experienced users understand that when troubleshooting thermal problems there are usually 3 areas of concern.

1.) Pressure
2.) Contact
3.) Application

Taking a thermal measurement on it's own is not enough without a relative basis of comparison. Application and use of thermal compound is a mechanical function though many electronically oriented people tend to attach little importance to mechanical measurements.

Every thermal compound has it's own unique ideal pressure/thermal performance curve. A highly liquid retail compound with great contact resistance will test well at low or moderate pressures as they hit their Bond line thickness (BLT) or average thickness relatively easily but since they are at optimum at these lower pressures adding more pressure fails to yield much of an improvement.

IC Diamond is of a much higher viscosity and has a very different pressure/ performance curve than a more liquid retail compound so at lower pressures and/or contact thermal results maybe equal or less. In paste reviews the major failure is in quantifying the mechanical aspects of what is being tested, are they @ 35lbs and 50% contact area? or 60lbs 60% contact? And how does that relate to multiple thermal/performance pressure curves of different compounds?

Note: 80% of sampled retail sinks were over 50lbs psi

Quantifying mechanical's is not realistic for the individual user but they get around that by observing multiple user results to mentally average a comparison unlike the individual anecdotal test they have a reference point(s) more or less like we are doing here providing samples for comparison.

I generated the attached charts Below from some forum user tests done awhile ago.

What they show is what most know already - that good C/P provides a good thermal result. What it also shows is that IC Diamond performance margins increase with over 50 psi force and with good contact margins are higher. Pressure is dominant and contact will increase with pressure.



Data point # 2 has weak pressure and poor contact resulting in a + 3.7C increase in temps over the liquid paste. Data point # 10 with great contact and pressure shows a -5C improvement over the more liquid compound. This is a nontrivial approx 9C spread and pretty well explains why people get different results.

Also observe Point # 6 it has a somewhat lower pressure but a higher contact hence the improved thermal

So to optimize your thermal result, apply enough compound, tighten your sink and do some lapping

[IC Diamond]

Aria Final IC Diamond Test Results - All in all this was a great group for a field test, give'em the compound and they follow instructions, post results. Most times running these things is like herding cats, this group was professional with no fooling around with a high return rate... What can I say other than Great Job people! I know what a PITA to change out compounds so I much appreciate the effort on your part, Thanks to all!

Andrew

Aria's test results were fairly consistent with the overall of 16 forums giveaways on a comparison basis with an average improvement of -3.934 C on a reported 40 results vs The gross average of -4.04 of the 16 forum 529 user test averages within .11C.

Aria's # were a little lower on the AS5 averages @ around 2C vs 3.3 C for the overall but the sample was small with only 8 reported tests. The only other group in that range was SilentPC where the comparison was made on undervolted systems.

I have broken out the individual compounds in this test series and added them to the 16 forum overall and the multiforum breakout of individual compounds for a more in depth comparison.

Notes On Test Result Comparisons

When comparing results more credence should be given to larger sample sizes, smaller less so.

1-5 samples 1 or two tests can flip results either way so usually get thrown onto the miscellaneous group.

6-10 samples may start to indicate trends but can be heavily influenced by outliers and so are lightly weighted other than a general trend indicator

11-20 samples - Starting to develop more of a confidence in the trend direction.

20-30 samples - Confidence level improves.

30+ Samples - High degree of confidence










Average is a little higher here on the breakout as in includes the GPU numbers.



Quick reference here for comparison















And as we had a couple of laptop tests we will include multiforum laptop results for comparison

[IC Diamond]

This is the final IC Diamond report although this is the end of the exercise anyone should feel free to post their results at any time, make comments or post a question. I am particularly interested in long term reliability those that keep the the compound installed for for any period of time 6months, 1-2 years etc. I have a guy @ xtremsytems @ 2years with no change in temps on several systems overclocked and folding 24/7 and would like to gather more data.

So to to close this out a little song for you. I have a guy I sponsor who is a great fan of the paste and he wrote a song about IC Diamond which is performed by his daughter on you tube. So I claim a first here- The first song ever written about a thermal compound and performed publicly.


Danielle sings IC Diamond