IC Diamond Test Results

[JofaMang]

I have experienced some scuffing from using the IC diamond. I have irremovable dark scuff marks on my CPU, running the width of the surface where my heatpipes make contact with the chip. My camera batteries were dead, and I needed to get my rig back together, but I will try get shots of it in the next few weeks anyhow.

Kind of destroyed the chance of ever making a warranty claim, but in reality, this chip only has to survive till bulldozer so perhaps it is just one less excuse keeping me from lapping the cpu, hah.

 

[builderbobftw]

^ Ouch.

Sorry bout that.

Ima check my CPU, and reapply after, Im getting worried.

 

[lucuis]

I apologize, i haven't had much time to do anything. But i did manage to squeeze in a GPU regooping on my 8800GTS 512mb.

Optional Info: Sucker is overclocked 740core/1848shader/972mem It's a modest overclock, but i was having temperature problems for a while with it. My highest OC was 752core/1921shader/1052mem for reference, some programs weren't stable though. Fan was set at 54%

Pre-IC7
Compound: STOCK
Ambient Temp: Totally forgot to grab this reading, but i've been monitoring temps over the last few days. So we got a guesstimated average improvement.
Idle Temp: 57-61ish if i recall correctly.
Load Temp:83 yeah...83

IC7
Ambient Temp:same as the first one, it varies.
Idle Temp:51-55 so far
Load Temp: 78 I should mention that i had another gamer friend over and we heated up the room quite a bit with our gaming. So if i had to guess, the IC24 compound had the tougher run and still pulled way ahead.

Both load temps where noted after a good day of gaming, HWMonitor keeps a max

Wasn't as precise and scientific as i would like, but you can definitively say the IC was a resounding improvement Good job guys. If i manage to finally find the time to get this stuff on my Q6600 i'll be a happy man...and should have some more results for you all.

 

[cmcghee358]

Just gott my IC24 in on Friday, applied today:

OPTIONAL INFO: AMD 955 BE C3, 1.5vCore/3.8GHz, Corsair H50 Cooler, Antec 1200 Case

Pre-IC7
Compound: Arctic Silver 5
Abient Temp: 25C
Idle Temp: 39C
Load Temp: 53C

IC7
Ambient Temp: 25C
Idle Temp: 37C
Load Temp: 51C

I experienced a 2 Celcius drop across the board.

I will be doing my HTPC next weekend. Will post results.

 

[cmcghee358]

Ok , Here's what My system looked like before the stress tests and after 3 hours of running it, with the standard TP which came with my Board.


I must tell you that the ambient temp of the room that I work in is 45Deg C...... a very hot environment.

You're really in a room thats 113 degrees? That seems a little outrageous

 

[builderbobftw]

^

That's like 7F cooler thna Iraq in the dessert during summer

 

[Shadow703793]

^Yup. If you live in a hot area with no AC, things get pretty hot. Hell even in the US, some places get pretty damn hot (ie Texas, Nevada, Cali,etc).

 

[Computerrock1]

^ Florida

 

[JofaMang]

You don't even have to been that far south. Here in Winnipeg (60miles north of US/Canada border where ND and MN meet, on the Red River) in the summer we will routinely get heatwaves that push us to 40c 100% humidity, as far north as we are.

With no major bodies of water to moderate our climate, we got an +80c swing from peak to peak over the course of a normal year. Why the hell do I liver here anyway? hmm

 

[builderbobftw]

^

Wanna trade cliamtes?

I live in Massachusets, its ball-freezinly cold half the year, and the other half it's like 100F.

Blows.

 

[dEAne]

thank you for sharing I need these. . .

 

[IC Diamond]

Ambient temp is not a factor in the test result as the final number is the temperature rise/change above the ambient.

The C/W is a constant, at a 100 W you might see a 3C change at 20C ambient and @ 40C ambient it will be the same 3C

 

[IC Diamond]

Updated Results any corrections or changes let me know

 

[IC Diamond]

I have experienced some scuffing from using the IC diamond. I have irremovable dark scuff marks on my CPU, running the width of the surface where my heatpipes make contact with the chip. My camera batteries were dead, and I needed to get my rig back together, but I will try get shots of it in the next few weeks anyhow.

Kind of destroyed the chance of ever making a warranty claim, but in reality, this chip only has to survive till bulldozer so perhaps it is just one less excuse keeping me from lapping the cpu, hah.

Been busy and kinda neglecting things

In 3 years in the field I have had no complaint on the scuffing issue with the exception of a couple of reviewers who were doing multiple installs.

Those people that lap and understand abrasive grades will get what I am saying very quickly.

I have taken a piece of copper and lapped to a < 3 micron finish then with a 1 micron polishing compound take it to a true mirror finish.

Now if I take a soft abrasive like toothpaste and rub it in it will scuff the surface and ruin the true mirror finish.

If I substitute a number of thermal compounds and rub the finish for a few seconds they will also destroy the mirror

Many thermal compounds contain larger grit sizes say something 15 to 30 microns in size and composed aluminum oxides, ceramics and other hard particles - in other words many thermal compounds are composed of the same materials used for metal and other polishing, they are by default lapping compounds.

my daughter won the science fair this year, her topic? Substituting household polishing compounds for thermal compounds. She used liquid metal polisher, Cerama Bryte, toothpaste,soft scrub and of course some land o lakes cheese as a baseline. some were better than the stock compounds tested here

You will get wear and tear as a consequence of multiple installs and removals, cleaning etc. and as a consequence some care should be taken to minimize damage when the lettering is only a couple microns deep. Abrasives have to move in order to work. The less wiping, rubbing, installing you do the better. When removing let the solvents/liquids do the work.

In your case I would like to see before and after pictures, and perhaps buy a sink and try to see whats going on. maybe I will do a quick video on the abrasive comparisons between compounds

One possibility is rotating the sink under 50-60 psi pressure to make sure you have a good mount is the very definition of lapping

 

[JofaMang]

I did not do multiple installs. I witnessed the scuffing removing my HSF for the first time since applying IC product. This did not upset me, as I always meant to lap my CPU and HSF at some point. I will be sure to take it apart again this weekend with batteries charged for the camera, so you can see for yourself. To be fair, I do use a lot more pressure to mount and set the ICD TIM than I have had to in the past with other products, so that could be the reason for it, as you have mentioned.

If you want a before picture, I am sure AMD.com has lots of CPU shots available. It is a 965BE that I purchased about 3 months ago, and it was still immaculate before using the ICD compound.

 

[IC Diamond]

I did not do multiple installs. I witnessed the scuffing removing my HSF for the first time since applying IC product. This did not upset me, as I always meant to lap my CPU and HSF at some point. I will be sure to take it apart again this weekend with batteries charged for the camera, so you can see for yourself. To be fair, I do use a lot more pressure to mount and set the ICD TIM than I have had to in the past with other products, so that could be the reason for it, as you have mentioned.

If you want a before picture, I am sure AMD.com has lots of CPU shots available. It is a 965BE that I purchased about 3 months ago, and it was still immaculate before using the ICD compound.

What you see you see, I am not arguing that point however it is hard to trouble shoot a situation from a distance with zero info. I have no personal experience with your problem. I have done 12 giveaways over 3 years with numerous user pics of installations and the problem has not been observed.

I have to be able to see what you see

I test on mirrored lapped pure soft copper die @ 90-100 psi wipe liberally with acetone to clean- if i over rub it it will scuff but pressure alone (above the manufacturers recommendation by a substantial amount) does not mar the surface.

So if this is your experience I am trying to identify the conditions under which it occurred , before and after pics help gauge any wear and may indicate the problem source although I know a before pic is a long shot and anecdotal info does not tell me anything. Pics help allot.

I could duplicate your set up and try to replicate your result to better understand the conditions under which it occurred and make subsequent recommendations from there.

 

[JofaMang]

I may have the time to take off my HSF today, clean up the chip and take some pictures, depending on a client possibly postponing a meeting until tomorrow. Worst case scenario I will have some Pictures for you by the weekend.

I understand, I apologize if I seem snippy about it, I am truly thankful for being able to take part in your giveaway/testing, and am very impressed with the product, and regardless of my experience, will continue to use it (and will even pay for some when I run out, hah). Anything I can do to assist you, just ask.

Here is some more information to see if you can replicate the issue, should you chose to do so:

AMD 965BE c3 x4 OC'd to 3.825ghz@ 1.408v
CM hyper 212+
GA-MA780G-UD3h motherboard
Open case in steady 24c controlled climate.

I applied the ICD24 using the DHT method of filling the gaps and running thin lines half the length of the heatpipes. There was no TIM pre-applied to the CPU directly. I let the applied TIM sit for 10mins as directed before mounting to my motherboard. I placed the HSF on the CPU, and while applying significant downward pressure (don't know how much, didn't have my rig on a scale) I twisted the HSF back and forth half a dozen times for a total travel of about 20 degrees. I then secured the HSF using the mounting hardware included.

The posts were tightened in a criss cross manner (opposing corners), 3 turns of the screws at a time, for about 5 tightening cycles to reach the tightest position, which is a physically limited dead stop, not just too difficult to turn further.

The case has been open since I first applied the ICD24, in a room with steady ambient temps of 24c. I also had a single 4870 graphics card on that motherboard during the testing time, if that is of any significance to you.

 

[Computerrock1]

@ic diamond, im very impressed with the amount of care you put into customer support. I wish the majority of companies would respond to something as small as this with such interest, i normally get brushed off by companies when i ask questions about issues with their products. Good job!

 

[IC Diamond]

I may have the time to take off my HSF today, clean up the chip and take some pictures, depending on a client possibly postponing a meeting until tomorrow. Worst case scenario I will have some Pictures for you by the weekend.

I understand, I apologize if I seem snippy about it, I am truly thankful for being able to take part in your giveaway/testing, and am very impressed with the product, and regardless of my experience, will continue to use it (and will even pay for some when I run out, hah). Anything I can do to assist you, just ask.

Here is some more information to see if you can replicate the issue, should you chose to do so:

AMD 965BE c3 x4 OC'd to 3.825ghz@ 1.408v
CM hyper 212+
GA-MA780G-UD3h motherboard
Open case in steady 24c controlled climate.

I applied the ICD24 using the DHT method of filling the gaps and running thin lines half the length of the heatpipes. There was no TIM pre-applied to the CPU directly. I let the applied TIM sit for 10mins as directed before mounting to my motherboard. I placed the HSF on the CPU, and while applying significant downward pressure (don't know how much, didn't have my rig on a scale) I twisted the HSF back and forth half a dozen times for a total travel of about 20 degrees. I then secured the HSF using the mounting hardware included.

The posts were tightened in a criss cross manner (opposing corners), 3 turns of the screws at a time, for about 5 tightening cycles to reach the tightest position, which is a physically limited dead stop, not just too difficult to turn further.

The case has been open since I first applied the ICD24, in a room with steady ambient temps of 24c. I also had a single 4870 graphics card on that motherboard during the testing time, if that is of any significance to you.

Thanks for the detail

Just to note

I let the applied TIM sit for 10mins as directed before mounting to my motherboard.


these are old instructions from a couple of years ago what we used was a OEM mix for silk screening, an acetate solvent to thin it out a bit but was reformulated as the acetate interacted with some component of the syringe so no 10 min dry-out required

From our web site Application instructions


Proper application is critical to optimum performance. Squeeze onto the center of the CPU an amount of IC Diamond compound about the size of a pea (5.0mm to 5.5mm) on the center of the CPU where most heat is concentrated. Place the heat sink on the CPU and press it down to spread IC Diamond over the CPU's surface, (Do not remove the sink to check the spread as it will introduce air into the joint). Clamp the heat sink and power up the PC.


The squish and rotate was part of the early application procedure and I can not say why we dropped it other than perhaps we thought it was not necessary.

Rotating under pressure is lapping to what extent where it becomes an issue I do not have a clue. High spots can show accelerated wear with added PSI .


I must confess though that I rotate the sink on mount for quicker cure (about 5 min @ 90psi) my set up is a special spring load for consistency but my rotation is only a couple of degrees and I only do it once or twice. This is a good piece of info and gives us with something to to follow up on.

User feedback is half the reason we do this besides the promotional aspect. We may not like what hear and we will argue points but ultimately to refine our message and product we have to listen to the people that use it

Thanks for taking the time to test and provide us with some good feedback.

 

[jimmysmitty]

^

That's like 7F cooler thna Iraq in the dessert during summer

And Tucson when it feels like it. I have seen over 120F days here and in Yuma, AZ too. Only Death Valley beats out southern AZ in terms of heat.

^Yup. If you live in a hot area with no AC, things get pretty hot. Hell even in the US, some places get pretty damn hot (ie Texas, Nevada, Cali,etc).

When I lived with my grandparents here in Tucson for 13 years, all but one of those years had no AC since the last year was the sell the house year. We had a swamp cooler but man that thing sucked. Was always 90F in the house. I am suprised my PCs didn't fry.

Add in monsoon season right over summer and 100% humidity + 100-120F........

 

[lucuis]

Had a chance to swap goop again, this time on my little Q6600. It's running at 3.6ghz for all tests. I didn't take an ambient reading but both tests where completed within an hour of each other. I should also mention i have been having trouble getting the super flat surface of my water block to sit right on my bulbous cpu. So temps might look a little funny. I gave the IC24 a chance to warm up to idle, so i ran the load test, then went and ate dinner before taking the screen shot.

Pre-IC7
Compound: AS5
Idle Temp: 47 47 45 48 (46.75)
Load Temp: 66 66 71 73 (69)

IC7
Idle Temp: 43 43 34 36 (39) Ave improvement over AS5: 7.75
Load Temp: 62 62 56 56 (59) Ave improvement over AS5: 10

I'd suggest taking an average of the cores for the best reading, in brackets, seeing as how the mount wasn't perfect in both cases. Ave improvement between idle and load was 8.875, not bad at all BUT I must stress that the IC24 wasn't all the improvement. I believe i managed a better mount this time then with the AS5. Nonetheless, the IC24 is a clear improvement. Thanks.

 

[IC Diamond]

Had a chance to swap goop again, this time on my little Q6600. It's running at 3.6ghz for all tests. I didn't take an ambient reading but both tests where completed within an hour of each other. I should also mention i have been having trouble getting the super flat surface of my water block to sit right on my bulbous cpu. So temps might look a little funny. I gave the IC24 a chance to warm up to idle, so i ran the load test, then went and ate dinner before taking the screen shot.

Pre-IC7
Compound: AS5
Idle Temp: 47 47 45 48 (46.75)
Load Temp: 66 66 71 73 (69)

IC7
Idle Temp: 43 43 34 36 (39) Ave improvement over AS5: 7.75
Load Temp: 62 62 56 56 (59) Ave improvement over AS5: 10

I'd suggest taking an average of the cores for the best reading, in brackets, seeing as how the mount wasn't perfect in both cases. Ave improvement between idle and load was 8.875, not bad at all BUT I must stress that the IC24 wasn't all the improvement. I believe i managed a better mount this time then with the AS5. Nonetheless, the IC24 is a clear improvement. Thanks.

You do get biases both ways on the good mount/bad mount comparisons, some times it's a bad ICD mount but I believe they tend to average out with a large enough sample

Thanks for taking the time to test and post results

 

[cmcghee358]

I just did my HTPC with my IC24:

HTPC: AMD Athlon IIx3 425 2.7GHz/Stock Cooler/ 1.3 vCore

AS5
Idle-
Temp1: 42
Temp2: 42
Core: 31
Load-
Temp1: 48
Temp2: 43
Core: 40

IC Diamond 24
Idle-
Temp1: 39
Temp2: 37
Core: 28
Load-
Temp1: 46
Temp2: 39
Core: 38

Looks like a nice little 2C drop for all temps. Same performance I got on my 955

 

[4Ryan6]

@ The IC Diamond Rep

These tests are quite impressive, as a long term Artic Silver 5 user and knowing AS5 does require some cure in time for the best results, and seeing that a lot of these tests were performed with the IC Diamond 24s fresh application performance compared to fully seated and cured AS5, will the IC Diamond results get even better with cure time, or is what you see what you get.

If comparison tests had been fresh to fresh applications, then there would have been even greater performance gap results between AS5 and ICD24, seeing as how AS5 does need the cure in time period for optimal results.

So does the IC Diamond 24 performance improve over seat in time?

 

[jedimasterben]

I think that he said earlier that there is about a 2 hour cure time, with another .5C or so drop in temps.