Innovation Cooling's Diamond 7 TIM test results

[123bob]

We are starting a thread here for test results for a new TIM some of us are testing. Innovation Cooling developed this stuff made out of diamond paste. We are testing it here and posting results. This thread is for that purpose. We are giving our honest feedback on how the results go.

So, a little background info, extracted from an email by Innovation....

Application

Every manufacturer has their own prescribed method of grease application for this test I would use our method out of the gate, then when you know what the best possible result that be gotten you have a basis to fine tune or experiment with your own application methods if you feel the need to do so. The magic Cross is the best application method I have seen to date but tends to use more paste.

IC Diamond is composed mostly of diamond powder, and as such is quite thick. 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 - the center of the CPU is where most heat is concentrated. Place the heat sink on the CPU and press it down , then twist it back and forth a couple of times to spread IC Diamond over the CPU's surface. Clamp the heatsink and power up the PC.

We recommend a pea sized amount on center to increase the likelihood of optimum results for the average user. That amount with the proper pressure will spread to cover the entire area of contact between the sink and IHS 100% of the time. All paste application amounts are subject to a kind of artistic guesswork i.e a grain of rice is marginal with ICD7, it is highly bulk loaded for long term stability, very different than the "wet" style lightly loaded greases which are subject to bake out or pump out in the short term.

A squirt and squish with a twist to spread it around is just about as easy as it gets and from our experience effective and foolproof.

Mounting pressure

Watch or make note or just be aware of your mounting pressure, we have a chart on our site showing performance vs pressure. For example from 60psi to 45 psi thermal resistance doubles from an increase in Bond Line Thickness (BLT). Obviously I do not expect that measurement from the group but it is a performance issue that will affect results. Lightly loaded sinks will affect performance significantly and not just for ICD7 also on tests I have done on competitive greases

http://download.intel.com/design/Pen...s/30255304.pdf

Above Intel link has a discussion of heat sink pressure loads on an Intel stock solution sink. Some of the loads discussed pg 51 43 lbs. +/ 10lbs.

So referencing back to the ICD7 chart pressure vs thermal resistance Innovation Cooling

Good stuff to browse for key points. Intel says to cover the entire IHS with TIM. Designs should compensate for retention bracket fatigue. That the IHS gets flatter with more pressure, internal diodes stuff etc.


ICD7 Background Discussion

There are 2 primary considerations that are inextricably linked in TIM overall performance the two components are.

1.Bond Line thickness- The thinner the joint the less resistance to heat transfer

2. Bulk thermal conductivity- The fillers ability to conduct heat

So an ideal grease would comprise a very thin bond line coupled with a filler that has great ability to conduct heat like copper, aluminum, silver and in our case diamond. All formulations try to hit this perfect ideal with some combination of the two. Some lean more towards more bulk some lean more to low viscosity for a better BLT.

A number of the performance type pastes operate on what I call the peanut butter or toothpaste principle from user TIM experiments I am sure most of you have heard about at one time o r another. A "wetter" grease with less bulk and more liquid will spread and cover a larger area and give you a thin BLT but the thermal conductivity may be low, but the performance initially is pretty good. That kind of paste gets it's performance due to its relatively thin thickness, but it's performance is relatively short lived like the peanut butter. It dries or pumps out in a relatively short time frame.

To develop a better mousetrap compound the effort was focused on performance and longevity/reliability. Performance was the easy part, the reliability part while harder to execute could not be simpler in concept. It is harder to bake or pump out a solid than it is a liquid. The bulk loadings on ICD7 were pushed as high as we could go(94%) and still maintain some workability and resulted in a extremely stable long term mix

As it not reasonable to run multiple year-in system tests, we have run some accelerated tests that are indicators of reliability.

We ran a test ICD7 in a thermal cycling chamber - from -40 C to +85 C - for up to 250 cycles (with the copper block heated with 130 W of power) and found no change in thermal performance / thermal resistance of the grease

In a separate test we pulled "full vacuum" on ICD7 with a mechanical pump for ca. 65 hours at room temperature. The resulting ... weight loss was almost immeasurable - 0.3 mg out of 6.2222 g. This works out to me 0.0048 wt %.

Another test we ran ICD7 for 1000 hrs @125 C with no change in thermal performance.

I have one system where I have been running approximately 1.5 years with no change in thermal performance.

If people came to me 5 years from now and said that ICD7 was still working fine, I do not think I would be much surprised.

Requested Test Information


We would like the before load and idle temps along with the age or time in place of the previous mount

The ICD7 load and idle temperature and if possible some care on the ambient temperatures from both before and after readings.

Feel free to make any observations, comments or suggestions we ar trying to refine our message, application and problem trouble shooting so the information you supply to us is valuable to us and much appreciated

That's basically it you are pretty much on your own, whatever result you get is what it is. I may comment occasionally or suggest trouble shooting points but not much else. At the end of the exercise we will provide a chart of the results with links to the data for viewing and suggested corrections.

Thanks again all for your time and effort

Andrew
--
Innovation Cooling
350 East Main street
Ansonia CT. 06401
web site
http://www.innovationcooling.com/

I'll go first......

All temp data is fully 100% loaded with the BOINC application.

Farm-12 is a Q6600 G0 running on a Gigabyte GA-EP35-DS3R. The Q6600 is clocked at 2.97gig. It is running open on the bench. It has a Thermalright Ultra Extreme 120 with a Scythe SY1225SL12SH slip stream 110CFM fan. There is an additional 120mm fan blowing across the board.

**Before Data**

TIM is Ceramique, mount is 5 days old.
System has been running a few days.
Ambient temp is 23C.
Ceramique Paste pattern is two grains of rice in the middle, with a square box about 3/16s from the edge of the IHS. (Fills the "moat" these things have..)
Coretemp version 0.97.1 temps are 45,46,42,43

This is how I applied the new TIM. About two grains of cooked rice in size.


Pretty, ain't it.... I twisted and pushed the TRUE firmly into the board. It is mounted with the screws tightened all the way, as before.

Here's an overall pic of the setup.


<Drum roll please>

**After Data**

CPUzV 1.28, Coretemp VID 1.2875 (Yeah, I know I can get more clock out of this. It's playing "musical CPU" tonight.)
Mount 40 minutes old, running fully loaded with BOINC
Coretemp 43,45,40,42
Ambient temp is 23C

About 1.5 degrees (average) cooler. I'll let it run and leave a placeholder for later data. I don't know if this stuff requires a cure time.

**Far After Data**
Well, it held the above temps until I decided to mash around the cooler. I really squished and twisted it and the temps went up 1-2 degrees. Must have blown the mount.....I'll remember that for future testing and perhaps use a tiny bit more paste.


Looking good so far against Ceramique. I'll be testing against Shin Etsu in a future post...

Regards,
Bob

[Martijn]

My results:

I applied the TIM as I always do, the pea sized blob on the centre of the IHS. All 'before' temperatures are with AS5, the ambient temperature here is 19C. Coretemp 0.95.4 was used to measure the temperatures on the CPU. I will add more to this list as I put this on other systems in the house.


Q6600 @ 3.6
See signature for details. Cooler = CNPS9500 from Zalman. The mount is from the first week of January this year. Load test was WCG (doh).

Temperatures before:
69-67-67-68

Temperatures after applying ICD7:
63-62-62-63


Sapphire ATI Radeon X1950Pro
Stock cooler, still clogged with dust (which I did not remove for accurate testing) and overclocked to 607/750 from 580/700. The mount is from December 2006. Load test used was Call of Duty 4: MW.

Temperatures before:
~85C

Temperatures after applying ICD7:
~77C



Sapphire ATI Radeon HD2400Pro
Stock cooler (passive), overclocked to 710/560 from 525/400. The mount is from October 2007. Load test used was Call of Duty 4: MW.

Temperatures before:
95C+

Temperatures after applying ICD7:
~88C





Oh, and just a quick question: how on earth did you guys manage to get this stuff into the tubes? It's so thick lol!

[Kayin]

So this comes pretty highly recommended so far then?

[[XC] riptide]

Guys. PLease. ALL voltages need to be posted. All actuals before and after droop effects. Boards etc. Cmon nOObs.!!!

[tastymannatees]

""""Oh, and just a quick question: how on earth did you guys manage to get this stuff into the tubes? It's so thick lol!""""

ICD7 aka tastymmannatees

We use a really big pump and load it from the tip.

It's thick for long term reliability on the mount with thermal cycling.

It is just harder to bake out or pump out a solid than it is a lighter bulk loaded "Hand Cream" or wet type compound. Intel recommends > 90%, we push to 94 %. There were higher bulk loads on some prototype compounds we were playing around but But gave us conniption fits on the installation.

[Martijn]

Guys. PLease. ALL voltages need to be posted. All actuals before and after droop effects. Boards etc. Cmon nOObs.!!!

See signature for details.

Q6600 @ 3.6Ghz, 1.48v actual

Thank you


EDIT:

""""Oh, and just a quick question: how on earth did you guys manage to get this stuff into the tubes? It's so thick lol!""""

ICD7 aka tastymmannatees

We use a really big pump and load it from the tip.

It's thick for long term reliability on the mount with thermal cycling.

It is just harder to bake out or pump out a solid than it is a lighter bulk loaded "Hand Cream" or wet type compound. Intel recommends > 90&#37;, we push to 94 %. There were higher bulk loads on some prototype compounds we were playing around but But gave us conniption fits on the installation.

Thanks for the info. When I took off the AS5 on my X1950pro it had indeed almost completely dried out.

[[XC] riptide]

Thank you

.

martijn.... Your sig ain't gonna last forever. Ever go through 6 month old thread. and guys say,,, 'Ya so same rig as in my sig I put the 3800 X2 into ' . The you look at sig and he's got Q6600 + Asus Preium etc etc.

get my meaning. ?

[Martijn]

martijn.... Your sig ain't gonna last forever. Ever go through 6 month old thread. and guys say,,, 'Ya so same rig as in my sig I put the 3800 X2 into ' . The you look at sig and he's got Q6600 + Asus Preium etc etc.

get my meaning. ?

Sure, but hey... This Q66 is probably going to last me forever

[123bob]

If the only thing we change is the TIM paste, as we are supposed to be doing for this test, what would we need to post our voltages for? We aren't supposed to be changing CPUs, clocks, voltages, coolers, cases, fans, or anything else. We are only looking at temps here....or am I missing something? I can be a big noob sometimes, so let me know...

(Side note, for those who don't know, "tastymannatees" is from Innovative Cooling and is monitoring our results.)

Bob

[[XC] riptide]

Well.. Bob, becasue it also will act as a repository of results for diff configs also, as well as the TIM results comparison.

Also a guy posting more info can let us see if there are any outliers in the results eg. Bad mounting etc on both tests.

Also there may also complex systems at work here. For example if the TIM has better ability to move heat energy it will be more apparent on a high heat output (high voltage) + High end cooling system than on say a lower voltage + less effective cooling. Is my logic sound? Anyways.... carry on. I wish not to dusrupt the results.

[123bob]

Well.. Bob, becasue it also will act as a repository of results for diff configs also, as well as the TIM results comparison.

Also a guy posting more info can let us see if there are any outliers in the results eg. Bad mounting etc on both tests.

Also there may also complex systems at work here. For example if the TIM has better ability to move heat energy it will be more apparent on a high heat output (high voltage) + High end cooling system than on say a lower voltage + less effective cooling. Is my logic sound? Anyways.... carry on. I wish not to dusrupt the results.

Sounds good to me. I'll edit my thread and add CPUzV and VID. Probably one of those things that is better to have, than not..

Regards,
Bob

[SiGfever]

I received my paste yesterday. In my hurry tonight (wife got new HP laptop for work, daddy had to setup) I forgot to get the "before Idle temps", damn noob.


Specs:
Q6600 G0 vid = 1.2375
Abit IP35 Pro v1.1 bios = 14
Corsair 2 x 1Gb 6400c4 (5,5,5,15...1.9v)

Settings:
vcore listed: 1.4275v
vcore real: 1.35v
Buss Speed: 396
vMCH: 1.29
vICH: 1.09
Multi: 9
Core Speed: 3564
CPU VTT: 1.23
ICHIO: 1.55
CpuGtlref 0/2: 68&#37;
CpuGtlref 1/3: 68%
PCI-E Clock: 101

Cooling:
Dtek FuZion WB (MX-2 TIM, large BB size dolop) (mounted 3/18/8, 15.5mm from top of knurled nut to top of the mounting plate) w/Backplate.
D5 Vario @5
MCR320QP w/ 3 120cfm fans in pull configuration
T-Line
7/16"id x 5/8"od MasterKleer

Temps:
Load-Before: (Boinc-all four cores @ 100%) CoreTemp v95.4
Ambient: 23.5c

Core 0: 42
Core 1: 42
Core 2: 42
Core 3: 43

After: Same mounting distance 15mm, small pea size dolop w/twist 1-2 degrees.
Ambient: 23c

Idle:
Core 0: 35
Core 1: 32
Core 2: 35
Core 3: 33

Load:

Core 0: 46
Core 1: 45
Core 2: 46
Core 3: 47

I will update when TIM has had time to squeeze out properly and set, update in a few days. If temps remain high will pull WB to inspect TIM coverage and remount for further testing.

[phelan1777]

ill post mine shortly here.

qx6700 currently air cooled under a Big Typhoon using ASC using speed fan for temps.


I will try to follow Sigs example though for details.

(will get more technical data @ a later time. (will be busy the next three days)

Specs:
Board: Evga 680I A1 P31 Bios.
VGA: Single 8800 GTS 640 w/ Stock cooler (ASC in place of stock TIM) I will switch out that TIM as well, and see what temps I get there.
RAM: 4 GB G.Skill PC2 6400 5-5-5-2




Settings:
vcore listed:
vcore real:
Buss Speed:
vMCH:
vICH:
Multi:
Core Speed:
CPU VTT:
ICHIO:
CpuGtlref 0/2:
CpuGtlref 1/3:
PCI-E Clock:

Cooling:
First run Thermaltake Big Typhoon
Fan: San Ace 120X38m pulling through Heatsink (will get RPMs)
Board is laying open on my desk, not in case.

Temps:
Load-Before: (Boinc-all four cores @ 100&#37 CoreTemp v95.4
Ambient: c

Core 0:
Core 1:
Core 2:
Core 3:

After:
Ambient:

Idle:
Core 0:
Core 1:
Core 2:
Core 3:

Load:
Core 0:
Core 1:
Core 2:
Core 3:

[123bob]

OK, after looking at SiG's detail, I'm with the format. As Rip suggested, this info is useful, for more than the TIM test. The caution here is that this thread is about D7s performance. Let's try not to get too far off that.

I sat back today, and had to ask, what the heck could a rock like diamond do for heat conduction that copper can't??? I googled "thermal conductivity diamond" and got from Wilkpedia the following..details here

Thermal properties

Unlike most electrical insulators, diamond is a good conductor of heat because of the strong covalent bonding within the crystal. Most natural blue diamonds contain boron atoms which replace carbon atoms in the crystal matrix, and also have high thermal conductance. .999-12C monocrystalline synthetic diamond has the highest thermal conductivity of any known solid at room temperature: 2000–2500 W•m/m²•K (200–250 W•mm/cm²•K),[5] five times more than copper. Because diamond has such high thermal conductance it is already used in semiconductor manufacture to prevent silicon and other semiconducting materials from overheating. At lower temperatures conductivity becomes even better as its Fermi electrons can match the phononic normal transport mode near the Debye point,[6] and transport heat more swiftly, to overcome the drop of specific heat with the fewer quantal microstates, to reach 41,000 W•m/m²•K at 104 K. The same diamond at .99999-12C is predicted to 200,000 W•m/m²•K (20 kW•mm/cm²•K).[5]

Diamond's thermal conductivity is made use of by jewellers and gemologists who may employ an electronic thermal probe to separate diamonds from their imitations. These probes consist of a pair of battery-powered thermistors mounted in a fine copper tip. One thermistor functions as a heating device while the other measures the temperature of the copper tip: if the stone being tested is a diamond, it will conduct the tip's thermal energy rapidly enough to produce a measurable temperature drop. This test takes about 2–3 seconds. However, older probes will be fooled by moissanite, an imitation of diamond introduced in 1998 which has a similar thermal conductivity.

Being a form of carbon, they can burn in the presence of oxygen if heated over 800 °C (1500 °F). In absence of oxygen they can stand higher temperatures, but will convert to graphite eventually.

WOW!! 5 times the conductivity of copper. This is worth pursuing to get it right.... And also of note, diamonds can burn....

OK, to the real test, and the data we were all really looking for. At Fugger's non-request, and at the risk of getting banned, I tasted D7. Somebody had too.... It is best described as a mix between mineral oil, bacon grease, and dirt, with a healthy hint of peanut butter in the background. It wasn't as bad as I thought it would be. I've gotten old axle grease in my mouth, from working on cars, that was much worse.....


Specs:
X3350 vid = 1.0375
Abit IP35 Pro bios = 14 (Taken from bios screen, not CPUz...CPUz shows it as 10..SiG, note the date in bios. You may have V14, the latest, in your rig...)
Corsair 2 x 1Gb regular old beat up TWIN 800mhz stuff

Settings:
vcore "actual" (CPUz reading) : 1.240
Buss Speed: 438
Multi: 8x
Core Speed: 3.5 gig
PCI-E Clock: 100
Vcore bios: 1.305
DDR2: 2.025
CPU Vtt: 1.27
vMCH: 1.44
vICH: 1.15
ICHIO: 1.65
DDR2 ref: 0%
CpuGtlref 0/2: 67%
CpuGtlref 1/3: 67%

Cooling:
CPU Block/Cooler: ZALMAN CNPS 9700 NT - Standard install
Before TIM:Ceramique
Mount age: about 6 days old. (and no, we ain't talkin' about no sheep here... )
Paste pattern: 2 rice grains in the middle with a thin square ring around the "ditch", about 3/16 from the outer edge.

Before Temps:
Idle-Before: CoreTemp v0.97.1
Ambient: 24.4C

Core 0: 54
Core 1: 54
Core 2: 38
Core 3: 38

Load-Before: (Boinc-all four cores @ 100%) CoreTemp v0.97.1
Ambient: 24.4c

Core 0: 63
Core 1: 63
Core 2: 56
Core 3: 53

After Paste pattern:
Two slightly squished rice grains in middle. See pic below.




After Temps:

After Idle: Coretemp 0.97.1
Ambient: 23.5c Opened Garage door in between....
Core 0: 54
Core 1: 54
Core 2: 39
Core 3: 38

After Load: 30 minutes (Boinc-all four cores @ 100%) CoreTemp v0.97.1
Ambient: 23.5c Opened Garage door in between....
Core 0: 63
Core 1: 63
Core 2: 56
Core 3: 53

I got nearly identical temps before and after, on this round, with this machine. Ambients varied slightly, which would indicate slightly worse performance. Will update tomorrow after it burns in.

Yo, Tasty, does this application amount look about right? I may be squishing and twisting too much also.....

Regards,
Bob

[Bail_w]

Yo, Tasty, does this application amount look about right? I may be squishing and twisting too much also.....

Regards,
Bob

that amount looks about right. i tired 1 dot of grain size in the center with my main rig, the spread was not so good. i will update my results with pictures later this week.

[tastymannatees]

WOW!! 5 times the conductivity of copper. This is worth pursuing to get it right.... And also of note, diamonds can burn....




I got nearly identical temps before and after, on this round, with this machine. Ambients varied slightly, which would indicate slightly worse performance. Will update tomorrow after it burns in.

Yo, Tasty, does this application amount look about right? I may be squishing and twisting too much also.....

Regards,
Bob

The amount seems fine, with overkill applications excess just squeezes out, it is better to apply a little extra to avoid redoing the mount later.

The only issue with application might be the double blob. We have been testing grease application and performance by applying ICD7 and other pastes to glass slides and doing accelerated testing. Single blob seems to work best. their are some sensitivities/situations to application where air can get introduced into the paste and small voids form. Short term it is not a problem, Longer term the voids expand and push out the paste so the applied pastes end up looking like swiss cheese with maybe only 1/2 the area covered. The worst method for all compounds was the razor blade spreading. In your case you were not on long enough for this to occur so I think we safely eliminate an application problem.

In other giveaways 7-8 out of 10 of the total average get a positive result. Some that do not, are I believe(although the sampling is a little thin on it at this point some trends are starting to emerge) due to a lighter load pressure from either retention fatigue in the mounting bracket or a lightly loaded design to begin with. I have now on the Anandtech giveaway 2 Intel stock coolers with pretty much the same poor result with previous generic type compounds where I would normally expect at least some kind of improvement. But 2 are only coincidence at this point.

With proper mounting pressure and heat the Bond Line Thickness (BLT) will resolve(process of curing) to the approximate particle size optimally for ICD7 @ approximately 30 to 40 &#181; . above 50 and and you start to lose performance.

Some of the newer compounds out wet the IHS\sink very well and reach an optimum BLT in literally minutes and are very competitive performance wise to IDC7 people close the case and forget about it. Thermal load and cycling pump or bake out the thinner mixes in general, sometimes indications of failure occur in as little 4 weeks.

This is one of the reasons I asked to include the age of the old mount. Some of the exaggerated improvements may be due an initial bad mount on the comparison compound or what I believe is usually the situation, the installed compound has broken down or in process of breaking down +7-8 C is common in the older mounts. On fresh installs Of the more competitive pastes I usually expect to see 2 to+4C.

"""WOW!! 5 times the conductivity of copper. This is worth pursuing to get it right.... And also of note, diamonds can burn...."""

The high conductivity of the diamond allows us to highly bulk load the mix for reliability without losing performance. A highly loaded metal, metal oxide, ceramic etc. tend to lose on that end which is why on the OEM side of thermal compounds the TIM's have high reliability but lower performance. Similarly Bulk loaded to ICD7 and you probably are looking @ a 3-5 times improvement with ICD7

Most of the thermal products out these days are enjoying the similar effects of Moore's Law. On my test setup I can put a thermocouple in the IHS and another in the sink so I am measuring the just temperature drop across the compound. AT 100W the measured result is around 1C. I guess you could call this a kind of singularity point, end of the line. You could make The heat sink and IHS one piece and at best you would only see a 1C gain, most might not even see it on their PC.

You can apply this to water cooling. Say for example you have a water system with an overall thermal resistance of .06 C/W, .003 for the radiator and .003 for the water block. So here at 100W your temp rise will be 6C. Double your radiator performance/size and you will only drop 1.5C. As a practical matter the 4-5C leaps in performance are pretty much at the end of the line in ambient cooling.

I am very impressed with this group, great attention to detail, very professional, high quality. Your input is quite valuable and I appreciate the effort that it takes to do this kind of testing. Thanks

[tastymannatees]

Here ya go some thermal paste between slides

http://overclockers.com/articles1529/

[123bob]

You're very welcome Tasty. There is a great bunch of folks here. We work hard to get everything we can out of our gear. From what I'm hearing from others that you have sent paste to, they will be breaking rigs down this weekend for test. So, I hope we see more results here to compare to.

Thank you for commenting. I'll go to the single blob as you suggest on the next mount. That one will be against a 6 month old Shin Etsu mount on my mass water cooler. I'm looking forward to see what happens there.

On this last X3350 mount, note that I have two hot cores. I think this is native to the heat spreader on this chip. I tried many mounts, patterns, etc to try and get core 0 and 1 down, no luck thus far. Latest data with 8 hours of run time is Coretemp 62,62,53,48 Ambient 20.8C. So, core 2 and 3 have cooled a bit with the ambient drop overnight. Core 0 and 1 have stayed up a bit more, I think this is a manifestation of a possible IHS problem on the die.

As you mention, it is possible the 9700 cooler on this last rig is a bit whimpy. I'll mod it to increase the pressure. I think that's a weak point in the Zalman design. They are somewhat whimpy right out of the box.

I think the ThermalRight Ultra Extreme (TRUE), in the first test, has first class pressure. I REALLY like the metal backplate they put on that. I can crank it all the way down and not worry about warping the board so much. I wish I had some of those plates for my Fuzion blocks...

Your pressure explanation fits with my test results. The TRUE mount performed better than the 9700 mount. I'll go back and beef up the pressure on the 9700 and re-apply in the single blob technique.

Also of note to other testers, you want to really clean the bee-jeebers out of the old mount. Get all of that old paste off so you don't skew your results. I have been using acetone in between all mounts and have been careful to clean thoroughly.

Regards,
Bob

[123bob]

Here ya go some thermal paste between slides

http://overclockers.com/articles1529/

Good info there. Thanks. It's the pea for me....

[phelan1777]

OK, to the real test, and the data we were all really looking for. At Fugger's non-request, and at the risk of getting banned, I tasted D7. Somebody had too.... It is best described as a mix between mineral oil, bacon grease, and dirt, with a healthy hint of peanut butter in the background. It wasn't as bad as I thought it would be. I've gotten old axle grease in my mouth, from working on cars, that was much worse.....

AND this is someone we young-ins are supposed to look up to?

[123bob]

AND this is someone we young-ins are supposed to look up to?

I never said that anyone should look up to me....... Just tell me when I've spoiled my DEPENDS.....

You always need the example of what NOT to do in order to maintain some balance...... However, I had a hard time with this one. It was a complex taste. The bacon grease part I'm still not positive on.....

Bob

[phelan1777]

I never said that anyone should look up to me....... Just tell me when I've spoiled my DEPENDS.....

You always need the example of what NOT to do in order to maintain some balance...... However, I had a hard time with this one. It was a complex taste. The bacon grease part I'm still not positive on.....

Bob

<pats bob on the shoulder...............you will be okay, after we get you to your nice padded room.............."ORDERLY!"

[tastymannatees]

On this last X3350 mount, note that I have two hot cores. I think this is native to the heat spreader on this chip. I tried many mounts, patterns, etc to try and get core 0 and 1 down, no luck thus far. Latest data with 8 hours of run time is Coretemp 62,62,53,48 Ambient 20.8C. So, core 2 and 3 have cooled a bit with the ambient drop overnight. Core 0 and 1 have stayed up a bit more, I think this is a manifestation of a possible IHS problem on the die.

As you mention, it is possible the 9700 cooler on this last rig is a bit whimpy. I'll mod it to increase the pressure. I think that's a weak point in the Zalman design. They are somewhat whimpy right out of the box.


Regards,
Bob

That's one question that always bothers me. "Why go to the trouble of making a great heat sink and knowing how competitive the performance market is and take a chance on losing some of that performance with just a good enough mount" with Just a couple Of degrees and people will pay more or you will sell more.

-----------------------------------

That's a very clever observation of the individual core temps as a possible indicator. I will comb back through some on the 80 or so results so far that I have when I get a chance to see if I can make a connection.

[123bob]

OK, here's the remount on Farm-09, the one with the Zalman 9700.

Used a 1/2" SAE washer to beef up the pressure on the mount.
Used pea technique, pic below.


Before Temps:

Load-Before: CoreTemp v0.97.1
Ambient: 20.4C

Core 0: 62
Core 1: 62
Core 2: 53
Core 3: 47

After Temps:

Load-After: CoreTemp v0.97.1
Ambient: 21.5C

Core 0: 62
Core 1: 62
Core 2: 53
Core 3: 47

Note the ambient is 1 degree higher. So, I gained one degree in performance with this test. Overall, that means the mount is 1 degree better with D7 than Ceramique.

[SiGfever]

@Bob,

It has been my experience that wide variance in core temps are highly influenced by uneven mount pressures. With a waterblock it is easier to get the same pressure per post by measuring the distance as I described above. A "fixed mount" such as your Zalman leaves you at the manufacturer's mercy.

Oh, and mine is v14, I rechecked bios, first time CPUZ missed it.

@Tasty,

What is the cure time for this TIM? I want to give it a fair chance before I pull it to inspect the spread pattern.
Thanks,
John