IC Diamond 24 Giveaway/ Reliability Survey

I participated in the test 2 years ago and I have got a new laptop which I have repasted with ICD7:

Before
CPU (i7 3610QM) - ~50C (Idle) ~71C (everyday intensive load such as video rendering and photo work)
GPU (GTX 680m) - ~49C (Idle) ~68C (gaming load)

After
CPU (i7 3610QM) - ~47C (Idle) ~67C (everyday intensive load such as video rendering and photo work)
GPU (GTX 680m) - ~44C (Idle) ~64C (gaming load)

Ambient hovering around 24C.

Very consistent improvements however, in my experience, ICD does have the same issue that all other thermal pastes suffer and that is longevity under very hot conditions. My temperatures gradually worsened as my previous laptop had a 2920XM that I would constantly be running at over 80C sometimes venturing into 90C. ICD was the best at suppressing the temps but eventually it wore out, a repaste helped it. My mate's laptop which never rose passed 70C on any of the components stayed consistent throughout the 2 year period.

I will keep updating this with more info as time goes on.

I hope this helps. I will register in the email and keep the testing going

 

Sorry for this double post but you have the same post on two threads..............

It is not just the paste. The problem is the harder you stress out heat pipes the faster they wear out. A repaste may help somewhat but what you will find is you never get back to the original temps.

The problem is not just the pipes but contact. Since the majority of the HS's rely on sprung metal legs these legs provide less presaure over timeand excess heat actually accelerates this loss of contact preasure. There can be other forces involved as well so there is no real clear cut answer as every system is different. I can tell you with my prior installs, at least one last 12 months, the temps held up just as good as the day installed ubt thus was a shimmed install taking care of leveling and the sprung preasure................

 

To TANWare's point

TANWare is 100% correct, besides adequate application amount the only 2 things anybody has to know to understand thermal compound reliability and thermal compound performance is good contact/pressure (C/P).

It's that simple, if the sink's barely touching the CPU Performance will drop.

If the sink's barely touching the CPU, perhaps only 25% contact area you have 3X heat transfer through 1/4 the area. The solder iron effect with the increase in heat density will kill any paste or LM in short order, even solder joints fail under these thermal stresses.

If you had a 100 Identical notebooks with all with a great mounting system and running the same wattage results would be pretty much identical in performance and reliability. The issue in variability of mechanics in mounting are reflected in the variability of results.

Following charts show a non trivial 9 C spread- To qualify a result it has to be quantified.


[/QUOTE]

 

Well sorry for the delay:

MSI ex625

Before:

Load: 95c

Idle: 50c

After:

Load: 90c

Idle: 48c

Ambient temps were 28c

MBP 13 early 2011

Before

Load: 95c

Idle: 55c

After

Load: 90c

Idle: 50c

Ambient temps were 28c

The temps are similar, I have the done testing without my custom notebook cooler on the msi, and no cooler at all on the mbp. Load temps were achieved while gaming, no overclock applied to the msi. The drops have been consistent on the same systems.

More to the point, the temps remained the same on the MBP throught 1 year. The msi 4670m has gone caput so I have only temps after 2 weeks of application.

Again sorry for the delay, I used the paste on a desktop as well, it was packing a e5200, drop in temps have been significant there, due to the poor/non existent thermal application.

 

I had the graphics card that I originally provided the test data for fail, (6970m) and was sent a replacement 6990m that I used IC Diamond with.
I fit the card around 6 months ago, and my temperatures are as follows:
Ambient: 20degrees celsius
Idle:
CPU: 45 degrees celsius
GPU: 42 degrees celsius

Load:
CPU: 65 degrees celsius
GPU: 78 degrees celsius

Cheers



Mike

 

thanks for the updates guys - Still 14 short all candidates for the WALL OF SHAME I am constructing....

 

Time for an update.

Ambient temps are probably around 78-80 degrees F. I don't have a thermometer to check right now.

Core 2 Duo T5800:
Idle: 53 C Load: 83 C

Nvidia 9650m GT:
Idle: 54 C Load: 79 C

The temps have gone up compared to a year ago: http://forum.notebookreview.com/har...veaway-reliability-survey-22.html#post7681549

 

avanish11 Thanks for the update much apreciated

Following the links this a 3 year old application?

 

Andrew, wonder if you are interested to do C/P testing for my Macbook Air 13. In fact I can also monitor the MBA temp. I need to repaste, so you can PM me regarding this.

 

Your Compound and paper hit the post today.

I will accept new applicants for free compound as long as they agree to do a Contact and Pressure test prior to application and update their results

 

sorry for slacking on updating my data. have had a lot of switching of systems and mods going on over here.

i will get you a data point with in the next few days as im selling the system that compound was used on and it will be the last time i can update on the temps.

if you want to send me contact paper i will be tearing down my new M14x the week i get it to do a re-paste. not sure if i need more paste tho?

i'll post back soon

 

No significant change for this month either, gonna post a full result update in october.

 

I wonder if the results of my test came back yet, or did I overlook it somewhere?

 

You mean pressure contact tests, i think i saw them a while ago. Raw data is there, analysis isn't though.

 

I've received my pressure contact test today, I'll just need to read back a little to remind myself on how to do the test, I'll try to get it done tonight or my next day off

 

September 2012, no change.

 

1.5yr update

CPU full load : 59c
ambient: 21c / 51% humidity

Delta change: +6c

CPU idle : 35c
ambient: 21c / 51% humidity

Delta change: +1c

Note: this system was power cycled on avg 2 times a day (pump out data?) and the heatsink was blown out about once every month.

 

I would love to try it on, please send me a free compound.
How do I do the contact and pressure test?

 

These are my results before and after running Orthos for 30 mins, the room temps are 25 degrees C.

 

I would like to try this out, right now I'm using Indigo Xtreme, but I'm just about to try that indium heat spring stuff, then I can try the ICD.

 

18 months later

at Idle

CPU:
53°C
56°C
52°C
50°C

GPU:
60°C


under load

CPU:
91°C
86°C
88°C
83°C

GPU:
84°C

 

Time for an update:

Spoiler

Some strange result son the CPU, i have two cores that peaked higher than I expected for a second, but i went with the max readings anyways, the stdev on the CPU average temp is higher as a result. Ambient: 23.5 ^oC.

 

Thanks for the updates people- Been distracted as doctor likes removing bones from my feet.. Recovery is not bad but when I lose a few days it takes weeks to catch up.

Just got back the latest C/P images, UPS lost the package at it's Parsippany, N.J location for about a month then was delivered to Sensor Products from UPS in a DHL package fairly well mangled with tire tracks but fortunately contents were not damaged. I will post results later today.

 

yknyong1- Apple Macbook Air13 i5-3427V

 

I've been trying to find time to do another test run but between moving across the US, looking for a job, dealing with the hurricane and now helping some family from long island, it's been a mess of time at the moment.

 

Just to note I have a couple of retailers that are going to be carrying the C/P paper and should be available in about a week

 

Not to worry when you get to it

Having lost power for several days in last years Halloween storm you have my sympathy fortunately this storm we were lucky and did not lose power but played host this morning to relatives in need of a shower. Long Island was hit hard glad we are a few miles inland.

 

I have a question.. I wasnt apart of the first survey but id like to be apart of this one.. any chances i could get in on this to try with my M18x-r2?
feel free to email me and i agree to follow the rules and post any and all results needed..

 

I would be happy to join the tests too - Dell XPS L501x.

 

Previous readings.

No change at all from last temperature reading.
Its maybe 1 to 2 celcius cooler but thats because its cold outside so ambient temps are a bit lower.

 

yknyong1- Apple Macbook Air13 i5-3427V - test 2

 

yknyong1- Apple Macbook Air13 i5-3427V -

 

Blaze Sempai -gforce 9600

 

BlazeSempaip 8700 - best yet

 

Hi,

I was wondering if anyone could suggest how to remove the scratches from my GPU/CPU dies. I was very careful when removing the old IC diamond (used iso propyl alcohol and did not rub the old compound when removing) but alas I still have fine scratches, and a strange residue on the CPU I can't seem to remove.

Thanks

 

It's amazing to me how few people understand the function of solvents and how to use them, many squirt a cleaner then immediately start to scrub the compound off polishing as they do so.

All and/or most thermal comounds contain aluminum oxide (AS5, MX4, Shin Etsu etc.) and is actually the preferred agent used in cutting glass with an MOHS hardness of 9 vs Diamond's MOHS of 10. IC Diamond is only slightly harder then your typical compound so any scratches is likely due to rough handling, contamination etc.

We recommend Acetone - isopropyl alcohol just does not have enough cutting power and we use it only as a final wipe as it leaves less residue.

Isopropyl alcohol Is mostly ineffective with IC Diamond and when applied it just sits there especially with older applications that have a loss of moisture almost the same as using no solvent at all you do need some physical effort with it to remove vs. acetone which cuts the grease in a few seconds.

Acetone is more cost effective than the thermal compound removal systems for sale, I buy it at home depot for $15 for a gallon only 2X of the cost of the cost of a 50 or 60ml cleaning system(probably same cost if you include shipping) but you get 60X the amount!

The proper way is to allow the solvent to reliquify the compound then a single wipe or two and the follow with an isopropyl alcohol wipe for less residue. Let the solvent do the work it was intended for, just good shop practice.

Just to note on the cosmetic side why one would would worry about a scratch hidden in a joint buried in a box? I mean it's not like a defect on a fine piece of art on display in your living room, it's more like being on a road trip and being obsessed with a scratch on the radiator - if it functions fine then why worry?

 

Brief summary of results to date

Some details starting to emerge about 70% seem to be holding on the various time frames which is nice.

There does not appear to be any correlation yet as to higher initial temps leading to earlier failure ( 10+ C just to pick a point)

C/P impressions to date have been almost universally poor mostly below 50-30% contact with some as low as 10% suggesting the root cause of early failure.

The 30% failure rate of notebooks in the first 3 years is an interesting number comparison to our approx. 30% result not sure there is a connection but worth noting.

I'll graph up the results over the next couple of days to get an over all picture

Some attrition with sale of laptops, system changes etc. should be noted as well as some non reporting end users who are candidates for the dreaded wall of shame.

To those who have taken the valuable time to test much appreciated

Any mistakes or ommissions in the data let me know so I can correct, it is somewhat of a hard/confusing slog going through all the posts and I am bound to have missed something

Andrew

 

Just some quick notes

Looking at the data to date it looks to be trending that if you hit a year with no change you are probably clear for extended 1.5- 2- 3 year use. A temp rise above +5C within the first year is a pretty good indicator that you will hit + 20C by year 2.

A kind of a go- no go indicator, those that hit in the 20C range I would call a failure and would repaste and while you have the system apart consider lapping the sink or some mod as TanWare's to improve contact. To note I have a couple of retailers that have agreed to test market the Contact and Pressure indicating film and should retail around $5-$8. Note that we are not selling the product and will not profit from it, we are only promoting it's general use for end users as a tool for improving performance/reliability

I am assuming a 5C +/- error on these tests as most are sloppy about reporting ambients.

As noted in my previous post there does not appear to be any correlation yet as to higher initial temps leading to a failure. Technos started with an initial temp of 58 C and at close to the one year mark ended up 19.25 C while Karamazovmm whose initial temp was 95C ended the year at 90C basically unchanged. There are other examples but this happens to be one of the more extreme. This was a suprise to me as typically notebooks runan average of 15C higher than the overclocked systems so I was expecting some indication in that direction at this point perhaps needs more time.

Applied thermal compound is simple stuff as things go, two mating surfaces with some goop in between and there are not too many avenues to explore to explain a difference in results other than amount of compound applied (we assume everbody gets this one right) Then the contact and pressure between the two joining surfaces.

As we discovered with the C/P testing the weak point in this simple setup is in the contact and is a reasonble cause for good result or bad result or something in between.

Note below yknyong1 contact area being worst case most of the heat transfer would be in the corner/edge area, a significat heat concentration more like a soldering Iron this edge corner/area would be the part of the joint to fail first and as the remaing paste is lightly contacted would then run at higher temps baking out initially then delaminating with further thermal cycling. This kind of contact is a likely canidate for early failure

the following BlazeSempai example at the other end is full contact with thermal and mechanical loads in sync with each other fully/evenly distributed across the entire processor and would be a likely canidate for long term use.

This is as simple as it gets.


yknyong1- Apple Macbook Air13 i5



BlazeSempai p 8700

 

Sorry no posting about temps. My wife's laptop went in to Gigabyte for repairs and I had to return my Clevo for an MSI. I'll post some initial temps this week once I receive my new laptop.

 

This was pretty interesting- go to 4.08 min. on the video and freeze it.

It is pretty easy to correlate paste impressions to what you see in the contact and pressure indicating film. Basically he only has contact on approx 50% of the chip with the higher heat load on the reduced area combined with higher pressure pumped the compound to the edges of the contact area. Repaste with IC Diamond netted a 30C drop at idle

Now look at around the six min. mark on the GPU he has full contact but is light on the pressure indicated by the compound being so thick, with good pressure it would be a light glaze and in any event temps were fine on the GPU at idle.

<iframe width='640' height="360" src="http://www.youtube.com/embed/2zetIRCILVE?feature=player_detailpage" frameborder='0' allowfullscreen></iframe>

 

My daughters Playstation's fan was making allot of noise when playing movies or games so I decided to open it up and see what was going on and do a video on it while I was in process. In any case it only had a few hundred hours of use and I thought it might be dust accumulation however was surprised to find almost none but when I pulled the sink off I found the compound in poor condition.

Most failures are a combination of bake out & pump out often not purely one or the other. The 2 contact patterns exhibit bake out from the pattern but have a good amount of compound that was forced out to the edges of the joint so is a bit of both. The compound tended to adhere to the sinks either due to temperature or surface finish.

You can see from the first contact impression that it had degenerated to about 50% +/- contact with the voiding due to material shrinkage. The voids are now pockets of air and are an insulator no longer conducting heat.

More liquid like compounds than IC Diamond shrink and leave voids as a kind of first phase of failure as it drys/pumps out.

All compounds dry out being the nature of the beast - Liquid in a porous media. It's just the physics of it.

IC Diamond does not fail at this point (equivalent time frame) as there is so little liquid (about 4%) it does not shrink even though a dry joint so that phase one of the failure is bypassed and due to better overall contact the following phase is extended for longer period of time. Note similar bake out voiding on picture 5, accelerated reliability tests.


As a kind of phase two the remaining compound with less contact accelerates the dry out or bake out of any remaining moisture due to the higher concentrated heat fluxes and then eventually loses the mechanical bond and deteriorates through thermal cycling ceasing working/transfering heat altogether.

I was going to do a full work up on the Playstation airflow measurements, thermal imaging, multiple thermocouples, noise measurement etc. but I have never seen a less friendly lab machine with processors sandwiched between componets, even cutting a window in the top would not work for the thermal imager so I will save it for the xbox and be satisfied with some contact images for now.

Contact Pattern 1



Contact Pattern 2

 

I took some contact and pressure images today of the Playstation and was initially surprised that that pressure and was so light mostly under the lower limit of the film (28PSI). This is partially due to the large surface area of approx. 1.75 inches square- applying 50lbs of force to to a couple of square centimeters on a laptop vs 20 centimeters on the playstation you will see a big drop in PSI with the larger area.

Contact being another issue I checked the sinks with a dial indicator on a surface plate and found one to be off .003 and the other .005, pretty much cupped along the axis the mounting screws most likely the aluminum bent on initial installation. Processors were off by max .002

These are the Contact images as acquired



Close up of the image on the left



Close up of the image on the right



So I made an effort to improve the contact and pressure on the second image and took a flat b**stard file to the high spots on the heat sink as well as introducing a .060 shim to the spring retention plate to improve pressure.

As you can see from the result an improvement over the original but falls well short of a good C/P mount.

When taking the Playstation apart I thought the compound was a little thick perhaps as thick as 2 or 3 sheets of copy paper reason being in hindsight the light pressure on the mount.

The zinc oxide grease is a fairly easy compound to spread and works well with a light pressure loading but the thick application coincides with the under scale pressure reading.

Even the zinc oxide should have lasted more than the few hundred hours of the light use it experienced so I believe the early failure was due to the overly thick compound increasing the thermal resistance perhaps 2X, 3X ......overheating the joint and accelerating the bake/pump out process.

So one cause of early failure can be attributed to inadequate pressure causing an overly thick bond line with the associated increase in thermal load.

 

What I've been saying all along with the bakeouts etc. This was also shown with the pictures I provided with the previous disassmbly. One year and no appreciable signs of degradation............

 

Finally have an off day and I managed to get some testing done on my laptop, still no major changes with 20 minute tests in a 28c room.



Edit: Nice to see that HP didn't fully screw up the cooling system on the HDX, although the northbridge runs along the same heatpipe as the CPU. I ended up replacing the melted thermal pad with a copper shim I had laying around sometime before I started doing the testing, but I might need to use a thicker one, and maybe sandwich the current one in between the GPU and its heatpipe, it might give better results, or will make it worse.

 

We are obnoxiously redundant for a reason-

Yes it was your post that set me off on this track. Why would you have a year with zero issues and technos a major fail after 6-9 months? and others go on to 2 and three years? Your grease was as wet looking as the day it was installed and having made a great effort for good contact and pressure it was obvious from your contact/paste impression you had it all and there in lies the key.

There are literally hundreds of papers on compound failure and most are done in labs under controlled conditions idealized contact, pressure etc. but they lack the variables we encounter here in the real world. Tracking those variables raises the bar and provides background for discussion

I am sure there exists among the thermal grease cognoscenti knowledge of these extenuating factors and perhaps in the published literature but read 10 or 20 reports and you think you should have stopped after the first one.

Nearly all of our competitors at the retail level are marketeers with little understanding of what specifications mean that they quote or how the data was acquired. For instance many might quote 5Wm-K or 10Wm-k performance for their compound with no indication as to how it was tested or calculated, so one man's tested 10W m-k is another's tested 4Wm-k as one extreme I have noted.

As far as I know we are the only one that publishes reliability data as at the commercial level it is pretty much of a requirement and most certainly the only one that provides it at the retail along with a substantiated argument for it.

Reliability is synonymous with quality which is often missing from most TC discussion's, for the over clocker that changes their sink more often than they do their underwear perhaps not so much a big deal. For the Notebook and Game Station people a re-paste every six months it is a PITA as noted here and other forums it is constant temp battle for some with overheats and yellow lights, if we can supply a solution with our compound as well as providing correct info for a proper fix we will have done our job. Which is why we come to the forums to be educated.

 

Thanks for the update Kuu much appreciated

 

While I don't use contact paper I am more than a bit retentive about my TIM installation. This is why it appears as new after a year. No other paste TIM used after a year with even great care could hold up as well.

About the only other thing that may hold up as well is the conductive liquid metal or pads. I am too chicken to try the liquid but maybe some day I'll try the new pads. For now ICD is my go to TIM for everything as I have plenty so why run to the store for cheaper stuff?

 

I bought IC Diamond 24 from eBay one year ago and applied to both CPU(i7-2860) and GPU(AMD6990) of M17xR3. After one year, GPU max temp is still around 75C but CPU max out to 94C from 84C. I reapplied ICD24 and CPU max temp drop back to 84C.

 

Hello,
so 2 years after repasting. There has been constant temperatures ever since the first test.
Temperatures are the same with the first test. Doesn't fluctuate even after a 30 min test for both the GPU and processors.

 

I think I'm due for a repaste after almost 2 years. My temps are at about 65 and when I don't blow the dust out of my docks it can go to 75-80. In any case, it's not bad at all after 2 years if you consider before repaste I was at about 80 cpu and 90 gpu. I bought a new AUO v5 screen and I have to open laptop so ill repast and re-report results.

Sent from my ASUS Transformer Pad TF700T using Tapatalk HD