High efficiency GRAPHENE assisted cooling

[PurpleMelbourne]

Hello all

My friend who is looking for applications for his graphene has supplied some 22mm diameter graphene discs for cooling experiments. They are like paper and so it can't be used lots of times as they will come apart. So I need to do only a small number of tests per graphene disc.
Its among the highest grade produced by any research team in the world at the moment and not available commercially, so its unknown thermal transfer properties are probably four times the transfer of silver, and or perhaps much better. Silver is about 7% better than Copper. Graphene is in the range of 400% - 1500% better according to what Google can tell me on this very new material. So maybe it can keep the heatsink close to as hot as the CPU and give a few degrees lower temps

I'm looking to put it to use on a not yet built Frankenpad with QX9300.
Its taking me while to figure out which is the latest information and advice on how to go about this due to the number of threads on the subject. But once I'm done I'll let you know the results of the Graphene Frankenpad (GF = GirlFriend? ). I'm wondering how much the PM965 northbridge chip would benefit from cooling.
Is it a problem, or should I leave my concerns to the CPU and GPU?


There are three stages in the CPU cooling.
First the beginning is CPU to heatsink boundary which you normally use Arctic Silver paste or similar. I'll be using dry graphene discs instead.
Second the middle is the movement of heat through the heatsink. That is fascilitated by copper heatpipes.
Thirdly the end is the exhaust where fan forced air blows over radiator fins.

So is the T400/ATI heatsink the BEST choice for a QX9300 in a T61P or T60P/T61P Frankenpad hybrid? For the moment I'll just be experimenting with a stock machine to see what happens. I haven't yet nailed down the exact step for 1066Mhz modification from reading in so many places written at such different times.

So as an interim I want to do a more conventional test. These graphene discs that I have are 22mm in diameter.

So what tests should I do that would be most useful to the community? I have T60, T61 and T61P that I can test on. Also my workhorse is a Dell M6300 with a T9500 which could also serve to provide some feedback as to how good graphene is for cooling laptops.

PS I do have a T9600 for the initial 1066Mhz testing before moving onto the very hot QX9300.

[RealBlackStuff]

That would be a T500/W500 heatsink for discrete graphics (not T400).
GPU-test: FurMark
CPU-test: Prime95

Before you use graphene, the heatsink surface might want some "lapping" to become absolutely smooth.

If you do need a thermal paste in the end, consider Thermal Grizzly Kryonaut

[zoltan87]

Wow, I am eagerly looking forward to the results. I was thinking about making a custom heatsink configuration. I would have been cutout from a copper block with cnc machine, but haven't found the time for it, also in the meantime I upgraded to a T500 heatsink and made some additional modifications, so now my system is extremely cool compared to what it was before the mod.

[PurpleMelbourne]

That would be a T500/W500 heatsink for discrete graphics (not T400).
GPU-test: FurMark
CPU-test: Prime95
Before you use graphene, the heatsink surface might want some "lapping" to become absolutely smooth.

Thank you for those very important details
Do those programs allow a timed, or continuous, stress test so as to get an accurate comparison of temperatures?

Wow, I am eagerly looking forward to the results. I was thinking about making a custom heatsink configuration. I would have been cutout from a copper block with cnc machine, but haven't found the time for it, also in the meantime I upgraded to a T500 heatsink and made some additional modifications, so now my system is extremely cool compared to what it was before the mod.

I wonder if a custom heatsink modeled on the T500 but specifically for the T60 or T61 may introduce further improvements.
What is the T500 fan like compared to the T61 / T60 fan?
Are there room for improvements be made on the fan? Maybe a slightly larger or perhaps a more powerful one?

I'm thinking to do sets of 3 tests:
Current performance with existing paste / heatsink
Cleaned heatsink + graphene
Back to heatpaste and heatsink to compare to original "old paste" performance
I'm curious to do a dry heatsink test too, but I'm worried about maybe overheating the cpu.

The first machine will probably be a Dell M6300 or D830 as I'm more familiar with Dell's.
Hopefully that will give me exciting results to enthuse me to do further testing on other machines

Any thoughts on the testing regime welcome and encouraged

[RealBlackStuff]

Read the descriptions for
Furmark: http://www.ozone3d.net/benchmarks/fur/
Make sure you run it at the screen's native resolution.
Keep an eye on the temperature.

Prime95: https://www.mersenne.org/download/
https://en.wikipedia.org/wiki/Prime95

Never checked timing options, just let 'em run and keep an eye on the clock/stopwatch.

FYI: I don't think the ThinkPad fans will work on a Dell...

[zoltan87]

I don't think there is much difference between t60 and t500 fans. Also the fans that I bought from eBay for future use (£1,60 each) are t60 and t500 compatible, which would indicate that they are the same. The t500 hetsink is definitely an improvement from the original t60 heatsink.
I wrote down my experiences with my cooling mod in the pinned thread "Make your Ati Thinkpad t60 run cool and quiet".

[pidjones]

I had read that graphene's improved heat transfer rate is in the plane of a sample, not through it. How thick is this sample? The accelerator community would be interested in ~1000 ugram/cm^2 for stripper foils. The problem with the samples I've seen is that they are not dimensionally stable under bombardment, and shrink where the beam strikes. Their enhanced heat transfer in the planer direction does increase their lifetime and strength, however. They are amazingly strong compared to amorphous carbon foils.

Sent from my Le Pan TC802A using Tapatalk

[PurpleMelbourne]

Still running tests. I'm having difficulty with stable results after removing the factory fresh (from 2007) hard heat paste pad from the T9300 CPU.
Temperature disparity of the 2 cores has risen from 5C degrees to 11C degrees with the change from factory stock paste + heatsink cooling to graphene + heatsink cooling.

The test involves a Dell Precision M6300 with Windows 7 and a fan control program (I8KFanGUI) to force the fans at a constant full speed of ~4,000rpm for consistency of testing. Manual execution of a batch file runs a Prime95 torture test and Furmark simultaneously. After 50 minutes Furmark is manually stopped with Prime95 left to run for another 10 minutes on its own for a combined 60 minute test with the temps being read and recorded by the fan control program (I8KFanGUI) both before and after the CPU / GPU heating phases.

Results without heat paste so far have been poor. Once I have stable results (hopefully interesting ones) then I'll show my data. For now I'm not finding graphene discs to be a replacement for paste. But the doctoral researcher is urging me to try the thinner samples, rather than the thickest which I went for. They are all less than 0.5 mm thick.

I'm busy trying to get you some hopefully exciting results, but at least worthwhile data even if just to rule out a potential prospect.
Also under consideration is perhaps starting again with a T9500, first with arctic silver as my control instead of the factory test as the graphene film discs will be likely destroyed by mechanical process of removing heat paste from them.

One potential anomaly to be removed from testing is that tests have been performed with case apart to reduce rebuild time of laptop. Though seemingly unlikely, perhaps poor aerodynamics may affect cooling efficiency. One fully constructed test shall be performed to rule this out before proceeding to next phase of tests.

Once stable results are obtained, then a T61P will be the next test subject.

[RealBlackStuff]

Before you use graphene, the heatsink surface might want some "lapping" to become absolutely smooth.

Like I said before...

[QWERTY Andreas]

I don't see how this will work. You are still reliant on the copper for the heattransfer, so it will effectively only work as "thermal paste".

Thermal paste is there to close all the microscopic air gaps between the die and the heatsink. Graphene is not a fluid, so i dont see how it does that?

If it doesn't close the airgaps, then its actually worse off than just using thermal paste. Thus its just another layer of material, and will make the cooling worse.
If you can level the heatsink, so that there are no "gaps" between the die, graphene and heatsink, you migth as well just slap the heatsink onto the die.

So assuming it doesn't close the gaps, its just making matters worse. Assuming it can close the gaps, then go for it

[RealBlackStuff]

+1

You got those graphene discs for experiments.
You want to build a FrankenPad, but play with a Dell to see how good/bad the graphene is.
Did you properly clean off the 2007 thermal paste?
No idea how good/bad that Dell heatsink is, but on a Precision 6300 it's apparently a P Of S.
And the matching hardware control seems to be just as poor, see http://en.community.dell.com/support-fo ... t/19332947
I would consider those Dell experiments failed (or too premature, if you want to be nice).

Start afresh, but this time with a T601 Frankie.
In Windows you can run TPF or ECW to monitor temps and fan control.
But I think that you are probably playing Don Quichotte here.

[jaspen-meyer]

The guy who authored this article, http://overclocking.guide/thermal-paste ... rogen-ln2/ , is set up to do a proper test of graphene used as 'thermal paste'.

[Thinkpad4by3]

I find graphite really interesting for for the meanwhile until I can get graphite for my self,I'll strap an ac unit to a Q9000

Sent from my E6553 using Tapatalk

[PurpleMelbourne]

Update:
OK this is turning out to be harder than I thought.
The graphene needed to be furnaced by the graphene doctor to chemically "fix it" using more proprietary process.
And I'm straying into areas that I can't disclose, so sorry the details are rather scant at the moment.
Now the CPU is running stable at 55 C which is something to work with. I'll keep you informed as progress occurs.

Now that initial pin the tail on the donkey in the dark work is mostly over, I'm looking to transfer from the Dell M6300 over to a T61P.

The Dell has served well allowing operation while under conditions of vivisection.
And though the Precision M6300's replaceable 512MB Quadro 3600M is tops, I see that indeed the T61 build is much better. CPU closer to exhaust, plus simpler AND better heatsink, also looks much easier to modify.
T61P not quite as outstandingly compact as the T61.

Still it would be lovely if someone would LEAK (cough cough) the NDA protected Intel PM965 Northbridge design / programming details to WikiLeaks so that proper voltage controls could be added via a Coreboot or Libreboot BIOS akin to modified Kernels coupled with TWRP recovery boot on Android smart phones...
Obviously not urging anyone to do that, just musing what a wonderful world it would be

Probably months rather than weeks.

[axur-delmeria]

Now the CPU is running stable at 55 C which is something to work with

Is that idle, or full-load temperature?

[PurpleMelbourne]

Now the CPU is running stable at 55 C which is something to work with

Is that idle, or full-load temperature?

Full load with Prime95 for an hour. Room temp around 20C
What temp do CPU's run at in a T61P?

[zoltan87]

Now the CPU is running stable at 55 C which is something to work with.

Wow that is impressively cool. Is this usual for t61 laptops? I have a t60 (which uses the older 60nm intel cpu vs 45nm cpu in t61), and before I modified the cooling with a t500 heatsink assembly and undervolted the cpu, the idle (!) temp was around 60 C, under load it went to the high eighties.

After all the mods I made (heatsink, thermalpaste, undervolting, tpfancontrol custom cooling profile) on my t60, I have idle temps of around 42 C and under load around 62 C.

[PurpleMelbourne]

Now the CPU is running stable at 55 C which is something to work with.

Wow that is impressively cool. Is this usual for t61 laptops?

The test is on a Dell M6300 with the fan running flat out. This has been to work out how to use the graphene properly.
Next step is to calibrate the temp of a T9500 T61P with paste, and then switch over to the custom graphene for comparison of performance.

QX9300 Frankenpad is the objective. So I'll be going down the T500 CPU cooler route or custom cooler if someone makes better ones available.

I have a t60 (which uses the older 60nm intel cpu vs 45nm cpu in t61), and before I modified the cooling with a t500 heatsink assembly and undervolted the cpu, the idle (!) temp was around 60 C, under load it went to the high eighties.

After all the mods I made (heatsink, thermalpaste, undervolting, tpfancontrol custom cooling profile) on my t60, I have idle temps of around 42 C and under load around 62 C.

14" NVidia T61 with P9700 and Broadcom Crystal HD would be a great travel computer.