[Liquid Metal Showdown] Thermal Grizzly Conductonaut vs Cool Laboratory Liquid Ultra / Pro

50$ / foot

 

Question: Wouldn't any spilled off liquid metal land on the electrical tape and then roll around once the laptop is vertical (in a bag or whatever]?

I'm about to do this repaste and I'm really confused how smooth shiny electrical tape will keep a little ball of LM stationary.
I get how the tape protects the contact points that may be around the die but what about this little bowling ball of destruction playing pinball through my motherboard circuitry?

 

Isn't that tape shiny and smooth? How does it keep spillage from rolling away?

 

ok, so i did it. I ordered grizzly online and went and got the 33+ tape and carefully followed all the threads instructions. Amazing. On a Aorus x5 v6 with a not very good cpu (silicon lottery wasnt that great for me) i now am at 4.0 ghz on all 4 cores at a max temp of 77c on cpuz stress test with fans on auto max. I previously with ICD7 was in the 90's at that clock speed.

The electrical tape isnt as shiny and smooth as regular stuff. I have some faith in it. I hope ill be ok.

Misjah was right... he had a great chip. If I did so well with a mediocre chip i can see how a silicon lottery could run cool at 4.3ghz +

Im very very pleased and amazed at how much thermal headroom ive gained. Thanks to everyone who contributed to this thread and helped me do a flawless LM repaste.

 

@knibbler did you repaste the gpu, if so what are your temps like?

Great results so far mate.


Sent from my iPhone using Tapatalk

 

Ok so I have both Conductonaut and Kryonaut. Any recommendations on whether I should repasted cpu/gpu with the same paste or a combo of the two (ie conducto on the cpu, kryo on the gpu)

thanks!

 

Good choice!
Usually for best and safe appliance is:

Liquid metal type thermal compound on CPU (application on die and on heatsink to reduce gap);

normal type thermal compound on GPU (this because usually there is a higher gap than CPU one, for big die "X"method, for small die "pea"or "small dot" in center).

Everyone feel free to correct me if I'm wrong please ^_^

 

I use LM on both GPU & CPU ... Never had a problem.

 

Roger that. I think I'm gonna try LM on the CPU and kryo on the gpu. The pascal chips seem to run cool enough as it is.

 

Pascal prefer lowest possible temperature. If you can reduce your gpu temp by 15C, you will get 39MHz higher Turbo boost

 

Guys, I'm trying to gauge my LM job. Do we check temps with fans at full blast or do we let it run with the standard profile?
(I'm guessing full blast coz otherwise the temps will be the same and just the fans will spin slower, please correct me)
And is it done with CPU load only or CPU+GPU load?

 

Full blast allow for relevant repetition comparison (if ambient temps are the same of course).

As for cpu only or both, you can actually do both to compare the two cases

 

I check at full blast and usually doing both, CPU (OCCT or AIDA64 FPU) only and CPU+GPU (OCCT + 3DMark or AIDA64 FPU+GPU or AIDA64 FPU + 3DMark). Despite this loads will be unrealistic in comparison with gaming, they show quickly uneven core temps and bad repaste attempt.

Edit: stress GPU with 3DMark.

 

And you still test with Furmark? Although all advices not to use it. Oh'well

 

Because it's harmful or doesn't provide a proper heat?

Got it. 1st

 

Well known vBIOS modders and the graphics manufacturers advice not to use it. You risk killing your graphics. A well known topic!! And mobile graphics is even more fragile than those used in desktops. @Mobius 1 @hmscott @Phoenix

 

Right now I'm not doing to poorly.
I'm now x264'ing a bunch of video's at 4GHz, 26c ambient temperature and I'm hovering between 60-65c with fan at automatic = inaudible.
At full blast it'll do much better than that but at 26c ambient I won't have good low temps anyway, I'll try that at a later time.

I did use a soft foam dam on both CPU and GPU to block any LM runoff that I think worked real well. Pics to follow if interested.

 

Yeah I'd like to see pics... More details on soft foam dam.

I did a LM repaste on my Clevo P650SG, using the conductaunaut. Used the 3M 88 tape on CPU+GPU. About 3 months in, with moderate movement of laptop, no issues and excellent temps.


Sent from my iPhone using Tapatalk

 

will the liquid metal run? If I carry the laptop around in the vertical condition will it seep through and run onto the motherboard?

 

I'd also like too know this as well since I carry my laptop in my backpack all the time.

Another question I've had an I've search around a bit with no answer. How does the liquid metal hold up too below freezing temps 0c an below, I live in Canada an the laptop gets left in my big rig in the winter quite a bit.

Sent from my SM-G935F using Tapatalk

 

no

don't put too much and you're fine

LM cannot be exposed to temperature under 9c

 

Those reasons are the ones keeping me away from LM. Because usually I carry my PC and use public transport.
@Mobius 1: Are you aware of hybrid paste consisting of LM held together with materials like ceramic or other material used in TGK or ICD or GC?
So in one word, traditional paste with LM like heat transfer.

 

Shock and handling will not make LM run off

By hybrid paste, bits of copper mixed in with a transfer medium. AFAIK it's coolaboratory liquid copper.

Not as good compared to real LM.

 

I can add my two cents, two years of daily bike commute (even in winter with sub-zero Celsius temps) did not produce any issue with LM.

If there are any complications, it will be right after applying the paste. If you take care of the amount of paste you out in, test thoroughly the laptop, and do not notice any suspicious sign of leakage, you're good to go.

 

So my $0.02 worth on all this then.

I LM'ed (Conductonaut) my P775 last week after I delidded 2 7700K's and a 6700K for the first time (one a day since I relidded them again as well). Delidding is easy so there's no real challenge there.

LM'ing the laptop was more... interesting. The heatsink seemed quite even but with a regular repaste (as a trial) I got a very poor connection between both the cores and the heatsink. I found that the cooling pads are quite thick, which is fine as a good pressure on those is better for heat transfer, but the poor clamping mechanism of the heatsink onto the cores means the cooling pads basically overpower that mechanism and keep the heatsink slightly off the core, partially. Compressing all the cooling pads at once by pushing down on the heatsink doesn't work, it takes too much force. So if you run into this problem the way to fix this is by shortly before mounting the heastsink again to press down hard on each individual cooling pad and squish it flat. Don't worry, it'll bounce back once mounted and still create a good seal between the chips and the heatsink. (also I added some sturdy foam behind the GPU board to push it up a little bit).

So the challenge is for a first timer like myself to LM the cores (CPU+GPU). People say 'don't use too much'. Great, not too much but not too little, what am I, Goldylocks? I don't mind taking risks but if I can I like to have insurance, especially if it's basically free insurance.
In another thread several people scoffed at my idea of using a dam, saying the LM won't run off... if done right, and that it'll just create backpressure onto the heatsink. Don't mind the big companies using a barrier of sorts as a safeguard when applying LM, but MY idea was stupid. Of course.

My idea was technically sound so I went ahead with it anyway. I found out through baking several foams in the oven for a while that most (in my case all) foamy material will withstand temperatures in excess of 100 degrees Celsius. So I simply chose the softest foam I could find, It was 27.5m thick to start with and would compress down to 1.2mm, that seemed all right (dunno where I sourced it, I just had it laying around, I like to have raw materials around that might come in handy). So I cut core outline shaped pieces of just slightly thicker than the GPU core (above the substrate) and a little thicker than that for the CPU as it had to go to below the substrate and fill up the gap around the socket to just above the IHS.

I nailpolished and taped the little SMD components around the GPU and 'glued' the foam on with a swab of nail polish just to keep it steady. For the CPU I simply stuck it in around the socket and made sure it was surrounded on all sides and about 1 to 2 mm above the IHS (but to the side of it, not directly over the IHS of course!).

The first LM attempt worked fine, buffing the heatsink to a mirror finish and then rubbing LM good onto the core and the heatsink first to create two shiny surfaces that would link up real well as liquids do, but temps seemed more like that of regular TIM. I wasn't impressed. So I opened it up again and added a drop, spread it out and closed it up again.

From the CPU delidding I learned what using too much LM would look like before I put the IHS back on, but that was with a great contact between the core and IHS where excess LM would simply be pushed out. This hesatsinks contact was however less than an absolute perfect fit (thanks to the cooling pads for one) and required more LM.
The second time around temps were crazy, 26c ambient and 65c at 4GHz with fans on auto, not even full blast. Another try at 23c ambient gave me 73c at 4.6GHz, also fans on auto and they were not screaming at all. So much headroom!

Playing Doom for a while and some other games at 23c ambient the GPU (+120core +120ram) got to 74c and CPU at 4.6GHz to slightly above 80c, fans on auto and not very audible. Again so much headroom left for cooling if it were needed, just a crazy difference that LM made...

Now, I haven't checked for any LM runoff yet so technically I cannot be sure the dam worked as intended, but there's no reason to doubt it gives that little extra peace of mind when traveling. (I am typing this by the way after a long trip with the laptop vertical and handled quite roughly several times and all is well and working.

So, pics.

Note: The GTX1080 loves the extra cooling and ramps up its speed all by itself to 1964MHz, minimum, in a laptop! (but then the fans need to go up high because the power consumption is just too much to keep the fans low. I'm not trying to bake my 6 grand machine (I have a lot of SSD's in there)).

 

So can anybody report on the temporal stability of the conductonaut? I am seeing higher temps, about 5 degC on my CPU, compared to when i had a new LM job on it. GPU also seems about 5 degC higher, same ambient temp (house set at 70 degF).

It's been about 5 months since the original job.


Sent from my iPhone using Tapatalk

 

Maybe change Thermal conductivity number for Coollaboratory Liquid Pro to 80 W/mK according to Coolaboratory's own web page.

Spoiler: COOLLABORATORY LIQUID PRO - 80 W/mk

 

Odd... CLUP is 80, then CLUU shoud be substantially more than that and as such really a lot more than Conductonaut?

 

Maybe a big mess on Coolaboratory web page. But there isn't a very big difference between them. Liquid Pro is more a mess when you apply it. Same also when you try to wipe it off. Probably the reason they developed Liquid Ultra in round number two.

 

Yeah but you suggest to change the conductivity in that list to 80 for CLLP when we all know from actual independent results that Conductonaut beats CLLP by a mile and CLLU sits in between that. So why accept so willingly and propagate that what seems questionable at best? It would mean that the actual values for CLLU and Conductonaut are wrong as well. It's all wrong!

 

I referred only to their website. All of the Liquid metal brands like Grizzly, Phobya and Coolaboratory probably mess with the thermal conductivity numbers. See as well... Phobya Liquid metal push Thermal Conductivity: 40 W/mK. I'm pretty sure the numbers should be around this or a bit under for all branded Liquid metals.

 

I agree that they're probably more alike than different and it's just PR BS. I had two syringes of Phobya laying here and I still bought Conductonaut for my delids just because it felt better with all the reports of it being better. Numbers didn't enter into it.

 

What happens if exposed to temperatures below 9c? Just curious since I live in Brazil and as such is not really an issue.


Sent from my SM-G930F using Tapatalk

 

it'll solidify, very hard to remove

also keep in mind that liquid metal is alloy, thermal transfer suffers as temp drops

 

Yoo forgot telling about density of solid Gallium. It will expanding during solidifying. See the Liquid Metal aka Gallium alloy break the bottle (Same for the heasink will get worse contact-fit with the Die/IHS).


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

Regarding Thermal conductivity:

Take for example Grizzly Conductonaut who promote high (73 W / mk) which is an alloy of tin, gallium and indium.

Tin: Has a melting point of 231.93° C and Thermal conductivity 66.8 W / (m · K)
Indium: Has a melting point of 156.5985° C and thermal conductivity 81.8 W / (m · K)
Gallium: Has a melting point of 29.7646° C and Thermal conductivity 40.6 W / (m · K)

Since Liquid Metal should be liquid and not solid, All types Liquid Metal intended as thermal paste can not consist of most part Indium or Tin. Aka Gallium (The unique melting point of gallium allows it to melt in the human hand) is the main ingredient. Forget about 73 W / mk thermal Conductivity, as Grizzly promotes. It's pure Scam!!! Same as Coollaboratory promote their older Liquid Pro around 80 W / mk.
Liquid Metal could be around 50-55 W / mk at max. Not more than this. As it have to be Liquid!!


Edit. Ekwb.com which is a well-known player in cooling is a bit more sober. They provide thermal conductivity numbers from 20 to 40 W / mK In their description of Liquid Metal. Aka closer to the truth than +70 W / mK.

 

You got the links mixed up I believe. Wiki says Tin is 66.8 w/m-k and Indium is 81.8 w/m-k

http://www.indium.com/thermal-management/thermal-k-list/
This one says 86.0 w/m-k for pure indium
http://overclocking.guide/the-truth-about-cpu-soldering/
In addition, solidifying tin will shrink by a large factor which results in a big thermal tension inside the material. This tension could already damage the substrate of the CPU.
Although there is tin in that mix, it also has a higher indium in theirs.
http://www.thermal-grizzly.com/en/products/26-conductonaut-en

 

Thanks for the notice No one know the amount of Indium Grizzly use. But it can't be very high, as it has the melting point around 156°C. And Liquid metal is wet down to 8C. Indium as you say is 81.8-86 w/m-k vs. Grizzly's Thermal conductivity at 73 w/mk. This mean the amount of Indium should be around 85%. No in Hell this will and can be correct. The main indigrents is Gallium.

 

Maybe it's different with these types of compounds with varying % of any of the other mixes?
https://en.wikipedia.org/wiki/Galinstan
-19C for this version of it


The melting point of Galinstan has been a source of much debate.[ by whom?] Many commercially available gallium, indium and tin eutectic alloys are advertised with a melting point of about +11 °C, which is significantly higher than the −19 °C featured by Galinstan. The official MSDS ( material safety data sheet) mentions only that Galinstan is an "eutectic mixture of the metal components gallium, indium and tin" with no further explanation provided. Additionally, a US patent to Geraberger Thermometerwerk GmbH [6] describes various related eutectic alloys and mentions that they may contain up to 2 wt% of antimony (Sb) to increase oxidation resistance and up to 2 wt% of bismuth (Bi) to improve fluidity. The resulting eutectic alloy is said to contain 68 wt% to 69 wt% Ga, 21 wt% to 22 wt% In, and 9.5 wt% to 10.5 wt% Sn, with small addition of Sb and Bi and an impurity level less than 0.001 wt%. The resulting material is noted to have a melting point of −19.5 °C and vaporisation point of above 1800 °C.

 

Is it normal that CL Liquid Ultra solidifies? This time I had it on my CPU for 3 months and there were no liquid left. Temperatures were horrible.

 

Yes, see my post a bit above. I think that because of the slightly weaker pressure that the laptops heatsinks apply on dies, the pastes are more in contact with air.

Add in temperature and your liquid metal oxidizes pretty fast and turns into this annoying cake.

 

Delivery just landed!

Will get it done later on today

Sent from my Nexus 6P using Tapatalk

 

What kind of temp improvements are people seeing over top teir Tim say Grizzly tim vs LM? Buddy of mine didn't use the 3m tape and cooked his GPU by little over spill.

 

So, have competed the LM repaste and been gaming for a few hours.
Honestly, I'm not as impressed as I had hoped for really.

Don't get me wrong, my CPU is maxing out at 67 (better by a good 13 degrees) which is brilliant. But when gaming for an hour or more the GPU is still hitting 74 (which is pretty close to stock paste temp)

I was extra careful when applying on the gpu, so could well be down to using too little? Still, not bad Temps. Just expected a lot lower on the gpu side.


Sent from my Nexus 6P using Tapatalk

 

Not the big difference with Liquid metal on Gpu. Max 1-4C degrees. Not in the same level as the processor. The bigger die and heatsink will dissipate the heat better to the heatsink grills. You need to know the facts before you jump into such wrong conclusions.

 

pascal gpu never really benefit from liquid metal, only marginal decrease in temps

 

And the reason is listed in my previous post. In addition, voltage for Pascal is fully locked.

 

Thanks for the info.

Going off that figure, its done it's job.
Still an improvement over previous Temps.

Just with seeing users here with Temps in the mid/high 60's on GPU, I did expect a more of an improvement.

Sent from my Nexus 6P using Tapatalk

 

Try HWINFO fan profile or enable CPU performance mode in BIOS. Either one, will give more temps drops.

 

Tried the HWINFO fan profile, but won't allow me to create a custom profile for some reason.

Sent from my Nexus 6P using Tapatalk

 

Enable HPET, EC Support and PCH Thermal Sensor or something. Ask @judal57 for fan profile, after using it my CPU temps <75C and GPU temps are <70C