Insane pump out effect with Phobya NanoGrease Extreme. What did I do wrong?

Applied it like... less than 2 months ago? Temps were fine the first few days then got worse and worse, today they were insanely high (reaching 100°C) so I opened it up and saw this mess:





If anyone can give me a quick reply as to what to do I'd be grateful, as I have my laptop sitting open next to me right now. It's a MSI GE65 with an i7 9750H and RTX 2070. I have Kryonaut, Conductonaut and Phobya NanoGrease extreme at my disposal. I already applied Conductonaut once but removed it afterwards because I was too scared. What do I do now? Maybe I applied the Phobya thermal paste wrong?

I'm going to clean this mess up and hopefully I get some answers by the time I'm done.

 

Maybe some sort of contamination by liquid metal residues?

 

The liquid metal was applied only to the CPU, and the pump out happened on the GPU too. So no, that's not possible.

 

Provided you applied the paste using the pea drop method, no you didn't do anything wrong. I'm actually having the same problem, and I think the thermal pump out from this paste boils down to 1 of 2 things:

1. We got a bad batch of the paste. I've never seen anyone else but you or me have this problem. It still is an insanely good conventional thermal paste when it makes full contact with the CPU and heatsink though.

2. The thermal pump out is a result of the thermal conductivity of the paste. Thermally conductive materials tend to have a larger expansion coefficient than less thermally conductive materials. This would explain the pump out, but I think this is the less likely reason as Thermal Grizzly Kryonaut does not exhibit this behavior.

I've found that liberally applying nanogrease significantly slows the thermal pump out, but you still have to add more paste after a while to fill in the holes left from the thermal pump out.

Kryonaut is second to Nanogrease in terms of thermal conductivity. It is also an amazing paste, but I had to switch to Nanogrease since I kept burning up my Kryonaut in my Ranger (it degrades very quickly at temps above 80°C or 176°F).

 

Honestly I'm getting kind of tired of this lol

First I tried MX-4, people told me it was crap so I switched to Kryonaut, then people told me Kryonaut was no good for laptops and that Nanogrease was the best for them because of the viscosity, but apparently not either!?

I just want a thermal paste that lasts. Put it and forget for like a year at least.

 

It's hard to guess what happened. It looks like you put a lot of thermal paste and it squeezed out the sides to me. There is still paste over the cpu/gpu die itself so I don't think it pumped out. Use whatever paste makes you happy. I've had good luck with all three pastes that you have (Kryonaut, Nanogrease, and MX-4) on laptops for months/years without the "pump out" effect...

 

The only solution that has given me stable, low temperatures was liquid metal. Only caveat is depending on how the copper was alloyed, it may eat more or less gallinstan and may require a reapplication

 

I'm really really scared of it moving around and turning my almost 2000€ laptop into a brick tho

 

Don’t be scared I have used liquid ultra on all my laptops since 2012 that I used to take to and from class and now to work, never had a problem. Only once I had to reapply to my machine early in because it sucked up a lot of the galinstan but after second application it has been perfect

 

How do you protect the parts around the CPU die? Photos would be greatly appreciated. Last time I did I just put kapton tape around the die but it wasn't enough for me to have peace of mind.
edit: also if I do it it would be only on the CPU. The GPU has WAY too many metal contacts right next to the die, while the CPU only has, like, 3.

 

I don’t use protection. I only paint a very thin layer on. It’s very cohesive molecule it doesn’t really fly around unless you’re putting on puddles of it

 

...what lol
And you take the laptop to class like that? You like living on the edge don't you?

 

Hmmm..... Taking into class and things might get worse if its LM.
I think use little paste and spread with spatula or q-tips or cotton buds. I'm using cotton bud to flatten/spread the paste from middle to the sides and if there's little excess apply it on heatsink that way you can prevent major chunk of paste being pumped out.

 

Try either Thermalright TFX or Kingpin KPX.
TFX is better than nanogrease extreme and a very thick paste, so its very good for sub-zero (won't have the cracking effect like Kryonaut can have at -125C, etc).
Kingpin KPX is also a thick paste and designed for subzero as well but I do not have this paste.

TFX is better than Kryonaut on my desktop (10900k, tested two chips and 360 AIO and NH-D15) but still worse than homemade Galinstan LM on my Vega 64.
You can get it on newegg, better to buy the 6.2g gram since its cheaper per gram. Yes it's expensive.
For Kingpin KPX, 30g canisters are available from kingpincooling.com

I suggest giving one of these a try. Try TFX.

BTW if your heatsink is concave and the contact between CPU and heatsink is not fully balanced and flat, almost all pastes will pump out. That's why some people use Panasonic Soft PGS+ graphite pads (Innovation cooling graphite pads are a commercial version of this, but its a bit cheaper to buy them directly from Panasonic/Digikey, but you have to know the correct part number; check the IC graphite thread as this was posted here a few years ago). You can use Innovation Cooling contact paper test to test pressure and balance, or try requesting a free sample of FujiFilm Ultra Low Prescale from the 'sensorprod' website (be prepared to do some diplomacy).

 

Eh I used to do LM for my laptops but the 5-10C difference between it and a quality paste isn't worth it to me. Even with all of the steps to protect it stuff happens. I see people losing their MBs on this forum more often than I would like. Sure if you want to gamble with your laptop then cool, but I'm no longer doing it.

 

It was scary at first yes but like I’ve said, never had a problem. Never applied kapton, conformal coating, or foam dam. Just never went crazy on application amount.

 

When it comes to liquid metal, a rice sized drop is enough. That's all you need. Don't apply any more than that or it will leak out.

The pressure between the heatsink and the CPU should be enough to keep the LM from ever leaking out assuming you applied a rice sized drop, barring some situation where the laptop constantly gets jolted around. Personally I'd put some kapton tape on the metal contacts on the CPU's topside and make a foam dam to make sure the LM has no chance of ever leaking onto any components.

@Raidriar Just curious, but did you ever vigorously shake your laptop with LM applied inside it? I'm just wondering how resistant LM really is to leakage assuming it's applied correctly.

 

No I never shake the laptop lol why to make crazy stress test like that? Only pull in and out of backpack and walking to from class and work

 

Because such a stress test, if successful, should quell everyone's fears about using LM in a laptop .

I know that sounds incredibly silly, because it is, but if someone were to do that and report no leakage, then that should quiet all the naysayers who say LM isn't worth it.

I'm betting even shaking a laptop vigorously won't make the LM leak out if applied properly. Keyword is properly.

Most of the people who have trouble with LM as far as I have seen have problems because they applied too much.

 

Rough or careless handling poses a risk of liquid metal escaping as well as potentially causing other problems unrelated to the use of liquid metal. If you slam it around a lot, like tossing your backpack into the car seat, or throwing your laptop bag into an overhead luggage compartment in an airplane, it can definitely get dislodged. Shipping a laptop with it applied without some kind of barrier is also foolish, as the shippers toss the boxes around like a rag doll. Using compressed air to blow out the inside of a laptop to clean out dust can also blow out droplets of the liquid metal. But, if you use a foam dam and cover the adjacent surfaces like most of us have there is nothing to fear but fear itself.

The thin foam inserts from cheap 120mm fan filters like these work really well to cut for foam dams. That is what I used most recently on my TongFang turdbook to keep the liquid metal from escaping during my business travels.

 

Sorry to say this but i think that this paste is just crap.... I've been trying multiple brands over the years and whilst some would dry out and others would have effect similar to yours i only found one that sort of does the job right which is Grizzly Kryonaut. It seems to hold out for a much longer period of time and it doesn't dry for months and months whilst preserving cool temps. Im not saying it super perfect but definitely better than most other brands i went though over time.

 

Phobya Nanogrease works better on desktop CPUs admittedly. The pump out effect is drastically slowed by the significantly larger contact surface area on a CPU with an IHS and the higher contact pressure from a heavier heatsink.

 

In my case it was too little. Copper will leech out the gallium over time and you be left with a hard layer of metal that flakes off or chips off. Apparently the best method is to us LM on the heatsink first and let it sit for a week, remove what you put on and then apply new LM. I suppose it would be better to use pure gallium so it saturate stains the copper surface as to no longer absorb the LM mixture.

 

You can't use pure gallium. It's solid at room temperature (melts around 28C or something).
My trick still works very well with galinstan/LM.
Sand the copper or nickel plated copper with 1500 grit sandpaper to roughen it up so it isn't so smooth. Clean it fully leaving absolutely no tiny flakes or dust anywhere. You are only trying to roughen up the surface so the gallium has something to adhere to and to absorb into.

Apply LM, pinhead size amount.
Spend 10 minutes with a lint free applicator wiping it around the entire surface, repeatedly, do not apply downwards pressure, just wipe. This will cause absorption of the gallium into the sanded microgrooves you created. You will notice as you keep wiping it that it starts feeling thicker, like a completely different consistency. This is what you want.

Once finished, apply a new layer on top of the old one. (same amount, small drop), wipe it normally and spread it (like 30 seconds), then reassemble.

 

I re-applied the NanoGrase thermal paste and for now the temps are great. However, if it degrades quickly again, I'm considering purchasing this:
https://www.amazon.es/Innovation-Co...oling+graphite+pad&qid=1593524431&sr=8-1&th=1

Any thoughts?

 

I think you better opt for ICD 7/24 paste because I believe you have warped heatsink that's not sitting flat with bare CPU and GPU die.

 

@Reonu I'd recommend Thermal Grizzly Kryonaut. After switching to it, never looked back...

 

You can use foam barriers to prevent potential spillage of Liquid Metal.
They seem to be the go to solution for this thermal compound (especially in laptops).
Here's a useful guide for that here:
http://forum.notebookreview.com/thr...r-liquid-metal-safety-insurance-guide.817207/

I also don't agree with the statement that Kryonaut is bad for laptop use - in fact, it seems to work pretty well on laptops, so I'd actually consider using that in your own laptop before you consider using Liquid Metal.

As for IC Graphite... I don't know how recommended that would be since Kryonaut apparently performs better according to the following:
https://www.igorslab.de/innovation-...eit-siegt-aber-was-ist-mit-der-performance/2/

Also, if I'm not mistaken, you need relatively even heatsink with high pressure for IC graphite to work properly... and in laptops, you won't be able to get that kind of even high pressure (desktops with high mounting pressure cooling solutions work better for this) - I'm sure the OEM's could certainly design a heatsink in laptops to such a degree, however, as you know, cost efficiency takes priority.

 

I also have the same problem with the same thermal paste. Temps went +10º after few days (I move the laptop twice a day) .
I thought of putting a copper pad between the heatsink and cpu/gpu, and make a sandwhich with thermal paste.

If you try, let me know how it works.

 

Graphite pads need a good amount of pressure to be anywhere near as good as regular paste.

 

Even with good pressure it won't be any better than quality thermal paste. Liquid metal FTW

 

has anyone tried putting a thin copper pad, between heatsink and cpu/gpu to add pressure? Adding thermal paste on both sides

 

You’ll get double layer with thermal paste. The extra layer with +-10w/mk vs 380W/mk thermal conductivity won’t exactly offset the gains with copper pad.

 

Still people having problems with pump out?
If you don't want to go liquid metal and balance the heatsinks, try this.

https://www.newegg.com/thermalright-tfx-6-2g/p/2MB-0014-00003?Item=9SIA7WF8UF9204

 

I just find it curious that besides Aliexpress I can't find that paste in stock in any decent/reputable vendor..

Really wanted to try it out.

 

Might try liquid metal pad from coolaboratory.

I think the whole "particle fly off" regarding liquid metal is highly exaggerated and in all my dangerous experiments with LM, I have never had that once happen, and that was when I did a repaste one where the gpu literally had no contact without a copper shim, so what was my solution? Pump in the LM till it was convexed like a dome and let that be good no it didnt fly off curiously.

In short, you will be just fine, just dont pump it on like its paste.

 

Surface tension does a good job holding it in place.

 

Yup, those foam inserts for 120mm fan filters work great.