*OFFICIAL* Alienware m15 Owner's Lounge

I'm not sure if Kryonaut is the best non-LM paste for this notebook. Phobya Nanogrease Extreme and Prolimatech PK-3 should be strong contenders too since they might deal better with uneven heatsinks. I would try one of those. Many recommend IC Diamond too, although it has lower rated thermal conductivity and can cause surface scratching.

You would need either 1.5mm or 2mm (the original) pads for the CPU, and 1.5mm, 0.5mm and probably some 1mm for the GPU.

That's def a very cool idea, although to be fair Gelid Gp-extreme is VERY soft so any resistance would be minimal.

Those CB20 CPU temps are amazing. Did you take note of your ambient?

I notice lower temp on core 1 on mine and other people's CB20 results a lot. I wonder if it's something inherent to the benchmark. No uneven temps in games for me.
BTW Notice the very cool PCH. It doesn't need any active cooling as apparently provided in the R2 for no real reason.

 

Where can I order the double size m.2 ssd heatsink for the m15?

 

Is there a diagram of the original pad thickness and another with the thickness they need to get replaced with? I'm thinking of getting some new pads, any recommendations? Would this void the warranty? Let me know.

Sent from my Moto Z (2) using Tapatalk

 

From Dell. Part number is R46DX

So pretty much every size I'm not sure about IC Diamond because I also read about surface scratching and someone that his heatsink went off with the cpu glued to it breaking the arm retaining the cpu.​

 

That tells me the sizes but which pads need to get switched out? I assume the ones with the impression are the 1mm and need to get swapped for 0.5 mm?

Sent from my Moto Z (2) using Tapatalk

 

Kryonaut is the worst thermal paste ever made! Temps start out really good the first couple of days then you watch them rise daily to a point where you can't even use the laptop anymore.

See:

http://forum.notebookreview.com/threads/thermal-grizzly-kryonaut.790919/page-10#post-10261106

http://forum.notebookreview.com/thr...s-before-i-start.741745/page-65#post-10249996

http://forum.notebookreview.com/thr...iquid-metal-paste.812596/page-2#post-10660500

Phobya Nanogrease is the best non metal thermal paste

 

It's not that bad for me, there is a consistent improvement vs stock (literally 3-6C depending on scenario), and there has been no deterioration so far after adjusting for ambient temp changes but it kind of feels underwhelming. By all accounts Kryonaut performs best in desktops with proper heatsinks, that's why I'm also wondering about Prolimatech PK-3 - not as good as Phobya on paper but conductivity similar to Kryonaut and is apparently designed to deal with crappy laptop heatsinks.

The only pad with a significant impression is the CPU VRM one. It's 2mm by default but 1.5mm probably works too. The reason it's like that is that the components it's supposed to touch are on two different levels above the PCB, so either @CptXabaras 's K5 Pro or a well compressible pad like Gelid sounds ideal.

 

I can't use a thermal paste that's not conducive right?

Sent from my Moto Z (2) using Tapatalk

 

? You should probably read some intro threads / articles on this. A paste should have as high thermal conductivity as possible, unfortunately the best ones are an alloy of the metal gallium which is liquid at room temperature (and even way below). Unfortunately this is electrically conductive and thus can wreak havoc on your components in case of any spills, and reacts with copper sinks so it's risky, and damages heatsinks.

 

I don't need the best, just good enough. I currently have some phobya nanogrease for example. I know I have some pads lying around but I can't find them

Sent from my Moto Z (2) using Tapatalk

 

Liquid metal won’t damage your heatsink if it’s made of copper.

 

Liquid Metal will stain but won’t affect performance of Copper or Nickel so it’s fine. Aluminum base plate heat sinks however are a big no though.

https://www.gamersnexus.net/guides/...cts-copper-nickel-and-aluminum-corrosion-test

 

It's not so rosy though, is it? The GaCu alloy that forms on the copper heatsink has lower conductivity than copper and therefore will compromise thermal performance, and can't be easily removed. It may also cause glueing effect making the heatsink difficult to remove, potentially causing damage during removal, and the alloy layer itself can be rough and uneven.

http://forum.notebookreview.com/thr...tibility-with-copper-heat-sinks.800890/page-4

 

I have used Liquid metal many years. And one thing for sure... Heatsink works exactly as the day it arrived with the laptop brand new.

Heatsink from my Alienware 17 from 2013 after near 4 years with it. The grayish color is what it is only a simple discoloration due the Gallium. But the cooling capacity is exactly as before. No change

http://forum.notebookreview.com/thr...ners-welcome-too.810490/page-58#post-10650404

Same also for the Heatsink in my Clevo P870

http://forum.notebookreview.com/threads/clevo-overclockers-lounge.788975/page-1618#post-10767615

 

I trust you measured this using specialised equipment?

Clearly both heatsinks had to be sanded to look nice, the scuff marks are obvious. The Clevo heatsink looks terribly uneven, the alloy has holes in it. In the thread posted above alone, and not just there, people are saying they wish they didn't go down the LM route.

I personally wouldn't unless I needed to run CB20 like loads for extended periods of time or lived in a very hot climate without air conditioning. For other scenarios a non-LM paste gets the job done for for this laptop with virtually no risk or hassle.

 

I trust what temp I'll get after new application of Liquid metal. If the heatsink wasn't flat or even after the change of liquid metal I would get temp I can't control in my benching. And the grayish color on heatsink help you get even better results the second time you put on new liquid metal.

If you don't know how to do it in first place, then it's a risk... http://forum.notebookreview.com/threads/misjahs-guide-on-how-not-to-liquid-metal-repaste.805338/

 

Thanks, @hmscott summed it up well:

http://forum.notebookreview.com/thr...how-not-to-liquid-metal-repaste.805338/page-3

All those measures people take just serve to reduce the risk rather than completely eliminate it.

 

As I said in previous post... If you don't know how to do it properly... Use ordinar thermal paste. Thats it. Not more difficult than that

 

Is it not now? There is always a risk with LM, and you are either in denial or not being honest about it.

https://www.reddit.com/r/MSILaptops...ng_never_xotic_especially_their_liquid_metal/

 

Have done Liquid metal since early 2014 and on many machines. What I prefer doesn't mean I force it on others. Re-read my post #2668

 

Kind of the big reason I never tried liquid metal. Not knowing if there was any long term risks to the heat sink. Glad to see more evidence that it isn't going do much more than stain the heatsink or affect its cooling. Might have to give it a try myself finally.

 

Great idea, again. Do let us know how any warranty repairs go afterwards.

 

No need being sarcastic brother. Doesn't fit in here. You know there are different skill levels on people/workers? You think all employees out there are equal good in their works, as well in the same company? You think every single heatsink is produced equal and come in perfect shape and form from the factory. See... No differences in tolerance level. All is created equal and perfect. Oh well. I think I stop here.

 

Even most boutique brand's LM application can go haywire during shipping. I'd recommend LM if the heatsink has enough pressure to stay put or the user is very careful guy/gal otherwise correctly applied standard paste should perform better than badly applied LM.
You will need to make changes to paste application which will vary from PC to PC.

 

Right. Most LM guides point out the increased risk if the laptop is used on the go, as is often the case with the m15. Even if the LM application professional is a Papusan-level artisan (never mind an amateur doing it for the first or fifth time), the laptop can be subjected to a shock, such as a fall or knock, and the dam may fail that way or for any other reason. Sure, the risk can be minimised but not eliminated. Can't really go around physics.

 

LM will always have inherent risks. They can be mitigated with proper technique and proper precautionary steps, but there is always a risk. For some people, it's worth the risk. I ran my XPS 9560 with LM for a year and took it to school every day and never had an issue. However that was my choice and if it ruined the laptop, I was willing to pay the piper.

Having said that, LM in the XPS was only 2-4C better than Kryonaut, even after long term use (well ~1 year, I don't really own laptops "long term"). Because of this, I don't run LM in my laptops anymore, it's not worth it to me. If I still had my 51M and I had time to bench it for fun, I'd probably do it there since every little bit helps though...

 

As a LM user my self, i'm aware that the risk is minimized and not completely absent. For that very reason i wouldn't personally let anyone do it on any of my machines, other than my self, since it is the only way i can be sure of how it has been applied, how the preventive measure the avoid or minimize disaster have been implemented.

Regarding the stain it leaves... well there are different theories on the matter. By my personal experience, after second application on an already "stained" heatsink, heat transfer does actually improve a little bit and the stain do not get worst over time. But i repeat, that's just my experience. YMMV.

Do i recommend it to other user? well, yes only if you are fully aware that there is still, even if remotely, a possibility of something going wrong, especially if you move the laptop aroud a lot, and if said user is willing to assume the full responsibility of it, like a nice grown up man. And i do move mine on a daily basis.

Ambient temp give or takes is around 23 - 24 °C . I've run the test from inside my site office, i'm in the middle of the desert and there are >44 °C outside, AC is working hard to keep inside temperatures low.

 

It's just about weighing up cost/risk vs benefit, which nice grown up men and women should be better at. In most scenarios even my amateurish Kryonaut application is absolutely fine. In lesser benchmarks such as CB15 or TimeSpy the CPU stays in the 80Cs, games don't even deserve a mention, and I don't need to run CB20-level extreme CPU loads on extended basis, therefore I don't see any point in assuming the LM risk essentially for no benefit other than temp bragging rights and perhaps slightly extended CPU longevity which is of no concern to begin with.

 

And you care what happens to my stuff why?

 

Knowing what would happen to your stuff would be very helpful for the entire community.

 

Didn't absolutely mean to sound as personally attacking your statements on the argument, if my post did sound so, i do apologize.

 

No worries, I didn't mean to suggest any offence on my part (none was taken) nor did I mean to cause any. Just a bit of a heated argument on a hot topic

 

Yeah, the fear of liquid metal is largely displaced. Other than the hazard associated with being sloppy and allowing it get onto something to cause an electrical short, there is no reason for anyone to fear it or avoid it. Being sloppy is going to surface sooner or later anyhow... if not with liquid metal, then something else. Being cautious and deliberate is important with all things where a good outcome matters. The people that are so dead set against it either don't actually need it or they are misinformed. The "stain" it causes to copper is irrelevant.

 

You lost me at "largely".

How cautious and deliberate does one need to get? Do I need an industrial robotic arm to perfectly reseat the sink? Do I need some sort of probe to examine the dam for imperfections that might cause it to fail? Do I need a clean room for this?

As for the copper sink reaction, it's not clear it's as inoccuous as you portray. For one thing it, the copper reaction causes the TIM layer itself to deteriorate. For another, it's not clear whether manufacturers would view the presence of LM alloy kindly during warranty repairs.

https://linustechtips.com/main/topic/1028665-warning-about-using-liquid-metal-from-thermal-grizzly/

Apparently, Conductonaut works best with nickel-plated sinks.

Again, it's all good if people proceed to use LM with a clear view of the benefits this will bring given their particular usage patterns, as well as very good understanding of the additional effort needed, along with the risks and downsides.

 

I've been using it for years with no issues and so have many others. So, yeah, "largely" is the right word. If results improve, and no damage occurs due to misuse, and it continues working well for many years, no point in being a worry wort. User induced damage of any kind is not covered under a warranty, but a warranty isn't predicated on using a particular product for thermal paste. If they are going to do that it needs to be in writing. The benefits of using it are clear and unmistakable, and for those who have benefited as greatly as many of us have, the risks are very small and most definitely worth it. The only real risks are applying it to aluminum or getting it on electrical contacts.

But, it's good to know what the risks are, and I wouldn't recommend anyone that does things in a haphazard manner or has a knack for screwing things up go into the situation thinking it is an idiot-proof proposition. It's not idiot-proof, but it's definitely worth doing if you want the best results.

Those considering using it also need to know that it will not be useful or beneficial if their heat sink fit is sloppy. Liquid metal cannot be used to plug gaps and their temps will be worse, not better, if their parts do not fit correctly. That kind of nonsense can only be partially accommodated by using a really heavy thermal compound like IC Diamond or Phobya Nanogrease Extreme. Gelid, Kryonaut, NTH1 and all of the similarly thin and creamy thermal pastes will also produce poor results with sloppy fitting heat sinks. (They may be good for a short time, but won't last very long if the fit is sloppy.) In fact, if you get decent long term results with those thin and creamy thermal pastes, you likely have an ideal candidate for liquid metal.

Repasting with CooLaboratory Liquid Ultra, any tips before I start?

[Liquid Metal & Traditional Paste] - Clevo P775DM2/3(-G)/P75xDM2(-G) (Sager NP9152/NP9172)

 

OK, so your argument is basically: anyone who has experienced an issue with LM had it applied by a clumsy idiot. Nice.
Also, not clear manufacturers need to specify every single user action that might have led to damage in writing. If the mobo is bricked or malfunctioning and there is obvious evidence of LM use, they would have an easy time assigning the blame to the user.

Regarding longevity, since we've skipped that thread, apparently LM needs frequent re-application in copper sink setups, even as often as every few months, due to the ongoing alloying process. Not particularly convenient, is it?

The temps you posted basically corroborate what @custom90gt said earlier about the difference in temps vs non-conductive paste often ending up being minor (that said, I'm super impressed by @CptXabaras 's results!). The major temp drop comes from deliding.

Anyway, thanks for the info on traditional pastes - another data point in support of Phobya on predominantly poor fitting laptop heatsinks. Has anyone tried it on the m15 and can post some CB20 temps?

 

Nope, didn't say that. Smart people can make stupid mistakes. Lord knows that I've made my fair share of them. That's not the fault of the product. It's called user error and it is quite common. But, it is absolutely true that clumsy idiots definitely shouldn't use liquid metal. In fact, they shouldn't work on their PC at all. Anyone that is a clumsy idiot would be demonstrating proper awareness of their shortcomings by paying someone that isn't (hopefully) a clumsy idiot to do the work.

If you are dealing with a dishonest manufacturer that is looking for a way to screw their customers and finding creative ways to make up stupid reasons to shirk their responsibility, maybe so. They are few and far between, and there would be no reason for them to blame the user unless the user applied the product in a way that actually caused the damage. It's not ethical or legal to arbitrarily void the warranty simply because they do not approve of the choice of products their customers elect to use. As much as it might not seem like it sometimes, most of them are smart enough to know that. It has to be relevant to be relevant.

This is totally false. It outlasts ordinary thermal pastes by an extremely wide margin, which has been proven time and again by many veteran liquid metal users in the enthusiast ranks. The exception is when the fit is poor. It will definitely dry out and fail, but not really any faster than most of the ordinary thermal interface materials that are low viscosity. They will also fail in short order, and deliver poor results in the interim. That's due to the sloppy garbage hardware that the TIM is being applied to and not the TIM. Sadly, it's very common for laptops to be sloppily-made pieces of garbage.

Yes, major temperature drop comes from deliding... if you use liquid metal between the die and IHS. If you do not, the improvement ranges from minimal to none. That is because the non-conductive TIM originally placed there is garbage and the aftermarket alternatives often aren't a whole heck of a lot better. What one considers minor improvement is subjective. A 5°C improvement in load temps (about half what is typical) is actually huge whether it sounds like it or not. It is a larger improvement than most will experience doing all manner of other home remedies in search of acceptable thermals. It can mean the difference between thermal throttling or not thermal throttling. And, 10°C (which is fairly common) is far more of an improvement than just about anything else one is likely to see from such a simple and inexpensive change.

It's OK if you don't like the idea of using liquid metal. You're definitely not alone. Neither are the advocates of liquid metal... we are legion. It's OK to have an opinion either way, and it's OK to be scared of things that are unknown or avoid things that one has no experience with if you're not the adventurous type.

 

"This is totally false. It outlasts ordinary thermal pastes by an extremely wide margin"

LOL? I've got one desktop pasted with Arctic Silver several years ago and still running fine OCed so you got carried away there, but it's good to know that those reported LM reliability issues with copper heatsinks are not as bad as many claim. Obviously you are particularly passionate about LM which puts the balance of your opinions slightly in question. Your thinly veiled and unkind suggestion that people who don't share your LM passion in scenarios where there simply isn't all that much to be gained from it (locked BGA CPU, running fine as is, likely suboptimal heatsink) are "scared" or "not adventurous" is fairly infantile, I'm sorry.

BTW Please ask @Flying Endeavor how his professional LM application resulting in 50C idle temps is doing, maybe you can help out.

 

you do now that @Mr. Fox and colleagues have laid a pretty substantial basis for all the most effective mods that can be applied to DTR's and now in respect BGA c*ncerbooks right ? have you googled his name just to see if he has any credibility ? you got one desktop while i can assure you @Mr. Fox has at least one desktop (for himself) one for the benching/tuning and maybe a dozen old DTR's laying around all LM'ed up. my alienware 17 R5 is thrown around when traveling (i always ship it to my destination in advance) so the fact the my alienware mx 17r2/alienware 17/alienware17 r4/alienware 17 R5 all survived LM without foam dams just the right amount you need and scotch 33+ tape proves his point.

Also true LM like conductonaught will not dry out, due to it being more metal than addative. The LM pastes that dry out have more addative than Metal within the compound making it brittle and crack when exposed to longer duration of oxygen. Metal can't dry out. he did not say that Artic Silver is bad or dries out. You won't need repasting if A - the paste is good quality and B your heatsink is flat and level on your CPU. Just like he stated Quote: "The exception is when the fit is poor."

LM is the way to go, and if it was so bad as you descibed Asus wouldn't use it out the factory: https://www.notebookcheck.net/Asus-...omises-13-C-cooler-temperatures.422700.0.html

It's been proven time and time, LM is the way to go especially in laptops where you have thin copper pipes and not much room where heat can flow (like a desktop with a good airflow for the mobo)

 

Calm down. You seem to have exceeded Mr Fox's level of enthusiasm by claiming that LM is the way to go in laptops period, whereas his statement was conditional on good heatsink fit which makes far more sense.

Manufacturers using LM is another matter altogether, as they would be taking care of the reliability of the solution and assuming the risk. If it works out in practice, which we'll find out in a few years - great! In the meantime Dell could stop using toothpaste as their TIM.

 

how you interpret it is your choice, we just give you the information so you can deal with problems yourself. you talk about enthusiasm and the next minute it's about a statement so maybe you need to look at how you define things, and not how we define things.

offcoarse his statement makes far more sense, i didn't make a statement lol. i made multiple claimes that can be seen as factual one including that all laptops need LM which i stand by.

There are almost no laptops with active cooling that can hold boost clock, while a few years ago Intel wasn't even playing the games they do today (like measuring TDP on the baseclock while it used to be boost) Because all laptops nowadays Throttle (not being able to hold their boost clocks due to them becoming thinner and smaller and OEM's not investing money) LM has become mandatory if you don't want a 1.2 cm laptop that get's as hot as 100C on you lap (like how Apple does with their i9 MACBOOKS)

his statement is more conservative then my claims that i presented as facts. a statement is a whole different ballgame than multiple claims

 

FYI My i7-8750H can hold 3.7GHz out of 3.9GHz under CB20 load without LM (plus I'm not sure how great my heatsink fit is given what seems like a minor gain from Kryonaut repaste, in the region of a few C) that's why I personally have relatively little to gain by embarking on the no doubt exciting one way LM adventure.

 

Great ! But only using your CPU and measuring your heatoutput is not the same as gaming. When gaming you're using your PCH, your SSD your GPU your CPU, so your whole motherboard gets hot because off the electric activity.

Since AW uses shared heatsinks, heat bleeds. So heat from the GPU goes to the CPU. Maybe in the first half our of gaming you'll see boostclock. But the current AW chassis are not build for sustained heat so it starts to throttle your components. With LM you're removing that bottleneck cause LM can tranfer more heat in a shorter amount of time.

When your chassis starts to get hot your system is throttling for sure.

 

LOL, we are finally talking about the m15! Perhaps admins will have mercy The only concern is performance in CB20. CPU temps during all other activities, including graphics benchmarks and CB15 stay in the 80Cs tops, max GPU temp is 67C at 25C ambient. PCH temp is of no concern, between 40-70C, likewise the Samsung SSD. There is no thermal throttling whatsoever, sorry.

 

Haven't mentioned the m15 cause i don't own one. Your temps seem fine but depends on which m15 you have. 3.7 out of 3.9 is not boost so you're kinda proving my point.

Your temperatures don't say much without context (which program used for determining when it throttles/if it throttles/if it was combined load/which paste etc) but those temps are mighty fine

 

Thanks, I did mention CB versions. By graphics Benchmarks I mean TimeSpy and its older but less taxing DX11 cousin FireStrike.

As I'm sure you know, anything above the base clock (in this case 2.2GHz) is Turbo Boost, the difference is whether it's short or long TB. By changing TDP limits they can be made equivalent if the cooling solution supports it.

This TDP throttling is only really relevant in CPU benchmarks. In games, the power draw is nowhere near the default long TB TDP and thus the clocks are at max 3.9GHz pretty much all the time (apart from loading screens maybe when Speedshift dials them down).

 

True. But i won't call 2.3/2.4 exactly turbo boost, as the two perform identical to 2.2 Ghz. i don't tune locked down chips so i don't have any information about them, but i've maxed out my turbo boost time, and TDP itself on for example my surface pro 3 and i can see it became hotter due to it being locked @2.9 Ghz but even when it crosses the throttlepoint it won't clock back but that's because i turned off all the safeties and limiters that clock the CPU down.

ubisoft for example games use Denovo DRM or AVX for the number crunching. Both will cross your TDP as they're using instruction sets within the CPU that generate heat. Alot of heat. Star Citizen, a game that's not even using AVX or Denuvo sits @an fairly comfortable 50/60 watts (when loading the game it can generate up to 75 Watt off heat) so i don't know which games you play but there are more than enough software packages that can cross the TDP easily while using it normally. But it could be that you have no need or interest in those kinda programs.

Flight Simulator X is an simultor that i play @4.5 all core due to it being very Mhz depended so FSX also generates around 65/70 watts easily. Especially when flying in dense crowded areas.

if i set my cpu 4.3 all core they'll stay 4.3 all core, except when i try to cross my TDP but that's another mather. Throttling due temperature is not the same as throttling due to a powerlimit that's been hit (for example 4.5 all core is around 90C in cinebench R20, so i have an 3/4 C before my CPU starts throttling but i still cannot go past 4.6 or bench 4.6 Ghz due to that clockspeed instantly crossing the 110 W tdp LIMIT aw imposed on the R5

 

Which laptop have you got exactly? 17 R5? What CPU? My 8750H draws about 63W during short TB. With LM I might be able to keep it at that power and 3.9GHz (maybe @CptXabaras can confirm), if I was willing to routinely run the CPU that much above the rated long TB TDP. As it stands, I can run it at 56W 3.7GHz without thermal throttling and that's fine for my needs. Perhaps one day I will feel like the extra 200MHz would move the needle in practice, bearing I mind this downclocking to 3.7GHz only happens during max utilisation of all cores lasting over 28s, i. e. very rarely in practice.

 

Alienware 17 R5 [Cpu - i7-8590hk @4.4] - [Display - 1440 gsync] - [GTX 1080 | 1789 Mhz core - 2045 boost | Memory + 400] - [Tobii Eye tracking] - [Kingston - HyperX 8 Gb Ram] - [Raid 0 500Gb - Kingston HyperX Ssd Pcie Predator] it's all in my signature.

both short and long are set @110w when you turn on OC. when it's turned off short is @65 and long @95. The standard Intel TDP specs.
If you don't have a need it's fine. 3.7 ain't bad that's hellaquick (i'm a tweaker and i tweak the shhhht out of all my electronics that i use) but for me personally i wouldn't be happy when i bought something and it can't be used @100% even though the price i paid was 100%.

but to each to his own, if i set a speed to stay locked, it should stay locked. There shouldn't be some c*ncer firmware that decides what happends with my laptop (while i'm the owner of the bloody machine !!). But we probably use very different software packages and have different needs. I for example cannot buy an gamelaptop/DTR with a locked down chip. Cause performance degrades over the years (like what happends with intel after al the mitigation patches have been installed) and when it degrades, i just OC my CPU or GPU a bit further down the road when the performance is needed. Apart from my GPU my CPU can last me a few years (3/4) or so, but i'm feeling that my 1080 is only going to give me joy for two more years before the games that come out cannot be played @1440P with unlocked frames.

luckily we have the option for the amplifier, but this is my first BGA laptop and i'm already tired of it (actually my R4 was the first BGA laptop i bought, but then it broke down on the last day of my warranty so Dell upgraded me to the R5)

After this, it's PGA AMD CPU in a laptop, or else i'm moving to an desktop.

 

Anyone who doesn't want to get stuck on BIOS 1.6.2 make sure you don't update to win 1903, it's now auto installing the firmware on every reboot. Sitting at 60% fans on battery with temps in the low 40s. The new BIOSes are absolutely atrocious.