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

I use both all the time, but I do not see that much variance between them. They are normally within about 2°C of the same. I think the data polling rate may be a tad slower on ThrottleStop and that could account for the difference when temps are on the rise.

 

that's interesting, on mine the difference is always around 6 - 8 degrees, not only on the rise or going down.

EDIT: ontopic: i've tested both methods of applying liquid ultra to cpu and heatsink and only to the cpu now quite a few times, and in my experience as of the temps i saw using both methods it really doesn't make a difference. at least on the same heatsink (where the contact to the cpu stays the same).

 

This stuff is amazing.

I think it was the old thermal compound I was using or I wasnt applying it right or the compound broke down over the years. I was using arctic sliver 5 for many years. I had the same tube, the biggest one they had, bought at Fry's.

For some reason I could not get the cpu, 3920xm above 4.5. I checked the temps and they were very high, around 100C+. I tried to get go 4.4 but the temps were still very high, over 95C+ and it would throttle.

I bought a tube of this stuff over 1.5 weeks ago but I was nervous in applying it. It must of been the warning in the back of the box, "do not apply any .....over aluminum heatsink or dont get any on it". Or just plain scared to ruin a very expensive laptop.

I paid $13 something on Amazon for a very small tube of it. I thought, wow this stuff is very expensive for the amount you get. I grabbed a tube of the largest Arctic Silver 5 tube at Fry's for about $27 many years ago. But I had to apply it since the arctic silver 5 has broken down. I repasted this alienware laptop many times to try to get the temps down but nothing happened. Even on my other work laptop, the temps were spiking upwards.

I applied it, I applied a very small amount like people said. The amount of less than a rice kernel. I reassembled the laptop up, it was the fastest tear down and built up ever for me, around 20 min. The alienware started right up, I was really thinking I hope I dont see flames or black smoke coming from the rear or flames shooting up from around the battery area. But my fears were gone when it started up without problems.

I checked out the temps with hwmonitor at stock speeds and "WOW!!" shoot right to my head. This stuff really works and it works really good. I dont remember seeing temps this low ever. I dont remember the actual temps, sorry.

I proceed to give it some overclock on the 3920xm to 4.1, temps were still very low. The fans are not even at high yet. I finally went all out and tried 4.5. With artic silver 5, at 4.5ghz, the laptop would get over 100C and throttle down. The temps with liquid ultra are at 92C at 4.5.

I did not go any higher since I dont need any more speed.

This processor, i7-3920xm, is sitting at
4.4ghz
vcore at 1.321--i might try for lower vcore in the future, but its stable and I love it
temps at 88C
room temp is kinda warm, maybe around 77F. I like it hot.

I might see over the next week how high i can overclock this CPU. BTW, I run this CPU 100% 24/7 running BOINC, its like Folding@Home. I run my processors harder in 1 day than most people do in several weeks.

I also repasted my work laptop with liquid ultra, the temps dropped about 10C. That one has a 2670qm and also runs 24/7 at 100%.

I wonder how long with it last at 24/7 operations.

 

So finally i exchange my M18xR2 3920XM CPU IC24 paste to C.Liquid Ultra compound.First time i use very little amount only CPU die.The results was on very bad.When run 1024M test at 4,3GHz 10-15sec later 2 die was ~100°C.Need again disassemble my machine...When i removed the heatsink,i see at once where the error.I looked brother Mr.Fox old test pictures and was on same the results my tests Thermal Testing - Liquid Ultra - Fresh Paste - Imgur.All die the temp not the same.2 die always higher ~5-7°C.I think we all use 0,5mm thermal pad the CPU transistors place and this not good idea
.Please see my pictures:


Here my results: All test was on still 4000RPM fan speed,my room temp ~23-24°C

ICDiamond24 vs Liquid Ultra 4GHz:


ICDiamond vs Liquid Ultra 4,2GHz:

 

Great work Well I have had experienced cooling issues in the past even using softest thermal pads of ~0.5mm on the transistors indicated by you, since I have removed those and have not used anything since. However, I may put some thermal paste next time as that wont disturb the CPU die placement on the heatsink and some extra heat transfer from components to the heatsink could be just beneficial for longlivety of the system in the long run.

Good job

out of topic question: can system board be removed from the chassis along with CPU heatsink intact installed on it or has it all go apart in order to remove motherboard safely? Thanks

 

So can i only 6 pictures upload per one post

Here some new results:

ICDiamond vs Liquid Ultra 4,3GHz

View attachment 109373

Liquid Ultra 4,6GHz fun test

 

I cannot remember now if the M18x, 18 or both. At least one of them you cannot access one of the screws without removing the CPU heat sink.

Great job!

 

well, its the case with m18xR1/R2 at least now Anyway I had a go in reverting back to black chassis and did looked in to checking how my LiquidUltra TIM holding on there, and to my surprise it was doing great, just like I have had it just applied still in liquid form. There were few things I didnt like and few things I must mention:

If liquid ultra was applied little too much or has squezed out and escaped the surfaces - expect fatal problems if it ran trough the system (thankfully not my case)
RTV silicone I used was soft and easy to remove from HS but not from CPU PCB. I had to unplug heatsink along with cpu, then gently twist the CPU away from the heatsink so it did came off pretty easy I must say. After cleaning and spending ~half hour gently scrubbing the RTV silicone of the CPU PCB I thought - I got to get a way around this awful procedure or I had to dump the LiquidUltra as TIM... well I loved low-low temps too much so I have came with brill Idea I have now covered entire CPU PCB surface with clear sealing tape (but not DIE) and so next time I need to remove this damn thing I just peal the tape off along with the goo and Liquid ultra run-offs. I am happy I have used this method and used extra care applying this silicone-dam, I encourage you all do same precaution next time repasting. This time i took extra measures and after inspecting the HeatSink I think I can see LiquidUltra eating bits of copper at certain areas, these are microscopic size holes if can be called so, but are traces of something that wasnt there before! I suppose its the quality of the heatsink to blame, as if it was more pure copper this wouldnt have happened, or maybe Im wrong and its the LiquidUtra doing its magic...? I wonder how will same heatsink will look 6-12 months down the line after one single application lasting this long...? Anyway Im not too stressed about this, heatsinks are cheap to come by, so I just primed both surfaces this time CPU and HS and added more silicone (different, softer, weaker automotive silicone) knowing it will be easier to remove with tape masking on CPU PCB too... and applied some MX-4 on mosfets nearby CPU socket, ones that heatsink hover over them, and last thing last - I have forgotten to place mSATA pcb back on the inside of the black chassis! pitty me!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Didnt wanna tear down again immediately as my TIM job has just been put together and that has to come off, be cleaned, reapplied again etc, and its damn late....Ah I suppose few months down the line maybe I will be more curious how is my TIM doing and I will do complete teardown again and then I will put things in place as they should have been.... well night is killing me so no surprise I have forgotten one of most important things.. At least nothing else left forgotten

Loading ton of photos of this evening job, so its for you to gaze and judge



























Nite!

PS: This teardown was done ~28 Days Later reminds me the zombie movie 28 Days Later

 

Great info here. I'm glad all is well I inside your beast. I don't think there is cause for too much concern with the Heatsink getting eaten. It looks like some serious staining though...meh
Nice little trick with the silicon too. That would be great for anyone feeling brave on the GPU.
Lastly what a Royal PITA you forgot to reinstall your mSata. The MB has to come out just to snap that cable on

 

After seeing UltraGSM's photos, I could not stand it any more and decided to take mine apart to see what it looked like after 6 weeks+ of heavy overclocking and it was awesome. Everything was perfect and paste evenly coated. There was no mess or "leakage" at all. I was very impressed with what I saw on the CPU and GPUs. Now the heat sink copper plate has a silver rectangle that seems to be smoother and less porous than before, so the gallium apparently filled imperfections. I did use the green scuff pad on the copper (did not need to on the die surfaces) and they looked very nice when I was done prepping them for re-paste. Now I wished I would have just left it alone and not wasted my time and thermal paste, but as they say, "curiosity killed the cat" LOL. Luckily, I'm not dead yet and neither is my über-beast.

 

I wonder if other thermal pastes would work with that residue stuff on it. Or if once you go gallium you can't go back

Sent from my SM-N900T using Tapatalk

 

They say that about other things, too.

More seriously, good question and I wondered the same thing so I used IC Diamond on the GPUs after putting it back together to see what would happen... and guess what... nothing. It works perfectly and no problems whatsoever.

 

I'm really happy to see there is no movement of this paste. It definitely has inspired confidence.
If I ever manage to source another PSU for the dual PSU mod I think I might decide to use this stuff on the GPU's since I'll likely run stupid high OCs for gaming sessions.

 

I guess I'll also add some more to this thread.

After 1.5 weeks of non stop 100% 24/7 operations at 4.4 at 1.301 volts. The temp is still sitting at 88C with room temp around 75-79F. It goes up 90C sometimes but it might be from some added dust in the HS.

I'll open the laptop up in about a month to clean out the fans and maybe or not, to look how the paste is doing.

 

Nah, don't look at the paste. I did so a couple of days ago and it was a waste of time and thermal paste. Based on your temps, things are going really great. It certainly would not hurt to clean out the dust if there is some.

 

Ok then, I wont look at the paste. I will still clean it, its a monthly shore to clean this boat anchor. I live in a desert city, out west, so it's normally dusty year round.

 

had to strip away liquid tim on one of my GPU and CPU for my heatsink mod. dual cooling fans baby!! im looking at 4.8ghz OC with ic diamond less than 84C with intel XTU for an hour straight. only problem is my 2ndary gpu also run at same temp as my CPU LOL. this is without any lift, laptop on flat desk, both fans at 3500 RPM.

will try liquid tim again when I get this heatsink to fit properly with palm rest piece. crossing my finger for under 80C at 4.8 on CPU

 

Any pics on your HS mod or dual-fan mod? By the way did you do any close-up visual inspection of your CPU/GPU heatsinks after liquid Ultra application removal and how long have you had them applied for before removal?

 

Well, I'm taking the plunge. Ordered some Liquid Ultra today. Plan on doing at least the CPU, maybe the GPUs too if I'm feeling courageous enough. (CPU has no exposed components, GPU does, so the latter is a lot riskier.)

The paste needs to be applied to both the CPU/GPU die and the heatsink, right? What's the best way to make sure you cover only the contact area on the heatsinks? I'm worried that I'll spread it over too big an area on the GPU sinks, and that it'll end up contacting the exposed components and blow the card.

Oh, also, the pack DOES come with one of the brushes I'd need to spread it, right? If not, what sort of brush do I need?

And in the unfortunate event that I DO get some of it on exposed circuitry, how do I safely get it off?

Appreciate the help.

 

I wanted to share my results with this TIM:

http://forum.notebookreview.com/ali...18xr1-r2-18-benchmark-thread-part-3-a-95.html

 

Only use a small amount, like half the size of a grain of rice applied to the middle of the GPU die, then using one of the two included brushes (which you can clean with alcohol afterwards for reuse) paint the TIM onto the heat-sink surface up to about 2 mm of the edge, you will need to eyeball the size of the GPU die and use that as a reference. It is amazing how little TIM you need, with that same half a grain of rice amount, continue to pain the actual GPU die. Be sure to get all of the former TIM off, using the included scouring pad for good measure. Don't worry about scratching your heat-sink BUT DO NOT USE SCOURING PAD ON CPU AND GPU DIE. Be sure to only have the thinnest layer of TIM on the each surface, and then carefully mate the heat-sink with each respective die. Try to avoid mating the surfaces and then aligning the mounting screws with their respective holes. The heat-sinks need to be applied directly in place from above with as little movement after the surfaces mate. I didn't even use thermal pads on the exposed contacts on the GPU die, just be careful to not use too much and you should be fine.

Thoroughly inspect the exposed circuitry surrounding each GPU die for any TIM and if some is discovered simply use cotton swabs and a jewelers screwdriver to get it all out before powering up.

See my post above for more information.

 

Thanks for the info.

Does the paste really need to be applied to both die and heatsink? Most of the application instructions I'm seeing online only apply it to the die itself. (Or, rather, to the heatspreader in the case of desktop CPUs.) Will it produce poorer results if not applied to both before the heatsink is replaced?

Also, your post above suggests that there's no point using it on the GPUs...I am considering applying it to the CPU only, but my GPUs are running far hotter than your 680Ms. My GPUs hit about 80C at 100% load. (Or, at least, my primary does. No idea how hot my secondary runs because 90% of the time, the temperature sensor doesn't seem to be functioning and nothing will show its temperatures.) Most of the reason I want to repaste the GPUs is just so the fans spin up less while gaming, I don't plan on overclocking them. Any thoughts on that?

Oh, and as you've used IC Diamond in the past, is it a pain to remove it? I've got some ArctiClean, should that be sufficient for getting rid of the ICD?

I'm assuming that brushes ARE included in the pack, right? I can't see them in pictures of the CLU blister pack. And what can I use to clean them if I don't have any near-pure alcohol to hand? Would the ArctiClean work, or perhaps a dash of vodka?

 

Applying it just to the die will suffice but ether meathod has be proven not to hurt.
As for cleaning the die, many products will work. 90% isopropyl alcohol is what you should look for. I use this kit

 

yep will do when i take it apart. i am in middle of switching my main machine from R1 to R2 and I just finished all my mods so I can finally set up my OS LOL. just glad all these experience will bring future machine hope even with DELL and bunch of other company locking down machine with crap BIOS and design I can still make it into a real DTR.

 

No actually there is a reduction of 4 C on the secondary GPU which is now in the primary GPU slot, the other GPU might be 1 C cooler but I am not positive. Look for my post about mid-way down this page where I compare two Metro LL benches:

http://forum.notebookreview.com/ali...18xr1-r2-18-benchmark-thread-part-3-a-95.html

I just used rubbing alcohol and an old cotton T-shirt that I tore up into disposable rags to get the old IC Diamond off of the die and heat-sinks.

Are you using manual fan control? Is the back of your laptop elevated? I usually set the fans to 3700 RPM for gaming, with some games, say Skyrim with an ENB or Crysis 3 needing a full 4200 RPM, this helps a lot with the temps. Also, I find that I have to go and blow out my fans about once a month, just doing this will reduce the temps by 4 C or so.

Yes brushes are included. Vodka might work, dont get too much in your mouth! (at least wait until after you change the TIM for that)

Oh and I just noticed you have 7970M which I believe inherently run about 5 C hotter than 680M with 80 C being normal expected operating temperature.

 

Alright, I think I'll do it on the GPUs as well. For the sake of quieter fans, if nothing else.

No, not using manual fan control nor is the back of my laptop elevated, I've never had the need for either. (How would I control the fans manually? The couple of programs I've tried haven't worked) I already blow out my fans regularly as it is. My current temperatures aren't a problem for me, I'd just like it if I can keep (or improve on) these temps while having a quieter machine.

Thanks again for the tips.

 

Well, I think I'm sold on Liquid Ultra. After I run out of IC Diamond, I'll switch over to LU.

This thread has been an amazing resource thus far. I appreciate all of the feedback from everyone about this compound.

 

Oh man, you really ought to be elevating the back of your laptop if youre seeing 80+ C on your GPU! Now I am not entirely positive, but I do remember seeing significantly more 7970 failures on the M18x R2 compared to 680M. Could it be the sustained temperatures of 80+ C? Its entirely possible. I mean to be honest, I am not running 1.05 V on my GPU's as I don't like seeing 80 C+ sustained, for example playing Skyrim with an ENB for hours on end.

Here's how the cooling adds up:

Simply elevating the back of the laptop: -7 C (CPU)

Manual Fan Control via Hwinfo64: -5-10 C

A good TIM, say IC Diamond: -10 C or Coolaboratory Liquid Ultra: -20 C (CPU)

Combined total:

~ -15 C GPU, ~ -30 C CPU

I HIGHLY recommend elevating the back of your laptop and manually setting your fans to either 3700 RPM or 4200 RPM (if necessary).

If youre seeing 80+ C and you have no overclock whatsoever, that is HOT. I mean, look, I'm running clocks that are ~25% faster than factory (~1k core vs. 758) and considerably more voltage (1.025 V vs .96 V) with performance that is roughly 25% faster (factory 680M SLI 36k GPU in Vantage or so, Im at 46k right now) and I don't see temps greater than say 71-73 C. In fact ~70 C are temps a factory 680M regularly sees sitting flat on a table with factory fan-control, If I was doing that with this overclock I could expect to see 85-90 C!

There really is no excuse for not elevating the back of your laptop or using manual fan control.

Here is the download page for Hwinfo64, you need 64 bit, when you get it going look at the bottom of the window and between "Reset Min/Max" and "Logging Start" there is a little square icon, click on that and in the new window move the slider to either 3700 RPM for gaming and if need be 4200 RPM for intensive games (Crysis 3, Skyrim + ENB etc.) and click on "Set Manual". When done simply slick on "System Auto". I go between 3700 and 4200 RPM as it has been my finding that 3700 RPM is the fastest RPM you can go before the fans hit a whole nother octave of loudness with 4200 RPM being the limit in fan-speed. I game with Astro A40's, they provide sufficient insulation against the noise.

 

If 80C is high enough to result in a significant GPU failure risk, then my 7970Ms are turds that will fail at some point anyway - the 80C will just make that happen a bit sooner.

Honestly, if they do fail...well, that's why I bought a 4 year warranty. But I really doubt that 80C is enough to kill the GPUs. If the 7970Ms had a significant failure rate, we'd have seen a lot more threads about it by now, I suspect. (Unless it's another 8800 situation, where the cards consistently burn out years down the line...but after that disaster and how it hit nVidia financially, I doubt we'll see something like that anytime soon)

I don't really want to elevate the back. I don't have anything particularly stable that I could use to elevate the back, and it would be a less comfortable typing position for me.

I'll take a look at manual fan control, though. Thanks for the tip.

I game with open-backed headphones (Sennheiser HD558's, to be exact), so if the fans are loud and my game isn't, then I can hear the fans. Trust me, after using a pair of proper headphones, you'll never go back to gaming headsets again - the difference in sound quality between the expensive headsets and similarly expensive headphones is astonishing. These things cost the same as my G35 headset did a number of years ago, and while the surround sound simulation on the G35 was impressive, they come nowhere close when it comes to sound quality.

Anyway, I should have my CLU in a couple of days, so I'll see how that affects temperatures.

 

80°C is not going to damage your GPUs. Running cooler is always better, of course... but that's not hot enough to hurt them.

You would not need to elevate the back very much at all. Half an inch is probably enough to enhance air flow.

 

Although 80 C is considered normal operating temperature with recent AMD mobile GPU's (and even desktop for that matter) one has to wonder if there is a correlation between these sustained operating temperatures and premature failure. To be honest, I do not know enough about computer technology to make any kind of authoritative statement but from what I gather there does seem to be a significantly greater failure rate of 7970M versus 680M which operates a bit cooler. Youre seeing 80 C with factory clocks I presume, imagine how many owners have overclocked these cards and how many again have not taken any precautionary measures as I outline above to address increased heat. Sustained temperatures of 90 C wouldn't be out of the ordinary and yes playing say Planetside 2 for 8 hours a day or more at these temps would certainly curtail the life-span of any GPU. Seeing 80-85 C during occasional benchmark tests, even if in cumulative approached an hour every other day is nothing compared to exploring Skyrim for upwards of 8 hours or more a day with the same temperatures, day in day out. But if you have a 4 year warranty and are not overclocked then you really have nothing to worry about and after 4 years 7970 will be obsolete and due for replacement anyhow.

Yeah Sennheiser is good, but I actually returned one of their newer models, the 363 D, as it didn't emulate Dolby 5.1 very well and had poor bass reproduction and this is in comparison to Plantronics Gamecom 780, a $50 headset. I now have Astro A40's which I consider to be probably the best I've owned in terms of audio, comfort and directional sound. I've tried and owned the aforementioned and also Logitech G35 (crap design, designed to break at the pivot point above ear-piece) and Steel Series Siberia Elite which did not gel well with my head-size which is quite large, really warmed up my ears and had distorted bass and sub-par directional audio oh and my ears felt suffocated. Comfort wise the Sennheiser 363 D felt great but I am disappointed that Sennheiser seems to be following the same design/engineering obsolescence trend (under-engineering products so that they break and customer foots the bill for replacement, i.e. the Logitech G35 mentioned above) and have made the 363 D entirely out of cheap flimsy feeling plastic, even the stress points. This is not quality one would expect for $260.

Anyhow, back to the original subject, I find that the typing comfort has increased a great deal by elevating the back of the laptop as now the keyboard and touch-pad space are at a much more comfortable and accommodating angle and not what feels like a keyboard lying flat on top of a thick phone-book with a 90 degree angle bend my wrists have to deal with.

And your fans will definitely be activated less as you will see cooler temperatures doing day to day tasks such as web-browsing etc.

As far as fan noise is concerned, while gaming with the default fan algorithm the CPU fans tend to kick on and stay on around 3500 RPM regardless of what I am playing with the primary GPU fan exhibiting the same behavior 50% of the time, so manually just setting them all to 3700 RPM isn't necessarily a lot louder but does keep the temps down as during moments of low load the components are cooled to lower than they would otherwise. Its actually very tolerable as you get used to the dull hum, whereas with the default fan algorithm when the fans kick on it is very noticeable. 4200 RPM now, that is a bit louder, fortunately the Astro A40's have a bit of sound isolation and even then it is tolerable.

I'm curious as to what kind of results you get with the Liquid Ultra with your GPU's and your laptop just lying flat on a resting surface with the factory fan algorithm. Please post back with your findings.

 

when did 80C became normal? my 6990m average mid 70s and peak at 81 which lasted maybe 2 seconds. my 7970m is even lower and wont get close unless I OC. I do have manual fan control though.

 

my CPU temps dropped between 20-30 C down from what it was versus ICD or PhobyaHegreaseExtreme before. I dont stress my CPU too much on everyday use hence its on stock clocks all the time, and my CPU idles between 29-35c most of the time while light-load and most I see is a very short peaks just at 50-55c, thats all... so I am ULTRA-PLEASED with the results of the temp-drop I tried runing it on 43xall cores for about a week, and it would never peak past 70-73c, and I must say I was impressed, but I do need battery life while on the move, hence got to stick to stock clocks and VID settings etc.

In general, Im very happy I did CLU on my CPU (GPU's still sitting on ~6+ months ICD with bare 24-48h or load time, and yet I was struggling to get much far than +80c after super-heavy GPU OC, usually GPU's peaked ~ 74-75c on constant OC'ed benching)

just my 02 cent

I have thought of a good option for creating easy-to-remove protective dam for the CPU-GPU die's - if yous apply some cheap (non-conductive & non-silicone) based TIM around the DIE, but away from the DIE just at the right distance to to seal heatsink aluminum part away from the inside impact area chamber, and to cover the exposed GPU/CPU components around the DIE itself, such application removal would be very easy to take apart and super easy to clean and re-paste if ever needed, plus its least stressful and least DIY required. All up to you's

Protect your hardware so you dont have to fork out big dollar replacing it, its not core-2 or pentium-celleron you's are playing with here

 

I have nothing much to contribute about temps, since I only play Dota 2 that runs on the primary GPU, and usually temps hover around 60C+ to no more than 75C. During 3Dmark11 runs on the hottest of days, I have seen primary GPU peaked at 88C, but only momentarily. I don't play much 3D intensive games now, but I think the last time I test run Blacklist, the max GPU was about 75 or so.

Personally, I'll want to keep the GPU temps below 85C. Anything above will perhaps shorten the lifespan of the GPU, not so much of the die itself, but it's usually the VRM and the capacitors that crap out first. A long time ago, I had the Sapphire HD 6850 in xfire on my desktop, and I oc'ed the core from 600+ MHz to over 900MHz. After about 2 yrs using it, one of the capacitor on the VRM just burst. I think it's due to low quality electrolytic caps, as I saw smoke and liquid bursting out of the cap back then.

And onto the headphone, his HD 558 is another class of product compared to the PC 363D. HD 558 is near high end, it's made for audiophiles. It's blasphemous to talk about 7.1 channel sound in headphones for the audiophiles. You only have L and R channel, to simulate more than that you use binaural techniques, for eg recording using a binaural dummy head. None of the high end headphones built for listening music have mics in them, that's another point. I have had quite a few yrs of experience meddling with headphones/earphones now, so one look is all it takes to determine whether a product is made for music listening or gaming.

There are benchmarks for good headphones/earphones, which most people don't know about. There're impulse response, third harmonic distortion measurement, SNR, driver matching, waterfall plot, square waves, resonance, etc etc, those are just like computer benchmark scores. Matching the drivers is just like binning CPU/GPU. You measure drivers, measure a crap ton more, and find a pair that match each other the most.

Personally, the ideal Sennheiser headphones with great price/performance ratio, would be the HD 600/650. Messing with headphones/earphones or audio in general, is like owning Alienwares. They are expensive hobbies that cost a lot. The difference now is, the really good headphones are really durable and very well designed. The latest AW, not so much. On top of that, expensive headphones generally keep their prices pretty well. You can sell second hand stuff at very good price, often close to original buying price..

If I didn't get the AW 18, I could have saved up and got this headphones instead, now that the price is only 3.4k instead of being close to 5k. I auditioned it once almost 2 yrs ago, and I could never forget how good it sounded.
PriceJapan.com

Even if I had gotten the 18, I should have stuck with the stock 4700MQ instead of upgrading it myself. That 1k+ money there should have been spent on a Haswell MBP. For me, the performance difference between a non MX/XM CPU and MX/XM CPU is negligible, unless you're an engineer or someone that runs a lot of simulation which require a lot of horsepower. If one really needs that much horsepower, then the person should just get a X79 desktop with LGA 2011 CPUs, and put it under watercooling.

I'm all for performance, but when the company doesn't endorse it, and make it painful to get the performance, I'd choose to seek alternatives.

 

Per the following response I was under the impression that 80 C on primary under 100% load is a regular occurrence although I could be mistaken. Worrisome aspect is that Corsair has no idea what kind of temps secondary is seeing which could actually be higher than 80 C at 100% load. Thing is, with many newer games I regularly see 100% load so if that is the case then OP may be seeing 80 C or more on at least one of the GPU's regularly if not both.

Again, although you can just sit your laptop flat on a table and rely on the default fan algorithm the many instances of 7970 failure would give me pause to consider taking precautionary measures against sustained temperatures above 80 C. Sure, you can see those temps in a benchmark here and there, but if youre seeing those temps for hours a day, day in day out, that is going to shorten the life-span of your hardware. And lets not forget that this heat isn't confined to your CPU and GPU's, no it is emanating throughout the cramped confines of your chassis, literally cooking your motherboard and other components. My system is two years old and I have been gaming at least 8 hours a day with the same CPU and GPU OC used to generate the benchmark scores in my signature: Skyrim + ENB, Grand Theft Auto 4 with ENB, Starcraft 2, Planetside 2, Natural Selection 2, Crysis 3, Metro LL, Far Cry 3 etc. No signs of deterioration or premature hardware failure as I have kept heat in check. I don't trust the infinite wisdom of Dell. There have been a myriad of 7970 failures. Were they heat induced premature failure? I think it is a possibility.

Although AMD claims their latest GPU hardware can withstand these temps that have hitherto been deemed unsafe what exactly makes their hardware exempt from the known laws of physics, particularly heat = entropy? Are they made out of heat resistant material, gold maybe? No, AMD is committing folly on a grand scale in an attempt to remain competitive with Nvidia, no computer component, unless being manufactured in some new unseen way with exotic material is going to live very long seeing sustained temperatures of 80-90 C.

One sure-tell sign of an increase in premature hardware failure would be the attempt by the manufacturer to distribute those losses onto the consumers shoulders in the form of an increase in hardware price, which we see with the pricing of the AW 18 for example ($5k? are you out of your mind?) and the recent dramatic hike in extended warranty prices.

 

I do know a bit about how these chips have actually produced, and as has been said in here, 80C is not in itself enough to cause a significantly increased risk of GPU failure. While it might wear it out faster, the lifetime of the GPU itself is far, far longer than the rest of the components on the chip. It is far more likely that a capacitor will blow before something on the GPU die does.

Well, I can't speak for the 363 D's, but my 558s have been nothing short of incredible. They actually emulate surround sound quite well (using the surround sound simulation in the R2's SoundBlaster/THX software, it does sound more spaced out and I prefer it). That said, I can't speak for how that compares to a true surround sound setup, because I've never actually properly listened to a surround sound setup. The audio quality is superb, incredibly clear, no distortion up to volume levels far higher than my ears can stand, and the bass reproduction is perfect. It's there, it's punchy, but it doesn't overpower the rest of the frequency range unlike certain other expensive headphones... *cough* Beats *cough*. As for build quality...I've lost track of the number of times I've dropped these, and they show absolutely no sign of damage, they are rock solid. Incredibly light and comfortable too, I often forget I have them on. As for the G35s I used to have...well, I never had a problem with them, but after giving them to my younger brother once I upgraded to these, it didn't take him long to break one of the cans off. I also have a friend that did the same to his.

I might give elevating my laptop a go if I can find something suitable to elevate it with. I'll post back with my results with CLU, in any case. (Well, hopefully it will be the results, and not a warning that I managed to blow a graphics card by getting a bit of it on a capacitor...)

I'm expecting greatness from it, considering how many people have had great results from theirs. Hopefully I don't end up disappointed.

My GPUs do seem to be running warm. They're currently on a fairly fresh ICD application, and on stock clocks. They hit 80C running Furmark, but don't run that hot when gaming. That said, I haven't been playing anything particularly intensive like Skyrim...

Protecting the die like that...that's a good idea, actually. Hell, gives me SOMETHING to do with the crappy MX-4 I have lying around from my first repaste attempt. Non-conductive, easy to clean...perfect for the job, I'd say. Particularly good idea on the GPUs with their exposed components.

You are quite correct in saying that it's a different class of headphones. With that said, I'm no audiophile. Not by any traditional definition, anyway...I don't play lossless audio, I don't have a separate DAC and amplifier, and I've jacked all of the 'sound enhancing' settings in the R2's sound control panel to their highest. That's just the way I prefer to listen to my music. I might not be obsessed with the purest sound, but I do like great sound reproduction, and these 558s do that like nothing else I've listened to so far. I'd actually argue that they're the best price-performance headphones on the market, they cost me £115, the same as my G35s did several years ago, and the same as some comparatively-crappy Bose headphones that I also bought several years ago. My 558s blow both of those clean out of the water. Getting better sound requires significantly more investment, though, with anything from the 600 series costing more than twice as much, which is why I think these are the sweet spot.

As for the mic, I've found that a separate £10 mic works well enough, although the Antlion ModMic looks like a great idea (mic boom that attaches to headphones, I plan on getting one when they come out with the 4.0 version.)

 

Well...honestly, I'm disappointed.

Repasted with CLU earlier. Thin layer of the paste brushed directly onto the CPU and GPU dies (I chose not to put any on the heatsinks), with a ring of MX-4 around the dies (a few mm away) to protect components from any liquid metal that might spill out.

The CPU performance has improved, but not as much as I expected from everyone else's results. Under 100% load, the temperature very slowly rises up until it hits about 87C (takes about 5 minutes), at which point the fan kicks into its highest state and brings it down to about 80C before kicking down again, so there isn't really a steady state temperature. This is still an improvement over what I had before, but still not the 70s people have been getting with CPUs clocked far higher than mine. (This is with no elevated back and automatic fan control.)

The GPUs, though...GPU cooling performance has gotten worse. In a 15 minute Furmark test, it rises very slowly, but never stops rising. It hits 86-87C at the end of the 15 minute test, which is...concerning, and marginally worse than I had with MX-4 or ICD.

I might open it back up again and have a look at the paste, perhaps I'll put some on the heatsinks as well to make sure it makes proper contact.

 

do you have 3pipe cpu hs? applying to gpu die may be more tricky than you thought. all due to imprint onto the gpu die

 

Yep, this is on the triple pipe.

The paste definitely covered the entire GPU die. Perhaps there wasn't enough and it's not making proper contact, but I didn't want to use too much. It definitely covered the lettering on the die, though.

EDIT: Yeah, appears there wasn't enough on the GPUs. On opening it back up again, some of the spots had almost no paste on at all, and barely any of it had stuck to the heatsink (compared to the CPU, where it seemed that pretty much half of it had transferred to the heatsink).

After putting a dab more of it onto the GPUs, as well as applying it to the GPU heatsinks too, thermals are looking better. A 15-minute Furmark run results in a slower climb this time, peaking at 80-81C on the primary GPU. (The secondary GPU's thermal sensor still refuses to co-operate with any software, it seems)

Going to test the CPU now. I didn't put any extra on the CPU but I spread it around a bit more, particularly on the heatsink, to try and get it more even. I'll see if that has any effect.

EDIT 2: Well, seems that I made the CPU slightly worse if anything. It now climbs faster to the 80s (although it might have started hotter in the first place), but the activity in the 80s seems to be the same as before, hitting around 88C, fan kicking into highest speed, then settling in the lower 80s with the fan sometimes kicking back down again for a bit.

I'm starting to wonder if there is something up with the heatsink. Perhaps one of the three heatpipes have leaked.

Oh well. Still better thermals than IC Diamond, and hopefully I won't need to repaste this again. Can't say that I wasn't hoping for more, but I can't complain. Still wondering how people are managing 70s on their overclocked XM CPUs, though.

 

Be sure to paint every part of the die, edge to edge, leaving no uncoated areas on the die. Then do the heat sink copper plates where they mate with the die. No need to extend beyond the area of the die on the heat sinks, but you want to be sure that you are mating paste against paste with a super thin coating on both surfaces. I tried doing just one surface and I was not happy with the results.

 

A useful analogy to consider: take a conventional gasoline engine automobile with an air cooled liquid heat-exchanger (radiator). Say the radiator is positioned in the trunk, not the front of the car that has access to ample fresh air. Replacing the radiator with a larger radiator will net little to no decrease in engine temperature as the heat-exchanger, irrespective of the size, is incapable of dissipating that heat as it is in the trunk. Even if you were to double the size of the radiator surface area, you would see no net improvement in reduced engine temperature. The same principle applies to the heat-exchanger in your Ferrari of a laptop. Given the inherent design limitations of a laptop chassis, and that unlike the comparative warehouse-like interior of a desktop PC, a laptop heat-exchanger needs to literally move the air-space of the room over its heat-exchanger as it isn't directly in contact with the air-space of the room. This is THE point of elevating the area where the air-intake fans reside. Youre essentially opening the trunk that the automobile radiator is sitting in.

I have seen a reduction of 13 C on an M17x R2 with factory TIM at 3.5 Ghz simply by allowing ample fresh-air to flow over the heat-exchanger. At 4.3 or more Ghz, with a superior TIM, this figure becomes exponential and you can see a reduction of 30 C but you must allow the heat-exchanger ample fresh air so that it can convey that heat out of the system. Think about it, yes your liquid gallium TIM is conveying the heat from the CPU to the heat-sink but the heat-sink is incapable of dissipating that heat.

It's either this or you may need to reapply the TIM as there is a remote chance that there is an air-pocket or maybe one of the screws is not fully tightened or you didn't follow the pattern outlined on the screws or when seating the heatsink onto the CPU die there was some lateral movement that upset the TIM or maybe you neglected to apply a very thin layer to both the heat-sink and the CPU die or maybe the heat-sink wasn't cleaned adequately with the included scouring pad and there is still some remaining IC Diamond on there, even if it is unseen by the human eye.

Doing it right, which includes providing ample amounts of fresh air, including manually controlling the fans should yield a reduction in peak CPU temps of at least 20 C, and possibly as much as 25-30 C.

See also:

http://forum.notebookreview.com/ali...ltimate-cooling-mod-end-all-cooling-mods.html

 

My 18 is propped on a Zalman cooler. Without it the temperature goes up by about 10C higher, at most. With it the temperature is lower. The CPU fan that's positioned at the middle of the machine usually gets less air compared to the GPU fans at the side.

 

Yes raising the machine lowers the resistance for the fans drawing in air, especially towards the centre of the machine.

 

As I said in my previous post, I did go back and apply some CLU to the heatsinks too, and now temperatures are fine, GPUs are down to 80-82C max after a very slow climb.

Not as low as I'd hoped, but again, I can't complain. It's not overheating, and if I really want to make it quieter, I can just prop the back up.

For the most part, my biggest hope is that I never have to re-paste it again, seeming as it is just liquid metal.

 

For those that are still apprehensive, Liquid Ultra application starts at 30:49 in this video.

 

Great video buddy. How's your temps now?

 

"Over-clock-a-doodle-doo" lol

 

Temps are fantastic.

 

Glad to hear it. I've been waiting on some new pads for a while now since after applying this I'm unlikely to ever need to remove the heat sink again.I can't wait to do it.

I noticed in your video you just went for it, as in, no screwing around with any pads or silicon to form a dam for the unlikely event of overflow.

 

No pads here either.