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

Hi @ll!

I will post benchs and pics before and after the repasting time.
Do any of you have any advice before I do the job?
This thermal compound is not usable with aluminum heatsink and as far as I know 3 pipe-heatsink is copper made right? So, no trouble at all with that according to technical sheet.

Thermal compound:


Thank you to all of you and I wish you happy new year!

 

Let us know how it goes. Personally, I don't trust that stuff. IC Diamond for life. :thumbsup:

 

Just put some on the die and spread it with the brush. At first they don't really spread and stick to the die, but after a while it will.

It will stain everything, so be careful not go get it on other surfaces. It stains the copper too, and after applying it if you want to use another TIM you will have to lap the copper.

Bare-die testing: A delidded 3770k, an H100, and 9 different TIMs - Page 3 - AnandTech Forums

It's liquid metallic gallium alloy, and it's conductive, so you don't want it touching any pins. Be careful not to get it onto the mainboard too.

 

You're a brave soul, Nikos. Good luck... I look forward to seeing how it turns out. Based on the reported results, I have been wanting to try this for a good while. I've heard too many horror stories to get excited about it and I don't want to have to deal with a mess getting it off if I am unhappy with the results.

In addition to the warnings from kh90123, be aware that it eats up aluminum. Avoid getting it on any part of the heat sink that is made of aluminum.

Here are my test results posted later in this thread: http://forum.notebookreview.com/thr...ips-before-i-start.741745/page-5#post-9520969

 

I like this video

How Thermal Compound Spreads - YouTube

John.

 

Well it's actually not that scary, I did get a few tiny drops here and there, but I wiped them off with tissue.

Imagine spreading melted solder onto the die. That's how it feels.

My ICD7 still not here. I am itching to repaste the GPU.

 

I look forward to seeing your results Nikos. I'm actually repasting right now with IC7. Half way smoko/coffee break before I put back together.

 

I actually read someone melted metal on to die to save some bucks versus buying those liquid metal kit, and it work just fine. He say it is cooler than using paste.

 

Reposted both GPU'S and CPU. Easy to do. Works like a charm. Run my GPU's at 900/1300 I hit 68oC on Valley Benchmark. My CPU runs 4.4Ghz. Running the ITU Stress Test, I hit mid 80's.

 

Have you experienced what it is like to clean up and repaste yet? Applying it should be a piece of cake. Clean-up issues is the part that is holding me back. I don't want to have to sand that crap off and it makes me worried about how well that will turn out in time.

 

Cleaning it from the die is no biggie. Just wipe it off. But there will be tiny droplets left behind, if you don't wipe it clean. Be careful with it.

Cleaning it from the copper is another issue. It kinda seeps into the copper. I assume that if someone is using the same TIM again then it's no big deal. Kinda neat isn't it, you get ppl to use it once, then they have to keep using it.

 

It comes with a little scrubby pad. I did use a piece of fine grit sand paper to clean the heat sink. Turned out fine.

 

If it seeps into the copper that might actually be a good thing if it doesn't cause issues if you decide to switch to another paste later on. It seems like that might potentially enhance thermal conduction by filling pores and increasing density of the heat sink.

 

Curious on the improvement this paste can provide

Sent from my SM-N900T using Tapatalk

 

Hi again mates, I've finished doing the repasting task. Still I have to do more test but the results are incredible compared to the stock thermal paste.

Here you have the results during the execution before applying new thermal paste (stock paste, 1hour XTU stress test):

As soon as I finish posting here, I'll start the same test with the same settings:

The only problem I can feel it's that one core is a little bit cooler than the others (8º or so) so I guess that I will have to tighten the heatsink again to achieve an equal temperature in all cores.

PS: So happy more room for oc'ing

 

Don't worry about the one cooler core. I consider what you are seeing to be normal. That is very common and you're not going to change it. It's cooler, so there is no reason to be concerned about it. You're getting good results so just enjoy the nice temps.

 

Here I go:
During the test.


After:


What do you think?

 

That looks pretty nice for 4.3GHz. Good job.

What CPU do you have? Is this an M18xR2 or an 18? I run my 3920XM and my 4930MX at 4.3GHz 24/7 right now, so I am more interested in the temps at 4.5GHz and higher. Can you run the same test at 4.5GHz and see how the temps look?

 

Those are some very nice results, i'm interested in seeing 4.5ghz performance too.

 

Sure! I don't have the stable config for 4.5Ghz if you help me with approximate values (like yours, or maybe an Intel XTU profile) It will save me a lot of time because I lost my 4.5 profile.
PS: 3920XM

 

Hmm, I'm surprised. I didn't think this stuff would work that well. Good posts! Hope to see more.

 

4.5Ghz - ThrottleStop 1024M's bench!

Current BIOS config:



I think it could be done better with more config because I just set high and high to get this bench pass. Too much voltage to what is gained don't you think so?

Also I repasted GPU's so I ran 3DMark Vantage to check everything is ok




A little bit higher:


PS: Awesome temps for this gpu's, this is by far the best thermal paste I've ever tried.

 

For now, anyway. We'll see how long it lasts.

 

This is messed up. I just saw your post.

Try setting 45x4, Flex VID 50, Pri Plane has to be at least 896, but 960 might be better for 45x4... see this screen. I like running with c-states disabled so my CPU clock speeds never decrease or fluctuate regardless of load, but that's optional. Be sure to disable all three options on this screen.

Here is my favorite XTU profile export for 4.6GHz if you want to try it. View attachment 3920XM-46x.zip

These results are pretty darned impressive, but "how long" is also a good question. I've used some other pastes that work similar to IC Diamond at first, but they are generally only good for a few days, then degrade severely and become pretty much worthless compared to IC Diamond. If this Liquid Ultra stuff will work the same in 6 months as it does now, that would be extremely impressive. IC Diamond generally lasts me a maximum of roughly 90 days with a mixture of 24/7 4.3GHz and extreme overclocked benching 4.6-4.8GHz

 

Today I opened up my 18 and put some IC Diamond 24 in there. More carat more better. Yeah.

Well that's just a joke. ICD 24 is the same as ICD 7, just a much bigger tube.

I noticed that compared to this post: http://forum.notebookreview.com/ali...tx-780m-installation-m17x-r4.html#post9458298, at stock some of the tiny chips at top and near the top right don't have pads on them. I proceed to put on some pads on them, and put on the ICD 7/24. I also swapped the 2nd GPU to be the main GPU again, since I play MOBA like Dota 2, and that only uses 1 GPU. What are those, VRM? Mosfet?

I heard coil whine some time before (with the palmrest removed), so I thought swapping the GPU would help. But in fact both GPU makes some coil whine sound when put under heavy load.

As for temps, it's at least 2C lower than the Shin Etsu. The ICD paste is perhaps the thickest paste that I have ever used. I guess one needs to apply the bead of paste by holding the syringe sideways, if you do it straight up it there's always a strand of paste as you pull the syringe away. That might form air bubble.

I actually repasted the main GPU (GPU 0) because I noticed that the temps are higher on that one, about 3-4C higher after the swap. I saw that the ICD paste spread very evenly, without air bubble. Very good paste indeed. Turns out one of the GPU runs a bit hotter. In GPU-Z the ASIC quality is 76% for GPU 1 and 71% for GPU 0. Perhaps that explains the higher temps.

 

^^ how do you check for ASIC quality? (thanks)

 

Be careful about putting too much stock in this. It has not proven to be all that meaningful or definitive in the grand scheme of things. My 680M and 780M cards have all shown low ASIC ratings, but they have been overclocking banshees. If I remember correctly, the same has been true for Johnksss and others that like to overclock. Just offering this an FYI... no need to make a big to do about nothing. Seems more like hype and a reason to be GPU hysterical than anything else to me.

That said, you do it like this... start with the little green GPU icon in the title bar.





My 680M cards were even lower, like 69% or so.

 

Hi!
Finally I've got it fully stable @ 4.5Ghz (As far as I know) thanks Mr.Fox (as always) and thanks to all that came here to exchange their opinion.
Now I have some extra questions:
I think I did something wrong with C-States, I turned them off just for trying but I want them on again(less heat and less consumption), but now I always have the 4 cores running at 4.5Ghz. Which options do I have to turn on again? I think I missing something.
Also I just realized that I have the chance to upgrade to Win 8.1 so I don't want to change it if it not worth or If it perform worse. Any Issues with current drivers?
This is my current Intel XTU config @ 4.5:

 

It's hard to do anything wrong with this. You only need to change one setting. Go to this BIOS screen and change C-States to "Enabled" and they will be working again. Or, you can add a check mark to the C1E box in the ThrottleStop main window and this will let you toggle C-States off and on at will. Add check mark=Enabled. Remove check mark=Disabled.

What were your max temps at 4.5GHz? Did it have any throttling?

 

Max temps were not higher than +-90 ºC
No throttling at all
Can't make stable your XTU config @ 4.6

Sorry, I've edited my previous answer don't know if you has seen the part of upgrading to windows 8.1 Do you know something about that?

Thank you ^^!

 

well, to my surprise my GPU's read 75.8% and 78.0% ASIC Quality, and yet I had been able to push them pretty high on OC, somewhere past 1.2GHz last time if remember right, however, its been a while and a lack of time...

this Liquid Ultra looks promising to me so far I will keep an eye on this thread for a while now to see the long run results and repaste experience

 

Really, ~ 90 ºC at 4.5GHz? That sounds pretty amazing.

Windows 8.1? Sure, it is pretty straightforward and there isn't too much involved in that. If you are already running Windows 8, then yes I would go to 8.1 (Windows 8 SP1) as it has been improved upon. If you are running Windows 7 and have not used Windows 8/8.1 already, you might want to create a drive image of your current installation so you can easily go back to Windows 7 if you don't like it. Or, you can set up Windows 8.1 in dual boot with Windows 7. I have that setup on the M18xR2 and the Alienware 18 and works really well. I seldom boot into Windows 8.1 any more because I find Windows 7 is a superior OS in most ways.

 

I'm in windows 8 already, don't need to change any drivers or so? I thought that I will have issues with this.
Thank you

 

You can turn your C-states on again in the bios. Now your reaching decent clocks it might be a good time to get familiar with Throttlestop - I can't say enough good things about the program. You can create different power profiles and switch them on the fly using hot keys without needing to go into the Bios or restart (C-states can also be toggled here too).
I believe there is a slight performance disadvantage moving to W8.1 - a waste of time if you ask me.

 

Nah, no problems with drivers. I am still using a fair number of Windows 7 drivers with Windows 8.1 without issue. You may have to boot with Driver Signature Enforcement disabled in some cases if the driver is not specifically designed for Windows 8.1 but that's easy enough. The Windows 8.1 UI is a little less ugly than Windows 8 and Micro$oft cleaned up a few things to make it a little less clunky and inefficient to navigate compared to Windows 8.

 

Why do you think it is a waste of time? I see some benchs with BF4 in which Win 8.1 is better than 7.

HARDOCP - Conclusion - Battlefield 4 Windows 7 vs. 8.1 Performance Review

I will not upgrade till I do more research

 

BF4 seems to have a lot of reports of improved performance but that game is broken mess right now anyway. Besides that it's actually not much more demanding to run than BF3 anyway. It will run fine on your rig.
Mr Fox has done a lot of comparative benchmarking which has shown W7 to still have an edge.
Me saying it's a complete waste of time might be a little obnoxious but all I've noticed people wasting time and money getting it to look and perform like W7 - so I kind of see it as a waste of time. Also screwing around with driver signature enforcement sounds like a PITA.

 

With respect to CPU performance, Windows 7 does still hold a bit of an edge over Windows 8.1, but not as much of an edge as it held over Windows 8. This may not matter to people that are not number chasers. As a matter of personal preference, I like Windows 7 better for aesthetic reasons as well, but this subjective. The choice of light colored pastel color schemes is tacky, loss of Aero transparency made it bland and the ability to customize many aspects of the interface in the desktop environment kind of stinks. I think they may have done that in an effort to drive people to using the Modern UI more.

Whichever OS you like to use more is the superior OS. Neither one offers anything particularly superior to the other in terms of features. None of the Modern apps are particularly interesting to me. Some of the adware-riddled apps me off every time I look at them. I usually end up in a bad mood running Windows 8/8.1 after a few hours because I think the UI is ugly and bloated compared to Windows 7. It doesn't help me do anything better on the desktop and the Modern UI is slows me down enough to become irritating. I've been using it since launch and made it my primary OS for a good 4 to 6 months to "adjust" and it just never won me over. I use Windows 7 about 80% to 90% of the time now.

I cannot say that I literally hate it with a passion because that would be a little bit of an exaggeration... it's not that big of a deal, but it doesn't really offer anything useful to me except for the Win-X menu. I like that specific feature a lot.

 

Just one question before upgrading to 8.1, I'm using unlocked A10 BIOS I don't want to change this BIOS, I saw this morning that Dell's support site released a BIOS A12 for win 8.1
I'm using secure boot and fast boot (also GPT not MBR) Do you know if I will have any problems with that?

Thank you

 

Not likely that you will have any issues with it. I am not aware of a compelling reason to move to A12. You might end up locking your machine into a state of permanent crippledness with A12 like the new Alienware 18. I don't see an unlocked version of A12 from svl7, so I would avoid it like a plague. There might be a good reason there is not one. You might have some legitimate reason for using Secure Boot/Fast Boot, but Secure Flash (if A12 introduces it) will basically make your machine worthless as a beast. It will be a castrated beast with Secure Flash, LOL.

 

Roger that! haha Not even think in touch A12 or above hahaha :

Now back on topic:
New drivers 332.21
3DMark Vantage:
3920XM @ 4.5
GTX680M SLI - svl7 vbios (not OV) - 980/2450 Mhz

 

CPU temps @ 4.5GHz look real good [Nikos]. Nice work. +1

 

How much improvement degrees do you see vs ic diamond ?

Sent from my SM-N900T using Tapatalk

 

Mr.Fox could tell you more about that I didn't use IC yet

 

I guess I will try to bite the bullet and try this one myself so far getting very excited seeing such low temps @ such payload on CPU

especially after I've seen this: http://www.youtube.com/watch?v=Z4v3k47Pdbw

got it, cost was ~16eur incl delivery, but I put my hopes to it that it will be worth to try

http://www.ebay.co.uk/itm/161068917349

 

was also thinking, if I got spare CPU heatsink, lapped it to mirror shine with diamond polishing paste as final finish touch (6000+ grit etc) would this give me any extra gain whlist using this liquid ultra compound... I wonder.... worth spending extra $ on spare HS and the hassle? What would you think?

Also, since that's pure metal compound I will not be just slapping that compound there and expect "nothing bad can happen", I was thinking to apply some RTV silicone (thin track of it) just around the CPU pcb close to the edges of it but not on the edges, so when I seat the CPU heatsink with new compound on to the CPU, the RTV silicone would create protective border sealing the compound in event it could possibly get loose or chip away in the longer run, precaution against metal particles getting loose and falling within the system causing disaster within the system, so called as a mystery failure's and BSOD's etc

what do you's say? better safe than sorry? Watched the removal of this compound youtube video looks like aluminium sort of like dust which believe it or not are very "crispy" when dry and tend to breakaway in event of impact or vibrations etc, especially after bonding compound dries out and just the metal compound remains (after curing)

 

Well, here are some tests we can use as a baseline for comparisons. I ordered a tube of Noctua NT-H1 Thermal Compound and a tube of Coollaboratory Liquid Ultra for testing. I will replicate the same tests using identical overclock settings with the other pastes to see how they work on a fresh repaste compared to an application of IC Diamond that has already been subjected to 30 days of use, included some pretty severe overclocking.

Control System Conditions
Thermal Paste	IC Diamond - applied on December 8, 2013 (30 days of overclocking) - see extreme benchmarks run on this fresh ICD paste LINK.
Tested CPU / Clock Speed	3920XM / 4.3GHz (43x4) - Note: this is my normal everyday clock speed for this CPU, running with c-states disabled 24/7
Flex VID / XTU Add'l Turbo Voltage	10 / 19.53125mV (Flex bumped to 10 for 3DMark11 and Vantage for improved stability)
Pri Plane / XTU Core Current Limit	896 / 112.000A
ThrottleStop Settings	TRL: 43, 43, 43, 43 - Flex 10; TPL: Long: 200, Short: 200, Turbo Time Limit: 0.0010s, Package Current Limit: 112

If anyone else, including [Nikos], would like to jump in and use these same settings and replicate the tests, it would be awesome to see how consistent the results are. We may see different temps, but from the results we might be able to calculate a reasonably accurate percentage of change. Even if we can't be reasonably accurate with calculations, we might be able to get an accurate pulse on which paste is the consistent winner among multiple systems running at the same or similar overclock speeds.

Be sure to run the tests before repasting, just as I am doing today. To keep it interesting, I ran some tests in both Windows 7 and Windows 8.
[parsehtml] [hr][/hr]</HR> [/parsehtml] Current Control Performance: IC Diamond - 30 days of heavy use
[parsehtml]<iframe class='imgur-album' width='100%' height='750' frameborder='0' src="http://imgur.com/a/YmXeo/embed"></iframe>[/parsehtml][parsehtml][hr][/hr]</HR>[/parsehtml] Fresh Noctua NT-H1 Thermal Compound
[parsehtml]<iframe class='imgur-album' width='100%' height='750' frameborder='0' src="http://imgur.com/a/0D8ly/embed"></iframe>[/parsehtml] First Impressions: Noctua NT-H1 is easy to apply. It has a creamier texture than IC Diamond. While initial results are impressive, I have reservations about whether it will be durable based on its viscosity.
Final Verdict on NT-H1: FAIL - IC Diamond is a better (more durable) product... see screen shots for results after only 4 days. See last screen shot of Darksiders II in-game temps on day 5. Removal in preparation for Liquid Ultra was extremely easy. "Pump-out" effect was obvious when heat sink was removed from CPU.
[parsehtml] [hr][/hr]</HR> [/parsehtml] Fresh Coollaboratory Liquid Ultra
[parsehtml]<iframe class='imgur-album' width='100%' height='750' frameborder='0' src="http://imgur.com/a/8wOSO/embed"></iframe>[/parsehtml] First Impressions: Application was much easier that expected. I "painted" a thin layer to both surfaces (CPU die and heat sink) after a thorough cleaning with alcohol. I was amazed at how little TIM was required. About the equivalent of 1/3 to 1/2 grain of rice was enough for both surfaces. Initial results are as impressive as what we have seen with ht_addict and [Nikos]. I look forward to seeing how long it will last.

Edit: 02/26/2014 - 6 weeks into overclocking hell and the Liquid Ultra show no evidence of degradation. Truly amazing... I am moving IC Diamond to second place on my list of best thermal pastes money can buy.
[parsehtml] [hr][/hr]</HR> [/parsehtml] Liquid Ultra Application
[parsehtml]<iframe class='imgur-album' width='100%' height='750' frameborder='0' src="http://imgur.com/a/JUQvK/embed"></iframe>[/parsehtml]

 

Do it... would be nice to find out if lapping helps.

I have a spare 3-pipe and when I get done testing the results with the other two pastes against the 30-day old ICD I may lap that one and see if it changes the results. I'm going to do it in order with Liquid Ultra last, so if the results are really great I might just leave it alone to see how long it holds up. Will be interesting to see at the 30-day mark how it is holding up. I'm eager to bench the crap out of that stuff at 4.6-4.8GHz to see the results.

 

pchub is out of stock with 3pipe hs for the moment and there gone up in price too, but keeping hot-link handy to check frequently when it gets available I will snatch one right away... (shouldn't have sold my spare 3pipe back in a recent day, would have had great use out of it now...well...)

for now will be waiting for delivery on the goodie. Maybe that keeps me running daily @ constant 4.5-4.x-x.x GHz just for fun (if thermal conditions allow, will see, will see)

 

I'm very keen to see the results. I'm starting to wish a didn't reorder IC7 the other day.