Coollaboratory Liquid Ultra vs Liquid Pro

[tw33k]

I've been unable to find a comparison between these two compounds so I thought I'd compare them myself. I've had it confirmed by the CEO of Coollaboratory that Liquid Ultra is rated at 38.4w/mk and Liquid Pro is rated at 32.6w/mk.

System
CPU: Intel 3770K @ 4.7GHz 1.224v (de-lidded)
Mainboard: ASUS Maximus V Formula
Cooler: Corsair H100 (2x Corsair SP120/2x Scyte Ultra Kaze 3000)
Memory: 2x4GB Crucial Ballistix Tracer @ 2000MHz
SSD Crucial M4 128GB
Graphics: Sapphire 7950
Case: Corsair 800D
Operating System: Windows 8 Pro x64

The case is very well cooled. There are a total of 9 case fans all intake. The H100 exhausts out the top.

I have a temp sensor measuring the room temperature. Attached to it is a sensor which is inside the case, next to the mainboard. There are 4 more temperature sensors attached to the fan controller at various points inside the case. The case temperature was recorded at the start, after 15 minutes, after 30 minutes, after 45 minutes and at the end of the testing. The average was then used.

I used the latest version of Aida64 running "Stress FPU" for 1 hour. FPU stressing means AIDA64 System Stability Test will use a floating-point calculation task that stresses the FPU part of your processor. Modern processors all have an integrated FPU, and from all the components that are integrated, the FPU is the most complex one. Hence stressing only the FPU actually stresses most of your processor, and usually drives the processor to its maximum temperature. The average core temperature is recorded by Aida64.

The compound was painted on the die, on both sides of the IHS and on the base of the H100. Each compound was applied, tested then re-applied and tested again. The best results were posted

Liquid Ultra

Liquid Pro

Liquid Ultra has a delta temp of 3.1c lower than Liquid Pro in my system

3770K @ 5GHz 1.416v

Liquid Pro

Liquid Ultra

This time Liquid Pro was 1.71c cooler than Ultra.

If we use a +/- 2c margin of error there is nothing separating these two compounds. Liquid Ultra is better to work with. I wouldn't use the cotton tips that are supplied with Liquid Pro because I found more compound stayed on the cotton than on the IHS. Also there is the risk of cotton fibers sticking to the die or IHS which would impact temps. I used the brush that is supplied with Liquid Ultra which makes the application very easy.

Cleaning is another issue. Liquid Ultra simply wipes off but Liquid Pro often forms tiny balls which when accidentally smeared leaves quite a big mess. I had the stuff all over my fingers by the time I finished wiping it off.

If you are using one or the other than stick with it. There is nothing to gain by switching.

 

[Idontcare]

:thumbsup: very cool test and results tw33k!

Thank you very much for running the comparison and taking the time to fill us in on the results

If we had a rep system here I'd "+1" you

 

[tw33k]

Thanks mate.

 

[FlanK3r]

nice, so I must to buy Ultra (I have PRO )

 

[tw33k]

To be honest, I wouldn't bother. They are so close in performance that I could have done a another round of testing and Pro came out on top. There's got to be room for error and a 3c difference doesn't leave a lot of room.

The reason I did this is because of the misinformation floating around (particularly on OCN) People are continually claiming that Pro performs better, they even throw up figures like they've run tests but never posting test results. I wanted to provide actual results.

 

[Idontcare]

To be honest, I wouldn't bother. They are so close in performance that I could have done a another round of testing and Pro came out on top. There's got to be room for error and a 3c difference doesn't leave a lot of room.

The reason I did this is because of the misinformation floating around (particularly on OCN) People are continually claiming that Pro performs better, they even throw up figures like they've run tests but never posting test results. I wanted to provide actual results.

I totally know that feeling, which is why I personally very much appreciate you taking the time to generate, and post, the results :thumbsup:

 

[tw33k]

5GHz results posted

 

[Pedroc1999]

Tw33k, Could you answer my PM please, you even have the same overclock

 

[tw33k]

Yes...I am on SF as well. It won't let me send PMs until I have 25 posts.

 

[Pedroc1999]

Yes sorry about that - the name reminded me but then the 5ghz overclock... That just confirmed it, i remembeed it from page 1 shoq us your rig

 

[Belial88]

Just wanted to say thanks (long-time member of OCN). A ton of people are saying in the Official Ivy Bridge Delid club that Pro is better.... because pro has been around longer, and no one gets both of these pastes (you spent $50 on both... wow, thanks for testing) so naturally people just recommend what they have, and given that pro has been around longer...

One source says pro has like 80w/mk and ultra is like 20s or something, but they are 2 different sources saying w/mk figures, and i dont believe w/mk is everything when it comes to thermal transfer (could be wrong on that, but i believe it's more complex than that, i mean just look at how many pastes with low wm/ks are ranked highly and high w/mks ranked low).

I mean it's only logical that Ultra (which is supposed to be Pro v2.0) would be better or equal (then again it wouldn't be the first time that an update was worse than it's predecessor). Obviously you can't blindly compare pastes, but for example I tested pk1/pk2/pk3 and pk1=pk2<pk3. with pk-2 spreading easier than pk1 and then pk3 just being better than both (but spreads the hardest, although it's still a ceramique so there's no need to spread anyways).

On a side note, do you HAVE to 'paint' these pastes? You can't just put a blob down and let the heatsink push it down?

I did a thermal paste review myself, I'm a bit surprised you got so much variance in your tests. I tested pK1/PK2/PK3 in the same session, air conditioned, temp controlled room (case with ton of ventilation too, 10 fans and side panel open kind of thing), and even after multiple remounts I didn't get any change in temps basically. The only thing that happened, was the chip ran about 5*C hotter when I used the spread method vs rice grain. I tried multiple amounts of rice grains too, and generally, as long as you don't use too much, it's all the same (it's very easy to use too much, a single fat grain or 2 small grains would cause 5*C+ than a reseat right after or before of the same paste in the same session, and the ideal amount of paste was always so little that the paste would only spread out to cover about half the IHS, sometimes 1/3rd, as long as you cover the die you are good).

http://www.overclock.net/t/1346069/...ews-hyper-212-h50-nh-d14-pk-1-pk-2-pk-3/0_100

edit: hey i didnt realize you posted this on OCN first. funny. just reply in the ocn thread lol.

 

[Ajay]

Awesome, thanks for taking the time to test and post results :thumbsup:
I'll probably try Liquid Ultra next time I do a build or cooling upgrade, seems to be much easier to work with than IX

 

[notty22]

I watched the firms video, on applying Ultra. It seems to look and almost spread like Mercury,(that was the first thing in my head), even though I have never played with mercury.

 

[tw33k]

Awesome, thanks for taking the time to test and post results :thumbsup:
I'll probably try Liquid Ultra next time I do a build or cooling upgrade, seems to be much easier to work with than IX

Even using it on a non de-lidded chip gives excellent results. I put it on my AMD 8150 @ 4.5GHz and got a delta temp of only 9.8c which is a fantastic result.

Just wanted to say thanks (long-time member of OCN). A ton of people are saying in the Official Ivy Bridge Delid club that Pro is better.... because pro has been around longer, and no one gets both of these pastes (you spent $50 on both... wow, thanks for testing) so naturally people just recommend what they have, and given that pro has been around longer...

Exactly. So many people over there claim to have used both but not a single one of them has posted any results

 

[Ajay]

Even using it on a non de-lidded chip gives excellent results. I put it on my AMD 8150 @ 4.5GHz and got a delta temp of only 9.8c which is a fantastic result.

I'm hoping my next CPU upgrade (IVB-E or Haswell) will not need de-lidding as with IVB.

 

[tw33k]

Finally got a response from Coollabs...

The heat conductivity value for Liquid Pro is 32,6 w/mk and for Liquid Ultra is 38,4 w/mk.

I knew the figures that people kept quoting were not correct and now we have confirmation (from the CEO no less)

 

[HexiumVII]

Are these pasts worth it if I already have arctic 5s? (will be using with Closed Water Loops)

 

[tw33k]

Both of these compounds out perform AS5 by a few degrees. They perform best when painted on the die of a de-lidded chip but you would still see an improvement using it in the traditional manner

 

[Idontcare]

Finally got a response from Coollabs...

I knew the figures that people kept quoting were not correct and now we have confirmation (from the CEO no less)

I'm very glad to see you drove that to a conclusion as well :thumbsup:

 

[Belial88]

as5 is terrible stuff, it's over 13 years old now (look at how much gpus and cpu's advanced in that time).

This stuff is probably at least 10*C better on IHS. Stock TIMs are nowadays getting better than AS5. AS5 is a terrible paste, any modern paste is significantly better, especially when you consider that 1-2*C better performance is considered a huge leap in performance, and that AS5 is regularly 1-3*C behind PK-1, which is itself outdated and is about 5*C worse than modern tims like pk-3 and masscool and nh-t1.

 

[Torn Mind]

Are these pasts worth it if I already have arctic 5s? (will be using with Closed Water Loops)

CoolLaboratory is boss. AS5 is overpriced and a waste of silver, imo, especially when Ceramique 2 is only barely worse yet half-price.

 

[tw33k]

I'm very glad to see you drove that to a conclusion as well :thumbsup:

Me too. It was so frustrating challenging people who had read that Pro was rated at 82w/mk and simply believed it. My thinking was, what company releases an inferior product and then goes and calls it "Ultra" It just didn't sit right. I must have filled out the "Contact Us" form on the Coollabs website a half-dozen times but never got a response. After some searching, I found an email address and it just so happened that the CEO read those emails and he responded after a couple days.

 

[nikosa43]

Congrats TW33K. Fantastic job and feedback for everyone that wants to find reference to explore his system. I also have a Formula V and a delided 3770k, that I have pushed to 5G with a Noctua NH-D14 and Ultra Coolaboratory. Posts like yours and of course IDONTCARE's (which I am a great fan), gave me the inspiration but also the base to start my experiments and get the satisfaction of fine tuning my system. Once again congratulations to both of you.

 

[tw33k]

I'm glad this was of some use to you.

I wanted to give a brief update. I thought I had applied too much Liquid Ultra for the 5GHz run so I pulled everything apart and used the smallest amount I could but still covering everything properly. I ran Aida64 for 1 hour but saw no improvement in temps. In fact, they were 1c higher than the previous run leaving me to wonder "how much is too much?"

I have recently received 3x Ultra and 3x Pro and will be testing again with various amounts to try to answer this.

 

[nikosa43]

In my setup I also used the smallest possible quantity, under and over heatspreader, but I left it like this with the first application. I was looking to the Vcore you used for 5 G and I was wondering something because I have the same mainboard. Which BIOS you have? I found a stable clock with 1406 and the Vcore I used was 1.425 with Extreme LLC for not having big Vdroop. Now with the update to 1604 I had to use 1.445 to stabilize it and not giving also errors in event viewer. In the first case I also had lower temperatures.(highest core spiked to 81) and now highest spiking to 87 which is normal with the increased Vcore. So please check this if it is possible because I was thinking to go back to previous BIOS revision. I also wanted to ask you another thing. Your Vcore was 1.416 (without any variation, from the app you used). I was used CPU-Z and during benching I had highs as 1.485. I think this comes from the Extreme LLC (overshooting), which I don't like. Is it CPU-Z false monitoring or BIOS settings that cause this variation. Thank you in advance.