Corsair H50 CPU Cooler Preview.

The Corsair H50 looks like a Swiftech H20-120 Compact but with LGA 1366 compatibility.
 
I would like to see a reverse fan review on this IMO to make it "fair". if thats a possibility for the next review to see it both ways?

I think that the radiator is just too small for an i7 especially getting closer to 4ghz they put off a ton of heat that a 120x1 just cant relieve.

I do admit that in my watercooled setup BOTH my 120.3 and 120.2 exhaust inside my case, but I have AMAZING airflow in the case and its only 5c above ambient at full load on the GPUs and CPU (24/7 folding@home)

Kyle have you thought of doing a poll on peoples ambient temps? I would like to know an average because I WOULD NOT play games at 80F, my comp room stays at 68F summer 70F winter which is a 10F difference from what you tested.

Thanks for the great review, loved the animations at the beginning!

On a side note, i would use this in a system build for people over the TRUE just because of the size constraints (if it performed as it did with a reversed fan)
 
So would this be considered an entry level watercooling solution? Because honestly, i can't see this cutting a name for itself as enthusiast grade, whether it holds its own against a TRUE or not. Obviously it performs, but I honestly cannot say this is enthusiast grade. To me, enthusiast grade is a bit more complex in setup than what the H50 is, though the basics are there. I'm sure plenty of people who do watercooling would love it to be as simple as what the H50 provides, but obviously some feel that better can be had by doing it on their own, and they're probably right. I don't think this should be aimed at the enthusiast, but at the person who is just getting into watercooling, or have very specific needs to which the H50 would be ideal.

I do have to give props to Corsair for simplifying the process, as it would make it easier on folks like me who want to get into watercooling without ruining anything, but I really really do not think it is at the level of enthusiast grade. I think what the H50 does is great for people who have cases that do not provide great airflow, those that are paranoid about warping their mobo over time by having a bigass heatsink like the TRUE or similar on them, or for systems that are built around the concept of silence or small form factors like most HTPCs.

For me, the H50 would make sense, as I don't have great airflow where my computer is, and no case, no amount of fans or configuration of them, and no placement of it would fix that. On top of that, the ambient temp in my room where I keep it never goes below 75 at any time of the year even with the AC set to Freezing Balls and no electronics on, so I don't really have any cool air to begin with. Its a phenomenon I've yet to figure out. But at the very least, the H50 would hopefully be an ideal solution to keep my CPU cooler than what its running at now.
 
How is this product dealing with the evaporation over time? Even a nicely sealed setup loses a little liquid over time. Or is this a disposable replace every year or so product?
 
I'm still watching through the video, but I had some thoughts about the other comments people have been leaving in the discussion thread so far:

- Airflow direction: You definitely want to have this drawing fresh air in across the radiator. It is simple mathematics / physics - if you have liquid at a high temp running through a radiator, more heat will be dissipated if the air coming through the radiator is lower (fresh air from outside the case) than if it is warmer (hot air from inside the case). I suppose if you could get the inside of the case as cool as ambient air outside the case then it would no longer matter, but that would be tough.

- Comparison to top-end air cooling: As I said, I'm still watching the video and I haven't gotten to the results section yet. I'd expect this to rival the best air-cooler, though; I don't have hands-on experience with this yet, but I do with the older Asetek LCLC design and it does a great job. Where I work we use them to overclock along the same lines as this setup (3.5-3.6GHz on a Core i7 920) in combination with a Nexus 120mm fan and in the Antec P182/183/190/193 cases. It performs better there than anything else we've used, though I'll grant we haven't tried the TRUE recently... we did once, but the massive size and weight of those coolers prevented us from using them (our systems have to survive cross-country trips at the hands of UPS / FedEx).

@Kyle - I know the Corsair Hydro is supposed to be an improved design from the Asetek LCLC, but if you are interested in a side-by-side comparison or adding another cooler to your upcoming Q3 '09 shot-out I'm sure we can send you a sample. Email or PM me if you'd be into that.

*Edit* - I just read Kyle's post a few spots up, so I suspect the offer of reviewing a LCLC isn't going to be very appealing. Sounds like the Hydro was a bit of a frustration for you, and I'm sorry to hear that. Still not quite done watching the video, but I really like the approach to doing reviews!
 
Last edited:
So would this be considered an entry level watercooling solution?

IMHO absolutely. I personally see nothing enthusiast about it's performance.

I don't know the retail price, but I can tell you for under $200 you can get yourself a very nice, much better performing unit.

Whenever air outperforms water, there is a problem.
 
IMHO absolutely. I personally see nothing enthusiast about it's performance.

I don't know the retail price, but I can tell you for under $200 you can get yourself a very nice, much better performing unit.

Whenever air outperforms water, there is a problem.


I really don't consider this a "water cooling" solution. It is just a long flexible heatsink solution. :) Check out Steve's ducting done right 7 years ago. Time to get your rain downspouts out again!!
 
I'll try to pop in a less than golden Core i7 920 (one that won't clock beyond 4.0 GHz reliably), and test an original Asetek LCLC on it when I have the time... shooting for the same clocks Kyle went for, and post the results here. Yes, a less than perfect comparison... but it's at least something. The previous numbers I posted in another thread were probably on a better than average chip, which Kyle's chip definitely isn't.

WilliamMGeorge, new OEM 120mm Asetek LCLC units will be changing to the new design momentarily... so you guys will be the same as Corsair eventually. Let's hope that's a good thing...
 
How is this product dealing with the evaporation over time? Even a nicely sealed setup loses a little liquid over time. Or is this a disposable replace every year or so product?

I would also like to know. It would be really stupid to be forced to replace 80$ cooler every x months.

 
Rebel44/Justys,

The technology has a rated 50,000 hour MTTF according to Asetek, and will require no maintenance. The (hard) plastic tubing used is resistant to evaporation; so, it will not lose liquid through it's normal lifetime. This is how big OEMs like HP can use it successfully in mainstream products that get stored in warehouses and shipped around the world in various extreme conditions.
 
Hey guys, some questions I'll answer for you.

The tubing used is fixed and designed for low-evaporation. The liquid should not evaporate significantly for foreseeable usage (meaning 2-3 years, on average, before any performance loss from evap).

As for the goal of this product, it was to create a competitor for high-end aircooling that would give you similar or better performance in a wide variety of cases.

It's designed to perform well inside a computer case, we know that high-end aircoolers (with superior surface area) will outperform it on a test bench or in a super high airflow case.

We hadn't tested the product like Kyle did - we tested it internally (against the TRUE and other aircoolers) in three cases: Cooler Master RC690, Corsair Obsidian 800D, and Silverstone TJ09. In all three of these cases, we used the same hardware:

Asus Rampage II Extreme
Intel Core i7 965 processor (engineering sample)

We overclocked by raising the multiplier on the CPU, and we didn't overvolt past 1.3V.

Initially I'd done quite a bit of testing on Core 2 Duo and Core 2 Quad, and our numbers looked very good against aircoolers, and my numbers look similarly good in the cases I mentioned.

I would say that the main benefits of the H50 over high end aircoolers are:
-Better performance in enclosures with medium levels of airflow
-Much quieter than most aircoolers (especially since you cut down on an extra fan)
-Good performance for moderate overclockers
-Very reasonable price point
-Smaller form factor makes it easier to install.


It's not going to outperform a high-end watercooling solution, it's not designed to.

Kyle's absolutely right when he says this is more like a "long flexible heatsink solution". It's a quiet, high-performance heatsink that happens to use water as the heat transfer medium instead of the liquid inside the heatpipes of a TRUE or Noctua NH-U12P or something.

Kyle's testing was done in a way we hadn't tested internally - but it's certainly not surprising to us that in a high-airflow case, aircoolers look better.

As a side note - the 28 minute video was definitely one of the most thorough "previews" I've seen. We'll be using his numbers and testing methodology internally (we've already purchased identical hardware he used in his testing) to better determine the performance benefit.

This product works differently from aircooling - so keep an eye out for our website where we'll be doing a lot of popular cases and showing people how to get the best performance out of the cooler.

We have a lot of cases out there that we have to evaluate - we've tested on about 6 or so internally, and we'll be hitting popular cases with compatibility and performance suggestions in the near future.

If anybody has any questions - feel free to post them here and I'll address them when I get the chance.
 
How about a comparison between Corsair H50 and Cooit Domino. Both priced at $80 and both target high end air cooler. Great vid and preview.
 
If you have a medium flow case, we would suggest this how-to, written 10 years 2 weeks ago. That was originally formulated from this work about 11 years ago. :)

Well, I think we actually accomplished quite a bit with very little. With 5 case fans and one heatsink/fan combo I was able to bring the system temperature down by 10°F. That is over a 10% reduction in heat. This allows my system to coast in around 87°F while running at 400mhz and around 92°F at 448mhz. This is the temperature taken by the Mobo
 
That's sorta disappointing. I was kinda looking forward to seeing how it performed, but I guess it makes sense. I have an FT-01 so water cooling is rather tough to do on this case unfortunately.
 
Very Informative Preview Kyle! I was really interested in the Corsair unit, but I suspected it might come up just shy in this comparison. What I also enjoyed was the discussion about ventilation. I clearly remember overclocking an Intel Celeron 330MHz on an Abit BH6, and back then we had to cut all our own blow holes. :D

To illustrate how big an impact a well ventilated case has all I have to do is compare your temps, with the True, to mine with a Zalman CNPS9900. Now the two situations aren't really directly comparable. My 920 C1 chip runs at 3.6GHz with only 1.2875v. And I'm running Linpack instead of Prime95. But even with an inferior "air cooler", the superior ventilation of my HAF 932 case yields lower temps at the same ambient temperature (~26.6°C) by about 2°C. And truth be told, if I didn't want a higher speed clock I might not have had to worry about replacing the Zalman.

Still, the Prolimatech Megahalems arrived this afternoon. So I decided to replace the Zalman with it and see just how well it does... (Even though it will eventually see duty with a D0 stepping 920 on an EVGA Classified board.) I have to see for myself that the new cooler is really better. :cool:


__________________________________________________________________________________________________________________________

Note: even without waiting for the Artic Silver to cure (yes I'm still trying to use up a 16g tube!) the Prolima-Tech Megahalems is about 5°C better than the Zalman. I'll still wait for it to cure to run a final series of tests, before I break it down and move the Megahalems to the new system. But I'm thinking I may not have to go with water, we'll see!
 
Last edited:
Kyle,

First, thank you for the video, its long and I did watch the whole thing. There has been a lot of hype about it and seeing first hand the testing conditions (in the case), is really nice.

As you pointed out, it has its good points, but clearly once it reaches its max thermal dissipation its game over for the little cooler. The price point is close, I think if they could get it to around $65 it would be a bit more attractive for a certain group of people.

I can't wait to see the HSF round up along with the Corsair unit being included and again thanks for the video!
 
I'm so jealous of you seeing all that HARDware sitting on your desk. I do run the same case and have all 4 fans in the door. Good video I'm looking in to water cooling and glad so see some info on that unit. I did not think it would keep up with my I 7 over clocked, it is good to know up front how good it does.
 
Falcon Northwest i7 with Asetek LCLC
falconcorei7_9_600.jpg

(Engadget photo)
 
For those of you interested, as of right now the H50 uses a different cold plate than any other LCLC-based unit on the market. I imagine Asetek will eventually move that plate into production but right now, we have the best performing plate on the market by a few degrees.

Also our fan spec is designed to be more quiet.

For performance guys you can replace the 1700RPM fan included with a faster fan (or two in push-pull) for significant performance gains. We just felt the 1700RPM fan was the best balance for performance and noise level.
 
Good job on the video , bit easier to watch than wading through pages of data.

The radiator to me looks like the weak link on the Corsair.
For load testing I found prime95 and intelliburn together put as much as 6c on top of just running Prime95 alone.

Looking forward to see more videos in the future. This one was great.
 
For those of you interested, as of right now the H50 uses a different cold plate than any other LCLC-based unit on the market. I imagine Asetek will eventually move that plate into production but right now, we have the best performing plate on the market by a few degrees.

Also our fan spec is designed to be more quiet.

For performance guys you can replace the 1700RPM fan included with a faster fan (or two in push-pull) for significant performance gains. We just felt the 1700RPM fan was the best balance for performance and noise level.

Corsair originally told me that it was a 1700 RPM fan then later told me it was a 1600 RPM fan + or - 10%. My fan ran at 1660 RPM.

I measured dbs between Thermalrights 1600 and the H50 fan and found no discernible differences.

I would suggest that the RPM and blade pitch of the fan makes it nice and quiet. Moving up to 2K RPM fan will drop you temps about ~5C in my testing....but 2K is considerably louder or has a more recognizable sound profile or pitch.
 
meanmodda[H];1034252180 said:
Good job on the video , bit easier to watch than wading through pages of data.

The radiator to me looks like the weak link on the Corsair.
For load testing I found prime95 and intelliburn together put as much as 6c on top of just running Prime95 alone.

Looking forward to see more videos in the future. This one was great.

Thanks for the kind words. Putting together something of this length and detail was NOT easy. Had three hours of video shot to edit through.

Roger that on your note above on loads. I am finding that even lonely Prime95 is producing loads on the i7 that are simply scaling above anything I have seen in the real world, so I think that pushing it beyond Prime95 is not realistic....but surely you can do it if you try. (Might have some Folding guys sound off on this.)
 
Fine video.
Through as usual.

Couple of observations, too bad Corsair gave up on the Nautilus. I have one on one of my systems and that bad boy could cool a Q6600 and an 8800 GTX without breaking a sweat.

The other thing about the Nautilus was you could use any block you wanted......I realize with this the pump mechanism makes that impossible.

I would think blowing heated air into the case would be a real deal breaker for alot of folks.
 
Fine video.
Through as usual.

Couple of observations, too bad Corsair gave up on the Nautilus. I have one on one of my systems and that bad boy could cool a Q6600 and an 8800 GTX without breaking a sweat.

The other thing about the Nautilus was you could use any block you wanted......I realize with this the pump mechanism makes that impossible.

I would think blowing heated air into the case would be a real deal breaker for alot of folks.

I have had a Nautilus 500 on my personal C2Q system now for the last year +. Been a great unit, sorry to see it go. This H50 is not comparable and referring to the H50 as a "water cooler" is a bit of a travesty. ;)

In regards to "blowing heated air into the case," this is exactly what almost all heatsink units do. The only difference between the H50 and a traditional HSF is the fact that you mount the fins on the side of the case to pull through ambient air instead of using already "hot" air inside the case.
 
youd think a slightly better fan (or better yet a noctua fan) and a slightly thicker rad and reservoir would help, I know on my antec 1200 id have the fan fliped exhausting the warm air, or pushing air from inside my case over the rad then out

but i could see a poorly ventilated case making that not work
 
Nice and thorough video.

That corsair thing has too small a radiator (looks thin) plus the "water" hoses look pretty small too. Not suprising the performance was just adequate. But the whole thing comes across gimicky.

The premise behind the corsair unit could possibly be re-engineered and something better created.

I've been running my 920 at 4.0 ghz (200 blck), 1.3v qpi, cpu (on auto) at about 1.41v. And my temps were only hitting 80c or so. From what I'm seeing on yours and others temps I must have gotten lucky on my 920. Cooling it with a TRUE 1366 :)

Any workshops planned for this summer? Demonstrations like this would be so cool.
 
That was awesome.

I can't say I'm surprised about the TRUE results. And I know you had excellent airflow set up for it, but it's a great sink if you've got the room.

Looking forward to the roundup. Nice job!
 
*cries uncontrollably*:(:( I'm so confused! I thought the intent was to get hot air out of the case, as fast as possible? With all the extra heat from other components and all... oh hell, why not just stick a floor fan on the open case and be done with it?:mad:

Maybe I'll just create a case that uses liquid nitrogen and you just slip in the mobo. :D

Hey Corsair, make the H50's plate out of diamond. I'll buy one. :D
 
@Kyle - One thing I was curious about in your testing was whether you load down the video cards as well as the CPU? I know I saw some posting about ways to stress a CPU beyond even what Prime 95 does, but it might also be interesting to factor in a video card or two under heavy workload to see if the additional heat in the chassis makes a difference. If you already do that, what program(s) do you use [just curious]?
 
hey guys, new guy here.

I am currently piecing together a core i7 920 system.

i am not planning on doing a large over clock, but i wouldn't mind bumping it into the low, maybe mid 3s. So i am looking at after market coolers. something like a TRUE, or a Noctua U12P. the big advantage im seeing here, is size. I have an antec 900, which with the big top fan and the general complaint that some of the bigger heat pipe coolers dont like fitting. this would seem to trim down the cooler alot, and if the top fan vents the hot air, would work pretty well.

thoughts?

cheers.
 
ivesceneenough, the Corsair H50, as well as the original Asetek LCLC, should have no problems cooling mainstream 3.5 GHz overclocks on even less than golden i7 chips.

I've had good luck with LCLC based solutions... also see the previous postings here from the Puget guy, who sells systems clocked at this speed using the supposedly lower performing original Asetek LCLC.
 
really nice video. i love the animation and the "lets see some shit". i would love to see more video reviews of products as well. i feel me and others can make a better informed choice seeing hardware in action on videos. im sure its time consuming but thanks for putting these few videos i do see together for us here at hardocp.
 
I copied this over from the other thread....just a passing thought I had.

I have a feeling the Corsair Hydro H50 will be a hot selling item. The going rate for them seems to be around $75 which puts it in line with high end air cooling which it easily matches or exceeds.

A good modder will position the radiator OUTSIDE of the chassis fully. Would likely be a little awkward, but could be done. In fact, if Corsair wants to sell more of these. Give me an extra 8 inches on the tubing and sell it with a plastic bracket that will allow me to mount easily outside of the chassis. Then you have the best of both worlds. Ambient temperature airflow on the radiator and no exhausting hot air into the chassis.
 
Back
Top