Corsair Hydro Series H50 Review

Add Your Comments

Corsair has been known for a long time for their great RAM. When they entered the power supply market, they established themselves as a force to be reckoned with. Recently they entered the case market with their Obsidian 800D to equal accolades.

Today we’ll be looking at the Hydro Series H50, their attempt at breaking into the CPU cooling market. Thanks to Corsair for supplying this unit for review.

Specifications

Per Corsair’s site, the H50’s features:

  • Pre-filled, closed-loop system is easy to install
  • Copper CPU cooling plate for maximum cooling performance
  • Integrated pump and reservoir is sealed for zero maintenance and improved leakage protection
  • Large 120mm radiator for fast heat dispersion
  • High-efficiency, low-noise 120mm fan for drawing cool air across the radiator
  • Two-year warranty

With specifications:

  • Model CWCH50
  • Cold Plate Material Copper
  • Fan Specs 120mm, 1700 RPM
  • Radiator Material Aluminum
  • Tubing Low-permeability for near-zero evaporation

We also asked Corsair some additional information that isn’t available on their web site:

  • Fluid: Deionized water with propylene glycol to prevent corrosion.
  • Thermal Interface Material Manufacturer: Shin-Etsu
  • Original Equipment Manufacturer: Asetek makes the radiator/pump/block assembly to Corsair’s specifications.
  • Fan Manufacturer: Akasa

Packaging & First Impressions

The outer box is strikingly similar to Corsair’s power supply boxes. It’s a sound design, so why not? Inside is a well packaged unit, immobile and encased in thin plastic flexible enough to absorb some shocks along its journey to the user.  We see in the upper right on the back of the box that the OEM is asetek.

The Box

The Box

Box Rear

Box Rear

Packaging

Packaging

When removing the unit from the box, we were struck by how light it was. It seemed to weigh about as much as the Apogee XT we reviewed. Weight doesn’t determine performance, of course.

When perusing the box contents, we see the cooler includes hardware for installing on LGA 775, 1156, 1366 and AM2+ as well as a 120mm fan to pair with the radiator. Not every cooler can claim such wide compatibility out of the box, so nice job Corsair!

Installation Hardware

Installation Hardware

Included Fan

Included Fan

The fan is an unknown make. It felt similar to a Yate Loon, so we compared. If you look where the blades meet the center hub, there is a small bend on the Corsair that differentiates it from the Yate. So the make is a mystery. As far as the fan itself, the sleeving was very well done and will keep your installation looking clean.

[EDITORS NOTE: Corsair got back to us – The fan is manufactured by Akasa. –hokiealumnus]

Fans Compared

Fans Compared

Now we come to the main event, the pump/block/tubing/radiator assembly. It’s one piece and you cannot disassemble it. It feels solidly assembled and not fragile at all; you probably couldn’t get it apart without a good bit of effort. It’s good to see they paid as much attention to the sleeving for the pump power as they did with the fan sleeving. Overall, it’s a nice looking unit.

Cooler Assembly

Cooler Assembly

Pump/Block

Pump/Block

Radiator

Radiator

The radiator appears to have a rather high 20FPI (fins per inch), which would put it on par with a HWLabs GTX radiator. Such high FPI radiators are generally associated with needing rather strong fans to get the most out of them. That’s not quite as much of a problem with this particular radiator because it’s not very thick.

FPI Measurement

FPI Measurement

The included thermal interface material comes pre-installed. The application looks a little thick but performance didn’t seem to suffer and when the cooler was removed, the contact pattern looked just fine.

Thermal Interface Material

Thermal Interface Material

Enough looking around though, let’s get it in and see how it performs!

Installation

The installation procedure for the H50 is painless and very well thought out. For this review, it was installed on an LGA1156 platform. The LGA backplate is one-size-fits-all, with three holes in each corner to accommodate all three of Intel’s recent CPU mounting hole configurations.

There are approximately 1/4″ thick “nuts”, for lack of a better descriptor, which you insert into the back plate. There is double-sided tape to keep it the back plate secured to the board for ease of mounting. You do not need to use the tape, but it may make things easier. Indeed, you may not want to use it if you plan on using this cooler in multiple systems.

The pump assembly hold down plate also has inserts that place the bolts in the proper position for your socket.

Back Plate

Back Plate

Hold Down Bracket

Hold Down Bracket

The beauty of this installation is that you install the back plate and hold down bracket before you install the cooler, which makes that step rather effortless. You partially tighten down the hold down bracket, push the pump/block assembly in and twist slightly, then tighten down the bracket the rest of the way.

Bracket Installed

Bracket Installed

Our only complaint about the mounting hardware is that the back plate is plastic. The modular design pretty much requires that, unfortunately it just feels less than solid when tightening. If you crank down it too hard, you could cause the nuts to strip out the part of the back plate that keeps them secure. It would take some effort though, so it likely won’t happen under normal circumstances. Just make sure you don’t over-tighten. We’d feel better about a metal solution.

Aside from that minor qualm though, we can’t say enough good things about this mounting. From the wide compatibility to the brilliantly modular installation assembly, it’s a very well thought out piece of hardware. Here’s what it looks like after the push, twist and tighten maneuver.

Installed

Installed

Installed

Installed


Testing Methodology and Results

The CPU being cooled for this review is an Intel i7 860, a furnace of a CPU. If you really want to test a cooler, an i7 is the way to do it. Tests were run at stock speed, overclocked to a moderate overclock of 3.6GHz and 3.8GHz, the 24/7 overclock of this system.

The H50 was compared for reference at stock speed against the stock Intel cooler. At stock and overclocked the main competition is a custom water loop, consisting of:

  • Swiftech MCP-355 pump with XSPC reservoir top.
  • Swiftech MCR-320 radiator, mounted externally with three Ultra High Speed Panaflo fans running at 7V (+/-.1V) for all tests.
  • Swiftech Apogee XT water block.
  • All linked with Primchill Pro LRT 7/16″ inner diameter tubing.

Unfortunately there are no good air coolers here to add into the mix. Bear in mind when viewing the results that the water loop the H50 is up against is way (way) more expensive, coming in at $270, and that’s not including tubing and barbs.

Testing consisted of no less than an hour and a half of Prime95 Small FFTs (except for the stock cooler, which only lasted about five minutes…we didn’t figure the CPU was a glutton for punishment). Most tests were run significantly longer than that (up to about 7 hours). CPU temperatures were measured using Coretemp’s logging function and ambient was measured with a common room thermometer to the nearest 0.5ºC.

The first hour of the temperature results were thrown out (to give the loops plenty of time to warm up) and the remainder were averaged for each core. Presented in the graphs are the average of all four cores at idle, the average loaded temperature of all four cores and the average maximum temperature of all four cores. Rather than present you with a deluge of data and screenshots, we’ve made some graphs to make it easier to digest. All temperature results were normalized to 22ºC ambient (per the esteemed Vapor, “…ambient and core temps scale perfectly fine (1:1) with i7.”)

Not to disappoint anyone that likes to pour through screenshots and excel spreadsheets, the test data can be downloaded in its entirety from Overclockers Tech here (it’s just under 4MB).

First off, we have the idle temperatures.

Stock Temperatures

Stock Temperatures

So, just in case you didn’t know, the stock Intel cooler is horrid. The good news is that the H50 did a good job at coping with this little heating element at stock.

Moving on, let’s see how it copes with a moderately overclocked i7. You’ll notice in this next graph that there are a few H50 results. It was tested in three configurations.

  • With the included fan.
  • Push-pull with the Corsair fan and a high speed Yate Loon placed in push-pull (and extremely under volted…more on that in a minute).
  • Finally, with the high speed Yate Loon by itself cranked to the max and a shroud in place of the Corsair fan.
3.6GHz Temperatures

3.6GHz Temperatures

Here we see the custom loop starting to come into its own, increasing its lead to about 13ºC. The H50 doesn’t do too bad for itself though. One thing is for sure, if you plan on putting this cooler on an overclocked i7, a fan upgrade is in order.

Regarding the push-pull results, we had to run the High Speed Yate Loon at 3.3V (which was amazing in itself that the thing still ran) or the Corsair fan put forth a rather loud, extremely annoying sounding protest. This stands to reason, as the Yate was pulling the Corsair faster than it was designed to go. What you should take away from this is that if you’re going to run this cooler in push-pull, get two stronger fans to do so. The Corsair fan just isn’t strong enough to cope well enough with an overclocked i7.

On the plus side, even changing it out for one improved fan is a nice boon to the H50’s performance, gaining almost three degrees over the stock fan.

So let’s see what happens when you give it a little more voltage and squeeze out an extra 200MHz. For this test, we did not run the Corsair stock fan or the push-pull orientation. When the CPU starts going too far north of 70ºC we start getting squeamish. Since the high speed fan was knocking on that door already, we decided to skip those other two configurations.

3.8GHz Temperatures

3.8GHz Temperatures

The custom loop actually gave back a little bit in this test, with its lead dropping to around 10ºC. The H50 certainly isn’t going to win any best water cooling awards but it does well for itself. These temperatures are about on par with the best air coolers but with the added benefit of being at a reduced noise level.

Final Thoughts and Conclusion

$77.89. That’s what you’ll pay for the Corsair H50 at Newegg. Overall, not a bad price for what you get. Right at what you could get a Prolimatech Megahalems with two good fans in push-pull ($60 for the Megahalems and $33.90 for two Panaflo fans at Jab-Tech) will get you the H50, an upgraded fan and some additional silence to help your long term hearing prospects. It will likely net you a few degrees better cooling to boot.

Is this water cooling? Technically, yes. But no, it will obviously not compete with a good custom water loop. It also doesn’t cost near as much. What it will do is replace your air cooler and give you some more peace and quiet. That’s where we see the market sweet spot for this unit.

Pros

  • Superb installation system.
  • Wide range of compatibility.
  • Quieter than air cooling with equal or slightly better temperatures.

Cons

  • Sorely in need of a better fan.
  • Unit cannot be refilled.
  • Plastic back plate.

With these things in mind, we award the Corsair Hydro Series H50 an 8/10!

– Jeremy Vaughan (hokiealumnus)

Leave a Reply

Your email address will not be published. Required fields are marked *

Discussion
  1. Yep, you summed it up. With fan improvements it's a good top notch cooler, possibly a bit better than the top air cooled heatsink. But not near as good as a full blown watercooling loop. The price difference sets the users choice.
    Ohh one thing you didn't mention, how quiet the custom loop was vs the H50 on HS fans.
    Thanks, looks like all of the questions are answered now.
    Well, even at 7v UHS Panaflo fans aren't the quietest fans around. I'd say one HS yate at full bore but contained in the case is roughly equivalent to the same noise level as three under-volted UHS fans. If those were internal instead of externally mounted, the nod for quiet would probably go to them. Likewise, if I didn't have those beasts on there it would definitely have been quieter.
    Nice review :) It would have been beneficial IMO to include a top aircooler in the tests as the Corsair really can't compete with a proper hand picked water cooling system.
    hokiealumnus
    Thanks! FWIW, I completely agree, just didn't have one available and couldn't afford to buy one...primarily because of my loop. :)

    If you can't afford something for a test (review) please ask me and I'll make it happen. :)
    Corsair responded to our inquiries. For anyone that already read it, the additional questions answered in the article are:

    • Fluid: Deionized water with propylene glycol to prevent corrosion.
    • Thermal Interface Material Manufacturer: Shin-Etsu
    • Original Equipment Manufacturer: Asetek makes the radiator/pump/block assembly to Corsair’s specifications.
    • Fan Manufacturer: Akasa
    Nice work hokiealumnus! :salute:
    For AMDs I'm sure this would match a high-end air solution but, bottom line, it's still only one fan, which can only remove so much heat. That was plainly shown when you juiced the i7 a little and got a 10°C jump - pushing the limits of the H50 for an i7 OC.
    The difference between high-end air and any water solution is very apparent for the average user who doesn't push his rig all the time. Few games will run a CPU at load for more than a few seconds at a time, which is enough to overload a heatpipe but not nearly enough to drastically effect the thermal capacity of a water loop. As such, the loop may show a very minor increase in temps while the air cooler will jump several degrees in this short time. So what we have with the H50 is the best of both worlds - low cost of high-end air combined with the heat capacity of water cooling to flatten out those temperature fluctuations ... :)
    I would bet this would hold an i7 overclcoked with a dual rad/fans notabley better...for say $99................Then it would expand a bit more into our 'demographic' of enthusiast overclockers.
    But with a 120x2 rad you've got a mounting problem with a sealed loop. Where could you get by mounting it without cutting up the case in some way? I don't have a problem taking tin snips to the rear fan guard but I doubt that would make for a good retail solution ... ;)
    Sealed heh, forgot about that somehow. But there are several cases with mounting available for dual 120 rads without *snip snip*, but that would cut down the size of the market available making it not as good (at least) of a retail solution.
    I'm curious but would it be possible to also look at this using a pair of Noctua NF-P12's in a push/pull setup? I'm looking to get one of these to replace my Zalman 9700NT for when I upgrade to a Phenom II 955 installed. I've heard a few success stories going with this setup for both cooling and noise so it would be nice to see that verified in a review.
    I figure that if this works as well as I'm hoping with mine I can look at trying to tame my hubby's system with an H50 in a similar setup that is a bit noiser than I like even with Noctua NF-P13's on the intakes. -_-
    Unfortunately not, but I can point you to a review of the NF-P12 on a TRUE. I think it would do well with two of those in push-pull. IMHO though, there are fans not equally as good or quite as quiet that will do the job very well (possibly even better) for a whole lot less money. But then again, I'm admittedly cheap and you're looking for quiet. For that, Noctua is bettered by very few, if any.
    The push/pull setup in this review was running ~1,500 RPM. That said, those fans aren't Noctuas either and the comparison wouldn't be accurate.
    Oh, and :welcome: to OCF! If I knew my wife signed up here to get advice on something for my computer, I think I'd faint.
    I've read that push/pull on the h50 makes for truly impressive gains. If i were them i would put a thicker core in and a beefier fan, though that'd raise the price above the air coolers it's competing with.
    hokiealumnus
    Unfortunately not, but I can point you to a review of the NF-P12 on a TRUE. I think it would do well with two of those in push-pull. IMHO though, there are fans not equally as good or quite as quiet that will do the job very well (possibly even better) for a whole lot less money. But then again, I'm admittedly cheap and you're looking for quiet. For that, Noctua is bettered by very few, if any.
    The push/pull setup in this review was running ~1,500 RPM. That said, those fans aren't Noctuas either and the comparison wouldn't be accurate.
    Oh, and :welcome: to OCF! If I knew my wife signed up here to get advice on something for my computer, I think I'd faint.

    Bummer since it looks like you guys covered a bit more about the H50 than I've seen from other sites, I mean all the ones I saw before this one used just the stock fan, and it would have been interesting to get your pro opinion on it being overclockers and all. ^_^ Very interesting review link you mention and it looks like the NF-P12's may fit the bill since I really would like to avoid using the stock one due to the noise levels some others have reported with using it.
    I've been interested in water-cooling for quite some time but this is probably the first time it's gotten down to a point where I can dip my toes into the pool so to speak without risking my hard-earned gear. Believe you me all the troubles some of you have passing hardware costs along to your spouse I deal with as well so I really have to make each bargain session count.
    BTW, I did wish to note that I am open to suggestions as far as fan selection especially if it nets me a fan that is quite and performs close or equal to the Noctua's. So if you've got some ideas I'd be glad to hear them. ^_^
    And thank you for the welcome to your forums I really only just stumbled across this place but from what little peeking around I've done this may be a second home next to TPU for me! In either case I'm really hoping this provides greater cooling for lower noise than the Zalman 9700NT I'm using.
    If I do manage to get this gear approved by the "budget committee" and get things going I'll try and report back here along with some pics as to how things went.
    It certainly should cool better, the 9700 is ok, but not really that great. Most of Zalman's stuff (all the orbs) is more for show then for cooling power. The H50 is on par with high end air coolers. It should still beat the 9700 with a nice quiet fan.