Sign in with
Sign up | Sign in

Conclusion

Corsair H70: Next-Gen Self-Contained Liquid Cooling
By

With the H70, Corsair brings a liquid cooling system to market that strives to address the challenges of liquid cooling while still exemplifying its benefits.

It relieves the user of the hassle of having to assemble individual components, which should appeal to those without much experience in liquid cooling. With a street price of around $110, it is reasonably priced compared to other liquid cooling systems. By pre-filling the closed loop system at the factory, it is ready for use immediately after unpacking, and does not have to be much more complicated to install than some air CPU coolers.

In addition, Corsair offers a two-year warranty, and the producer Asetek indicates that the average lifespan of the liquid cooling system is about 50 000 hours. Installing the liquid cooling system is easy, and the detailed installation instructions for the different supported processor interfaces should simplify the process for less experienced users.

There are a few weak points, however. We would have liked the tubing between the water block and the radiator to be a bit longer than 24 cm. The same goes for the fan cables, which are a bit short at just 21 cm. The fans included created mixed impressions. On the one hand, the radiator with its very dense fin structure requires a certain air throughput--something the fans handle just fine. But on the other hand, these fans are not exactly the most acoustically-friendly. Quieter fans with variable speeds would be something to hope for here. You could look around for alternatives, but then you'd only be adding to the cost of this unit, which already exceeds the upper range of the highest-end air coolers.

However, if you can live with the slightly elevated background noise level that comes with the Corsair H70 liquid cooling system, you get cooling performance in the good to very good range in the upper segment of CPU coolers.

In the light of air cooling alternatives like Zalman’s CNPS 10X Extreme, which costs roughly $69, Corsair’s new H70 remains a premium-priced product. It's perhaps best suited to environments where space around the processor is tight, and you don't have room for a cheaper, comparable-performing high-end air cooler.

Display all 83 comments.
This thread is closed for comments
Top Comments
  • 26 Hide
    chovav , October 29, 2010 8:47 AM
    where's the H50 in the charts?
  • 20 Hide
    Scanlia , October 29, 2010 9:34 AM
    I think its stupid, just look @ the Prolimatech's...
  • 14 Hide
    karma831 , October 29, 2010 9:40 AM
    Overpriced, equal to or gets beat by air cooling in temps, and is pretty loud...Doesn't sound like a very good product ?__?

    The prolimatech and noctua offerings are much better.
Other Comments
  • 10 Hide
    wribbs , October 29, 2010 8:30 AM
    "street price of around $110, it is reasonably priced" -too high.
    "slightly elevated background noise" -understatement

    You will be better served by the h50, yes its weaker but its cheaper and quieter and performs close to the same. Do yourself a favor and avoid this product although it is quite good.
  • 26 Hide
    chovav , October 29, 2010 8:47 AM
    where's the H50 in the charts?
  • 20 Hide
    Scanlia , October 29, 2010 9:34 AM
    I think its stupid, just look @ the Prolimatech's...
  • 14 Hide
    karma831 , October 29, 2010 9:40 AM
    Overpriced, equal to or gets beat by air cooling in temps, and is pretty loud...Doesn't sound like a very good product ?__?

    The prolimatech and noctua offerings are much better.
  • 4 Hide
    maxal , October 29, 2010 10:19 AM
    Would it be possible to post combined charts of idle temp/noise and full load temp/noise ? IMHO that would be more informative than plain bars, as many are interested in balance of those characteristics.
  • 5 Hide
    gkay09 , October 29, 2010 10:25 AM
    I had expected to see even the CM Hyper 212+ results...
    IMO $110 for such a cooler not worth it,...Other good option if going for a closed loop cooler would be the CoolIT SYSTEMS ECO...
  • 13 Hide
    cmcghee358 , October 29, 2010 10:40 AM
    I would have also liked to have seen comparison between a push/pull exhaust, and a push/pull intake.

    For my H50, it is quite explicit that the radiator should be pulling outside air through the fins. And yet, people keep it as an exhaust and then lament of their temps.

    One of the biggest perks of a closed loop water cooling system is you dictate what air is being used. Where-as big air is always restricted to case air(unless a side is removed and a box fan is added)

    I would have also liked to of seen it as a delta above ambient, rather than a flat CPU temperature comparison. I personally process the delta better, and lend more credence to it.
  • 0 Hide
    uruquiora , October 29, 2010 10:46 AM
    would have liked to see the temps in passive mode, with no fans, and on a more recent CPU...
  • 1 Hide
    mjello , October 29, 2010 10:59 AM
    I want one for my gfx card not the cpu. Its my gfx thats noisy. I don't care about cooling temps or overclocking as long as the cooling is adequate. I just want sweet silence :) .
  • 0 Hide
    cmcghee358 , October 29, 2010 11:03 AM
    Quote:
    I want one for my gfx card not the cpu. Its my gfx thats noisy. I don't care about cooling temps or overclocking as long as the cooling is adequate. I just want sweet silence :) .


    I guess you missed the part where it was the noisiest thing on the list with a 57.3 dB at max?

    Also considering the TDP of some of these GPUs are between 190 and 300W, you can expect the H70 to work much harder thant he sample 130W CPU. So it will most likely be loud all the time because of the higher operating temps of the GPU.
  • -4 Hide
    rottingsheep , October 29, 2010 11:48 AM
    pentium d's best performance!
  • 3 Hide
    d1rtyju1c3 , October 29, 2010 12:18 PM
    Someone explain why they are using a pentium D for the test.
    I want to see the results on 1366/1156 CPU not an old pentium d.
  • -7 Hide
    TemjinGold , October 29, 2010 12:32 PM
    Isn't the point of water cooling to not have fans to deal with?
  • 1 Hide
    ohseus , October 29, 2010 12:58 PM
    D1RTYJU1C3Someone explain why they are using a pentium D for the test.I want to see the results on 1366/1156 CPU not an old pentium d.


    The reason is clearly stated in the page about the test configuration. RTFM.
  • -6 Hide
    d1rtyju1c3 , October 29, 2010 1:29 PM
    Quote:
    The reason is clearly stated in the page about the test configuration. RTFM.




    I could care less if the Pentium D is less efficient than a 1366/1156 CPU. People dont really need to know the coolers results on an old Pentium D.
    Yes, I know that more than likely if it performs good on the Pentium D then it will definitely perform good with the i7, Blah Blah Blah, show me, That is what people want to see. For someone who dont know and is looking for a good review on how the H70 performs on the i7/i5 this review is pretty much useless to them.

    It is one thing to test new components on the Core2 CPUs but an old Pentium D come on. Even an old 45nm Core2quad or C2D would be less efficient than an i7/i5, not only that more people own the C2Q/C2D.
  • 1 Hide
    Anonymous , October 29, 2010 1:32 PM
    Perhaps I missed something but:

    - which configuration was used? push/pull
    - air from outside the case or the inside

    The orginal fans included with the H50/70 are crap. The firt thing I did, changed them for 2 Scythe fans. The temperatur is much better, the same thing for noise level.

    Want to see what this cooler is realy capable of:

    http://extreme.pcgameshardware.de/offizielle-pcghx-wasserkuehlungstests/112351-kompaktkuehlung-corsair-h70-im-test.html

    It's german but you don't have to understand the language to interprete the charts, particularly "H70 @ 2 Serienlüfter @ 1.600 1/min
    Spoiler:"
  • -1 Hide
    rubix_1011 , October 29, 2010 1:34 PM
    This isn't real watercooling, and H50/H70's are known to have weaknesses and flaws. H50's were prone to leaking, and on both units, the pump is INCREDIBLY underpowered. You are MUCH better using a high quality air cooler for your CPU.

    Yes, Corsair and a few of these companies making these *cough* CRAP *cough* self contained coolers also make crap GPU coolers for those of you interested.

    I'd suggest you check out the Forums > Overclocking > Coolers and Heatsinks for more discussions on Tom's (I am a regular for watercooling out there) as well as Overclock.net forums...or Xtremesystems forums. Creating an 'idiot proof' watercooling unit like this needs to be a much better performing unit if they are charging this...there are far better alternatives for beginners for watercooling.
  • 0 Hide
    socrates047 , October 29, 2010 1:46 PM
    im sporting a big goofy smile now that i reflect on my decision to go with prolimatech :) 
  • 1 Hide
    kelemvor4 , October 29, 2010 1:51 PM
    Have to agree with other sentiments expressed. If there are air coolers that match the performance of h70, then h70 is a failure at being much more than a gimmick.

    They should offer 240 and 360 models, 120 just doesn't dissipate enough heat.
  • 2 Hide
    rubix_1011 , October 29, 2010 1:55 PM
    @kelemvor4:

    You are quite correct. For the price/performance curve, it fails against other air coolers costing more than 50% less.

    Flat performance vs. even beginner watercooling is dismal...even against the ThermalTake bigwater kits...which we all try to steer others away from.

    The radiator needs to be at least a 220 AND offer better flow rates. Their site doesn't even state what the pump flow is...but a few have done some tests and its very pathetic.

    Gimmick? Yes, very much.
Display more comments