Temperature Results

Each Corsair Hydro Series cooler is mounted to blow air from outside the case over the radiator. We had to leave the top cover off of our 600T chassis during the H100 tests due to its restrictive design; otherwise the H100 results would have been horrendously skewed. The top cover didn't affect our other tests and was left on for those. Since the H60 has a PWM fan, we disabled any BIOS fan controls to ensure the fan was spinning at 100%. All three of the H80 and H100 speed settings are tested. We’ll start with our stock 3.5GHz test results and then move to the overclocked results.

CPU Delta Temperature - 3.5GHz (Stock)

The stock speed test doesn't provide much of a challenge for any of the Hydro Series coolers, leaving very little gap between them. Not surprisingly, all of the aftermarket coolers simply crush the stock Intel cooler; however, it’s also worth nothing that the Silver Arrow outperforms both the H60 and H80, and nips at the heels of the H100 set to Medium fan speed. Due to the relativity cool temperatures during this test, some of the speed profiles for the H80 and H100 never hit their maximum RPM. This keeps those coolers relatively quiet even when running on high or medium at stock speeds.

CPU Delta Temperature - 4.8GHz (OC)

Now that we're overclocked to 4.8GHz, these coolers are finally showing a temperature difference worthy of their price difference. Also note that the stock Intel cooler didn’t handle the 4.8GHz load reliably, so we don’t have results for it in this chart. The H100 is the runaway leader here, sporting a comfortable 3.3C lead over the H80. The H60 falls 5.1C shy of the H80—not terrible considering its slimmer radiator and single fan design. The cooling advantage of the H100 on the high setting does come at a penalty though; it's substantially louder as you'll see next in the noise test. The potential spoiler in the midst of these results is the Silver Arrow, once again coming very close to the H100 at a lower price. It’s not without compromises, which we’ll cover in the conclusion.

Test Setup and Procedures Noise Results
Comments Locked


View All Comments

  • bakedpatato - Monday, November 7, 2011 - link

    Good article; I now know that it's probably not worth it to replace my Xigmatek Dark Knight pre sandy bridge model because there's not that much of a performance delta.
    On that note, why do the big honking air coolers consistently outperform the sealed water cooling units?
  • LoneWolf15 - Monday, November 7, 2011 - link

    "On that note, why do the big honking air coolers consistently outperform the sealed water cooling units?"

    I'm guessing that the following limits hurt sealed H2O units, limiting their ability to compete.

    -Limited amount of liquid (as opposed to a system with a reservoir that would allow fluid to cool more easily during the cycle)
    -Limited volume of liquid that can be carried by sealed units tubing (both due to limited tubing diameter and the limited amount of liquid)

    Also, if you have that reservoir and larger diameter tubing, you can use a higher volume pump to deliver cool liquid faster. None of these are easy to do for a sealed kit, because it increases expense and reduces reliability. These limits prevent sealed systems from greatly exceeding the performance of quality air coolers.
  • Death666Angel - Monday, November 7, 2011 - link

    Unless your reservoir is in the region of several tens or hundreds of litres, the reservoir serves no cooling purpose. In sealed cooling solutions, you don't need it because of the sealed nature (at least if the manufacturer does its job of delivering an air-free system and handles loss of liquid due to dissipation from the tubes adequately). In open/standard water cooling you need it in order to fill and refill your cycle, to have a spot where the air can escape the system, to gauge if enough water is in the system or to get an estimate of the circulation (bubbling water can mean high circulation, flat surface can mean low etc.).

    Once a water cooling solution reaches a certain amount of water flow, increasing the amount will not help the cooling in any way, so the limited flow in these units does not negatively contribute to its lack of performance.

    You also make it seem as though these units use water to cool. That is false. They use air to cool, the water only transports the heat from the generator (CPU) to the radiator. Standard air cooling just cuts out the middleman (water).

    So your reasons are all false, I'm sorry to say.

    It boils down to simple thermodynamics. Look at the surface that these Corsair radiators provide for cool and then compare them to that Thermalright monster and you have your reason for why they don't do better.

    The reason DIY watercooling is so high end is because usually they have at least twice to three times the radiator surface compared to the H100. People with SLI/CF configurations usually go for a MoRa which has 9*120/140mm fans and the respective radiator surface to go along with it. But these tiny 120mm or 240mm radiators used by Corsair do not greatly increase the surface that is available for cooling.

    Also, geometry plays a role in cooling. In order to increase surface in the same dimensions, you can make the fins be more narrow, thus having more fins in the same space. This then influences the air flow through the fins. So narrow fins can be better at high air flow (provided by more powerful/louder fans) or less narrow fins can be used in order to obtain a passive or low-rpm cooling method, which is less powerful.

    (English is my 2nd language so I'm not sure I use all words 100% correctly. Hope I'm understandable.)
  • JarredWalton - Monday, November 7, 2011 - link

    "People with SLI/CF configurations usually go for a MoRa which has 9*120/140mm fans and the respective radiator surface to go along with it."

    Seriously? I've seen plenty of water cooling solutions over the years, but I have never seen anyone with an apparently 360x360mm radiator and nine fans (let alone a 420x420mm radiator with nine 140mm fans). I'm sure that some extreme water-cooling types go for this sort of setup, but I'd wager far more are using 120x240mm radiators, if simply for the fact that most cases won't easily accommodate a larger radiator. If you go with an external radiator tower, sure, you can have a larger setup, but then you really never plan on moving your PC. Other than that, though, I'd agree: it's about surface area, and that Thermalright Silver Arrow is a monster.
  • Death666Angel - Monday, November 7, 2011 - link

    Look through the water cooling forums (one I frequently use is this: http://www.hardwareluxx.de/community/f137/), look especially at the picture threads.

    People who only cool their CPU might go with a 240 setup, but that doesn't really work more quietly, more efficient, cheaper or easier than simple air cooling, as you have shown. So the people who invest the time and money to get water cooling going usually start with a 360 radiator or 420, depending on the case. They can be easily integrated in the usual water cooling cases that get recommended. People who go with SLI/CF will either get 2*360 (or similar), 420+280 (one in the bottom/top, one in the front) or make it easier and mount a 1080/1260 either externally on a side panel or use an external stand. The additional cost of more/bigger radiators is also easily offset, since the base components can remain (CPU heatsink, tubes and fittings/connectors, pump...).

    For my first water cooling I'm going with a 480 (2x2 120mm) radiator mounted on one side of the TJ08-E, which might get expanded to 2*480 once I start cooling my GPU (not going to buy a new card before the 28nm generation and my HD5770 is fine with air for now).

    Of course, my sampling of the water cooling community is limited and there might be a lot of people who don't frequent special forums. But I personally doubt it, since you do need kind of a lot of knowledge about the right tubes, fittings, sizes, water components etc. And the newbies who as in those forums and want to go with "just a little water cooling for the noise" usually get recommended (and convinced of) good air coolers like the Silver Arrow. :-)
  • sticks435 - Monday, November 7, 2011 - link

    I would look in the water cooling threads at overclock.net instead. Most people recommend a 120x3 for an overclocked CPU and GPU, and a 120x4 for an extra GPU. Maybe if you were running Triple or Quad SLI/CF you might need something like what you mentioned, but that is definitely the exception instead of the norm.
  • Death666Angel - Monday, November 7, 2011 - link

    Maybe if it is a high-end radiator or noise isn't that important. But like this review shows, a high end air cooler already competes with one 240mm radiator. And a CPU only uses between 100W and 150W depending on the overclock, whereas a GPU even at stock uses 200-250W. Using only one 360 radiator for a setup that can use 300-400W sounds very noisy and it looks as though air cooling could do a better job then, 2*140mm fans heatsink for CPU and triple 80mm fans for GPU for example are relatively easy to come by, fit in most cases more easily than a whole water cooling kit. :-)
  • Death666Angel - Monday, November 7, 2011 - link

    But, to nip this discussion in the butt: unfortunately neither of us have any real data, so we can only go off our instinct and experience so far. Yours is different from mine, no sense arguing that. :D
  • Etsp - Monday, November 7, 2011 - link

    to nip this discussion in the bud*
  • n13L5 - Monday, August 27, 2012 - link

    If I had a big case that could fit a silver arrow, I certainly wouldn't even be looking at these water things. Cause they are risky by comparison... a pipe could bust and flood the computer with whatever corrosive liquid is in there etc...

    But if you have a small case that has little room above the mainboard, these little systems don't need to beat the silver arrow at all... even the H60 will easily beat anything I could fit over the CPU in an SG05 or a PC-Q08 case. And, the heat goes straight out of the case, not spread around inside the case. I don't know if this test took into account that the mightly Silver Arrow just removes the heat from the CPU but not from the case.

Log in

Don't have an account? Sign up now