Performance Evaluation and Testing of Corsair H110i GTX Review Summary
Performance Evaluation and Testing of Corsair H110i GTX Review Summary
You're planning to handle additional AIOs too, right? Will everything fit neatly into this single section? You might want to consider revising the main title to reflect more than just one AIO. How much progress has there been on adding a dedicated water cooling part?
This is one of the 1st detailed reviews on a CLC that I have seen, and very well done (martins being the benchmark to date). I won't repeat the decibel "times as loud" and other clarifications that have already been addressed. One of my disappointments in CLC reviews is that most reviews ignore some very basic issues that I think users will like to see addressed. Things I'd like to see as these type of reviews go forward:.
1. Eight times as loud as ambient ... It is intended that we sit in the same room with these things right :0 ?
Here the H100i was shown to be about 12 times louder, not as compared to ambient background noise but as compared to a better performing air cooler under full load. The H110 does better than the H100 in this respect but just how far away (numbers included) they are from air coolers, which have better thermal performance should be a relevant topic in any review.
2. It was asked above about the better air coolers. The Noctua DH-15 and Cryorig R1 both outperform the CLC in the test above, the H100i was 8 times louder than the Cryorig 12 times the Noc. A teeny step down would be the Phantels PH-TC14-PE with the DH-14 and Silver Arrow just below that.
3. I have a concern about the "should never need maintenance". As an analogy I will use the Bic Style razor where you "never have to change blades".... because you throw them away. In the first place we have mixed metals which violates the 1st rule of PC water cooling. For those interested in the science (and seeing the pics), they can read this:
https://martinsliquidlab.wordpress.com/2...-explored/
In the second place, one can use corrosion bacterial / algae inhibitors to inhibit such action but they are subject to reduced effectiveness over time. Most water coolers flush and replace between 12 and 24 months (18 oft recommended). In professional applications (i.e. power generation), the coolant is sampled, tested and adjusted every 90 days. The effectiveness of these agents might be expected to last 18 - 24 months and without them I wouldn't expect the corrosion to start to have a serious impact on watertight integrity (see pics at above link) for a period of similar length. So like the Bic Disposable Razor, I think of it as a "disposable" component, not something I would take and use again after retiring the original PC it was installed in 3 - 4 years ago. With a radiator, or block (assuming mounts works) I have no such concern.
4. The high fpi radiator also means that a) it needs higher rpm fans which is the main reason the units are so loud. Traditional rads are usually 30, 46 and 60mm thick. The 30mm ones at more typical fpi do outperform these rads at typical fpi's. The Swiftech H220-X does outperform the same size H100i and the H240-X outperforms the H110 at lower rpms and lower fpi. Would be great to know whether the more capable pump or the copper radiator in the OLC is responsible.
5. One of the major shortcomings of CLCs remains the weak pumps. Be nice to know how the radiator on the H100 did with a more capable pump.
6. The part about "Of course, there are other high-end air coolers that will still run you less money with similar performance, but this is about liquid, isn’t it?" The question is should it be ? If the air cooler is 1/12th as loud, outperforms the liquid one and is substantially cheaper, what is the reasoning ? There are reasons for doing liquid cooling, increased performance being less relevant in recent years but aesthetics and sound reduction are both very valid reasons. The latter isn't going to come via CLC.,
6. The fact that OLCs are made out of all copper / brass, can be flushed and cleaned w/ inhibitors refreshed, as well as the ability to add more components AND have no qualms about installing in a 2nd or 3rd system over the years. These would be welcome topics / criteria when comparing pricing in future reviews. With the H110 GTX CLC at $140 being compared say with the Swiftech H240-X at $150 ... for $10 you get:
a) no mixed metals
b) ability to change coolant w/o voiding warranty if you desire
c) Pump is 5 times as powerful
d) Less than 1/4th as loud
e) thermal advantage of copper rads
f) included reservoir for level monitoring gives indication of any leakage
g) You can add GPUs, MoBo Blocks, additional rads and pump will handle it.
For your next project, I'd love to see a smackdown between the Swiftech and EK Predator lines.
While I'm waiting on the EK and Swiftech, I can do some delta tests against different flows using the H110i radiator, even modest ones - like 0.50 GPM and 0.75 GPM...maybe 1.0 GPM along with the same fan speeds (1200, 1800, 2300 RPM).
Maybe this can act as a decent comparison if the pump were able to push more than 0.25 GPM worth of flow. I can also check out the pressure drop on the radiator if we're going to give this 'supplementary' set of data a go.
Honestly, the information doesn't add value since you can buy aluminum radiators of similar size separately, and this isn't even an expandable AIO. The AIO comparison really centers on performance comparisons among the three models you're considering.
No one should invest that much in an AIO just to break it down or void the warranty for a different pump. It would be smarter to purchase the parts yourself, as testing such a scenario needs a proper independent water block for accuracy.
The type of pump used doesn't change much, but the overall capacity remains limited.
It's a complete waste of time gathering irrelevant data.
But it's your time too—remember, others are watching, so don't let them waste yours either.
Ideally, yes. My goal was to highlight the performance options if the pump in question were designed to work better. I understand this isn’t a major concern for most people, especially those who sell large quantities of these coolers.
For myself, this is useful for those trying to measure the downsides of closed loop coolers—especially the lack of suitable pumps. Having data on actual pump performance compared to slightly improved models helps illustrate the limitations of the cooler through chart analysis.
I regret if it seems like a waste of time, but from my perspective, it’s about raising awareness and understanding:
1) Why closed loop coolers are marketed at lower prices
2) The real difference in cooler performance versus pump capability, and how delta affects cooling
3) Encouraging community discussion about available options
Ultimately, it depends on whether Joe User decides to invest time in choosing wisely for a closed loop cooler.
And......how often do we encounter those Frankenstein setups where someone swaps an H50 for a GPU?
The reasons behind why closed loop coolers are sold at lower prices are discussed. It's about understanding how pump performance affects delta, which in turn influences cooling efficiency. The conversation also touches on the ease of installation and the benefits of simpler designs. Additionally, the discussion compares cooler performance against market options and highlights the importance of considering factors like manufacturing costs, convenience, and competition from heat pipe air coolers. The overall tone emphasizes making informed choices while acknowledging the challenges in fully understanding complex products.
The decision to start using the H110i came from the need to precisely manage the heat output during testing. Most reviews focus on attaching the unit and running standard software, but this doesn't capture the full picture since each test varies in CPU load depending on processor type and overclock settings. While a temperature chart is available, it often misleads by not reflecting the actual performance of the cooler.
We could instead follow a more thorough approach—installing the cooler, executing benchmarks, and creating a detailed graph. This method helps our audience move past the oversimplified idea of simply reporting CPU temperatures. Understanding overclock levels, voltages, and load patterns is essential for informed decisions.
I welcome any suggestions or feedback, as presenting data in this way requires a balanced perspective. If I weren’t open to it, it would seem dismissive and one-sided.
This approach uses watt-based thermal load tests rather than just comparing CPU models. This is because I can't test multiple processors at once, but I can demonstrate how overclocked CPUs behave under different loads. By doing this, I can clearly show the relationship between CPU demand and cooling capacity on a graph.
I remain receptive to ideas, so feel free to share alternatives if you think better methods exist.
Each entry will have its own summary and an index that connects all reviews, with results available in one place for easy reference.