Corsair H150i Elite Capellix

[crazyeyesreaper]

The cost of entry might be steep, but Corsair's H150i Elite Capellix manages to impress. In fact, it breaks the mold when it comes to the raw performance you can expect from an ARGB-equipped closed-loop liquid cooler. With a simple, clean design and good looks, it deserves a closer look if you're in the market for an AIO.

Show full review

 

[AnarchoPrimitiv]

The "Apple Tax" is well known to all, but do you think we can officially say now that there's a "Corsair Tax"?

 

[LocutusH]

I almost believed, that its a good cooler. Until i saw the noise results.

 

[Makaveli]

Looks good other than noise at maximum fan speed.

I have the original H150i which gives up performance to this by a few degree's but is much quieter!

 

[Tsukiyomi91]

performance is to be expected with those fans. If Corsair used the non RGB LED versions of the ML fans, the noise levels would be much lower & the overall price will be a little lower.

 

[Makaveli]

performance is to be expected with those fans. If Corsair used the non RGB LED versions of the ML fans, the noise levels would be much lower & the overall price will be a little lower.

Agreed.

Fan noise went up when they replaced the fans on the 2nd version of this aio.

H150I Pro RGB fans 400-1600 RPM 47.3 CFM
H150I RGB Pro XT fans 400-2400 RPM 75CFM
H150I Elite Capellix fans 400-2400 RPM 75CFM

 

[crazyeyesreaper]

I almost believed, that its a good cooler. Until i saw the noise results.

performance is to be expected with those fans. If Corsair used the non RGB LED versions of the ML fans, the noise levels would be much lower & the overall price will be a little lower.

Agreed.

Fan noise went up when they replaced the fans on the 2nd version of this aio.

H150I Pro RGB fans 400-1600 RPM 47.3 CFM
H150I RGB Pro XT fans 400-2400 RPM 75CFM
H150I Elite Capellix fans 400-2400 RPM 75CFM

So i am guessing no one read the conclusion?

Now, it is worth keeping in mind that these results are with the Corsair H150i Elite Capellix running at maximum fan RPM, which hit a highly annoying 55 dBA. However, with the noise level reduced to 45 dBA, temperatures only increased by 1 °C. This makes the cooler 10 dBA quieter and still has it rivaling or beating every other tested cooler. In fact, going a step further and dropping the cooler down to 42–43 dBA, which is the same as Noctua's NH-D15 in our tests, resulted in temperatures going up another 1 °C. Keep in mind that the Noctua NH-D15 hit 88 °C vs. the Corsair AIO at 84 °C at the same noise level. That is impressive considering most AIOs struggle to rival top-tier air coolers at similar noise levels.

 

[Makaveli]

So i am guessing no one read the conclusion?

Now, it is worth keeping in mind that these results are with the Corsair H150i Elite Capellix running at maximum fan RPM, which hit a highly annoying 55 dBA. However, with the noise level reduced to 45 dBA, temperatures only increased by 1 °C. This makes the cooler 10 dBA quieter and still has it rivaling or beating every other tested cooler. In fact, going a step further and dropping the cooler down to 42–43 dBA, which is the same as Noctua's NH-D15 in our tests, resulted in temperatures going up another 1 °C. Keep in mind that the Noctua NH-D15 hit 88 °C vs. the Corsair AIO at 84 °C at the same noise level. That is impressive considering most AIOs struggle to rival top-tier air coolers at similar noise levels.

No I read the conclusion.

I'm well aware that you can lower the noise just referring to out of box performance.

With both version 2 and 3 those adjustments will be needed to keep the noise level low.

 

[crazyeyesreaper]

consi

No I read the conclusion.

I'm well aware that you can lower the noise just referring to out of box performance.

With both version 2 and 3 those adjustments will be needed to keep the noise level low.

Considering how PWM works, if not manually overidden the CPU isnt getting hot enough to reach those levels at least on mainstream platforms (on HEDT etc that might not be the case).

So for LGA115x / AM4 etc its unlikely the cooler will ever hit 55 dBA in real life work loads. Plus the software does offer quiet / balanced / or performance modes as well which curbs noise. But I digress it is what it is not everyone will be happy but considering at 43 dBA it manages to equal the EK-MLC Phoenix 360 which is an all copper loop at 51 dBA I feel the max noise level is kinda moot. The only cooler remotely close to the H150i is the Enermax Liqmax III ARGB (much better price) but enermax has had issues with gunk build up so still not sold they fixed that but so far so good on my sample.

 

[Makaveli]

You are correct noise level will depend on what you are doing on the PC. And for the average desktop you probably won't hit that 55 dba level unless you are really pushing the system with an sustained load over a few hours.

For my setup I just leave icue set to performance mode on all fans and never have to touch it.

I'm not a fan of Enermax's aio for the reason you stated.

 

[Space Lynx]

$179.99 for a 360mm AIO that competes with closed loops from a great brand and long warranty. sounds like a great price to me. seriously considering buying a new case for this. i like my current case tho... bleh.

 

[wheresmycar]

If i;m buying a cooler for $180 i'm looking for a more balanced noise and performance profile. Isn't this the popular choice amongst most people? Not sure why these manufacturers can't offer options with more premium fans or no fans for less. Just bloody RGB RGB RGB

 

[Haile Selassie]

If i;m buying a cooler for $180 i'm looking for a more balanced noise and performance profile. Isn't this the popular choice amongst most people? Not sure why these manufacturers can't offer options with more premium fans or no fans for less. Just bloody RGB RGB RGB

This cooler has presets for balanced or high performance according to the reviewer so I'm not sure where's the problem?
You go all out and run max speed and suffer high noise or drop the noise level by 10dBA at a cost of 1°C.

I'm using the previous version, the H150i PRO XT and I like it a lot.

 

[krusha03]

So i am guessing no one read the conclusion?

Now, it is worth keeping in mind that these results are with the Corsair H150i Elite Capellix running at maximum fan RPM, which hit a highly annoying 55 dBA. However, with the noise level reduced to 45 dBA, temperatures only increased by 1 °C. This makes the cooler 10 dBA quieter and still has it rivaling or beating every other tested cooler. In fact, going a step further and dropping the cooler down to 42–43 dBA, which is the same as Noctua's NH-D15 in our tests, resulted in temperatures going up another 1 °C. Keep in mind that the Noctua NH-D15 hit 88 °C vs. the Corsair AIO at 84 °C at the same noise level. That is impressive considering most AIOs struggle to rival top-tier air coolers at similar noise levels.

I think this is why a noise normalized graph would be very useful (eg. Temp at 42 dBA)

While it seems that the performance is good, its quite expensive and if you dont need rgb the arctic liquid freezer II is much better value

 

[randomUser]

Why is there no Arctic freezer AIOs in the chart?

 

[Space Lynx]

Why is there no Arctic freezer AIOs in the chart?

it is surprising they have no reviewed that new lineup, gamersnexus said it is one the best

 

[wheresmycar]

This cooler has presets for balanced or high performance according to the reviewer so I'm not sure where's the problem?
You go all out and run max speed and suffer high noise or drop the noise level by 10dBA at a cost of 1°C.

I'm using the previous version, the H150i PRO XT and I like it a lot.

It was more of a generalisation. For eg, I bought a brilliant performing Kraken X62 in 2017 but the fans were schite. To hit 5ghz with a 7700K i was left no choice but to swap the fans out for something quieter. I'm just saying it would be nice for a handful of options across the board accommodating "performance" and "premium fans" as a bundle.

 

[FreedomEclipse]

Tbh, if im gonna be dropping this kind of money on a AIO then i would go for the Asus ROG stuff with the little LCD screen on the CPU block, even if its the 240mm model.

 

[crazyeyesreaper]

I think this is why a noise normalized graph would be very useful (eg. Temp at 42 dBA)

While it seems that the performance is good, its quite expensive and if you dont need rgb the arctic liquid freezer II is much better value

With new test benches being built noise normalized testing will be added. Along with Intel and AMD platforms. Need to finalize stock and oc settings then rebench a ton of coolers so will be a bit longer but that data is coming.

 

[zlobby]

Not tested on AMD platform. Why?

If i;m buying a cooler for $180 i'm looking for a more balanced noise and performance profile. Isn't this the popular choice amongst most people? Not sure why these manufacturers can't offer options with more premium fans or no fans for less. Just bloody RGB RGB RGB

$180 should warrant a full copper build. Sadly, not the case here.

 

[crazyeyesreaper]

Not tested on AMD platform. Why?


$180 should warrant a full copper build. Sadly, not the case here.

No AMD build because when this test bench was originally built AMD didnt have nearly the market penetration and OC as it still is was shit making it a not useful test bench. However the latest batch of processors are far better in that regard. So as I literally posted just above your post, AMD + Intel updated test benches are coming along with noise normalized performance.

As of right now I have the largest overall data set of CPU coolers tested on a modern platform. Right now reviews sit at 66 coolers tested.

As of right now I run 4x tests 3 times at stock and OC thats 15 mins per test. That works out to 6 hrs of benchmarking per cooler.

New benchmarks for both AMD and Intel platforms will take 14 hours. Now I am carrying forward 15+ coolers to start that works out 210 Hours just benchmarking coolers for new platforms. If I carry over more coolers that means more testing. I feel 99% of users have no idea how much time it takes to acquire this much data. So yeah no I was not about to add a friggin AMD bench halfway through a cycle just for shits and giggles so I can retest 60+ coolers at 6 hrs each its just not gonna happen lol.

Also keep in mind I maintain a set ambient temperature, have photo work, writing, then of course if there are issues I have to work with companies etc. Even for something like coolers its not a super easy process. I check thermal paste spread, how well the mounting process works. Issues like Corsair's A500 not mounting properly due to capacitors on the mobo hitting the heatpipes etc. They all add more time / complexity etc.

If you want AMD results, I would suggest waiting for other reviews for the time being. Someone will undoubtedly test the cooler on AMD. For those wanting AMD testing from me well your just gonna have to wait patiently.

 

[Makaveli]

No AMD build because when this test bench was originally built AMD didnt have nearly the market penetration and OC as it still is was shit making it a not useful test bench. However the latest batch of processors are far better in that regard. So as I literally posted just above your post, AMD + Intel updated test benches are coming along with noise normalized performance.

As of right now I have the largest overall data set of CPU coolers tested on a modern platform. Right now reviews sit at 66 coolers tested.

As of right now I run 4x tests 3 times at stock and OC thats 15 mins per test. That works out to 6 hrs of benchmarking per cooler.

New benchmarks for both AMD and Intel platforms will take 14 hours. Now I am carrying forward 15+ coolers to start that works out 210 Hours just benchmarking coolers for new platforms. If I carry over more coolers that means more testing. I feel 99% of users have no idea how much time it takes to acquire this much data. So yeah no I was not about to add a friggin AMD bench halfway through a cycle just for shits and giggles so I can retest 60+ coolers at 6 hrs each its just not gonna happen lol.

Also keep in mind I maintain a set ambient temperature, have photo work, writing, then of course if there are issues I have to work with companies etc. Even for something like coolers its not a super easy process. I check thermal paste spread, how well the mounting process works. Issues like Corsair's A500 not mounting properly due to capacitors on the mobo hitting the heatpipes etc. They all add more time / complexity etc.

If you want AMD results, I would suggest waiting for other reviews for the time being. Someone will undoubtedly test the cooler on AMD. For those wanting AMD testing from me well your just gonna have to wait patiently.

The amount of work is alot and appreciated.

 

[krusha03]

No AMD build because when this test bench was originally built AMD didnt have nearly the market penetration and OC as it still is was shit making it a not useful test bench. However the latest batch of processors are far better in that regard. So as I literally posted just above your post, AMD + Intel updated test benches are coming along with noise normalized performance.

As of right now I have the largest overall data set of CPU coolers tested on a modern platform. Right now reviews sit at 66 coolers tested.

As of right now I run 4x tests 3 times at stock and OC thats 15 mins per test. That works out to 6 hrs of benchmarking per cooler.

New benchmarks for both AMD and Intel platforms will take 14 hours. Now I am carrying forward 15+ coolers to start that works out 210 Hours just benchmarking coolers for new platforms. If I carry over more coolers that means more testing. I feel 99% of users have no idea how much time it takes to acquire this much data. So yeah no I was not about to add a friggin AMD bench halfway through a cycle just for shits and giggles so I can retest 60+ coolers at 6 hrs each its just not gonna happen lol.

Also keep in mind I maintain a set ambient temperature, have photo work, writing, then of course if there are issues I have to work with companies etc. Even for something like coolers its not a super easy process. I check thermal paste spread, how well the mounting process works. Issues like Corsair's A500 not mounting properly due to capacitors on the mobo hitting the heatpipes etc. They all add more time / complexity etc.

If you want AMD results, I would suggest waiting for other reviews for the time being. Someone will undoubtedly test the cooler on AMD. For those wanting AMD testing from me well your just gonna have to wait patiently.

Appreciate the hard work and looking forward to the new test benches

Btw as others have mentioned, any chance you gonna test the arctic liquid freezer 280 or 360?

 

[crazyeyesreaper]

Appreciate the hard work and looking forward to the new test benches

Btw as others have mentioned, any chance you gonna test the arctic liquid freezer 280 or 360?

As for Arctic I just havent had a chance I have been swamped with system reviews / building / acquiring parts for new test benches. I think right now I have 3-5 system reviews and 3-4 coolers to test as of right now. Hopefully once I swap to new test benches I can reach out to Arctic.

 

[Chrispy_]

The price keeps going up but nothing else really changes apart from more gimmicky RGBLED and the rats nest of cables that also mandates.

Is this Asetek rebranded or CoolIT rebranded? I've lost track of which ones are which now and they're all about the same anyway.

Does anyone even care any more? The only innovations are those sold outside the USA where the stupid patent-trolls Asetek can't have their monopoly. Global firms like Corsair are forced to tirelessly churn out Asetek over and over again until everyone is bored stupid and sick of it.