When I stop to think about watercooling for a bit, I see two basic performance reasons why I like it . One reason is the performance of it, I can overclock higher than with air. The second and probably more critical to me is the performance per noise level. Perhaps I’m getting more noise critical in my old age, but listening to a vacuum cleaner in the background is not at all acceptable or what water cooling is really about.
I can test performance and log RPMs, but that’s really not apples to apples either. Some fans have stronger PQ curves at like RPM levels and there is also a large variety of noise quality differences. How a fan interacts with neighboring fans can be both good and bad and box specs simply do a poor job at measuring noise levels in an actual case/radiator condition. IMHO, a review of performance without a measurement of noise is no different than testing one radiator with 1800RPM and the other with 2700RPM and calling it a review. You can compare like RPM levels as a quick and dirty test, but that too is generally a poor measure since fan performance can vary by as much as 400RPM producing the same air flow. To do a review and comparison correctly, you really need to measure and compare noise vs dT or noise vs Core temp or some sort of constant. That’s what I’m attempting to do with this first noise testing phase. I am also retesting in a top case mount condition so all kits get a equal and fair condition.
I started this kit testing using the bottom mounting location on my Switch 810 however I encountered a problem when trying to install the H100i….the hoses are about 2″ too short. I’ve done enough testing to know you can NOT change the condition, no matter how small between tests. A simple thing like placing the case on carpet vs a hard surface is enough to throw results out by 2C, same goes for noise measurements.
So in an attempt to get a true apples to apples noise vs DeltaT, I first need to compare noise vs RPM and then in Delta T testing I can convert the logged RPM average to noise level.
Unfortunately, Corsair Link is not compatible with Windows 8, so my only means to control the unit at all was to use the PWM fans and control speed via speed fan. Corsair Link simply does not monitor or control anything H100i related if you are running windows 8 as of this testing which was a bit of a surprise to me. I believe Win8 has been out since last October so a good 5 months later, still no Win8 support. Oh well, it really didn’t bother me since the new kit fans do have 4pin PWM control. At least I could control fan speeds using the motherboard and speed fan, I just needed to use a PWM splitter.
Corsair H100i VIDEO
Swiftech H220 VIDEO
That is what I have so far, I will let your ears do the talking… As a place holder you may start and stop both videos to match up RPMs if you would like and simply play one, stop, then play the other back to back.
Corsair H100i Chart
Swiftech H220 Chart
For reference use this to compare to similar daily noise types:
|Sound sources (noise)
Examples with distance
| Sound pressure
Level Lp dB SPL
| Jet aircraft, 50 m away
| Threshold of pain
| Threshold of discomfort
| Chainsaw, 1 m distance
| Disco, 1 m from speaker
| Diesel truck, 10 m away
| Kerbside of busy road, 5 m
| Vacuum cleaner, distance 1 m
| Conversational speech, 1 m
| Average home
| Quiet library
| Quiet bedroom at night
| Background in TV studio
| Rustling leaves in the distance
| Hearing threshold
Using the following chart for perceived noise level differences:
|Perceptions of Increases in Decibel Level
| Barely Perceptible Change
|Clearly Noticeable Change
|About Twice as Loud
|About Four Times as Loud
Top Mount Noise vs Core Comparison
While I only have the H100i for comparison at this point, this is how that combination of noise testing and thermal testing lays out. This is the meat and potatoes results as it pulls out all the variables such as fan performance vs RPM differences and compares a direct apples to apples noise vs performance.
And that plots out like this:
I’ve conducted a few different fan polls in the past before moving forward with radiator testing and typically the majority of watercooling users out there were using fans at about 1350RPM with a distribution that predominantly covers the 1000-1800RPM range. There are a few people that run higher speeds, but the vast majority are more in the 1000-1800RPM range. That makes sense as those speeds are only producing noise levels in the lower 40s dBA which is acceptable by most.
With that “Silence” preference in mind, the Swiftech H220 did outperform the Corsair H100i by roughly 2dBA at the 1350RPM mark, and about 3-4dBA better at the 1800RPM mark. The H100i does have higher speed capabilities, but at 60dBA you are approaching vacuum cleaner noise levels (70dBA). There are some users that don’t care about noise though, so I’ll leave it to you on what’s important.
For me personally, minimum noise level, dynamically throttling the noise and performance, and a smooth noise is priority. I find the H100i fans (like most high speed fans) to produce quite a bit of motor noise and there was quite a bit of harmonics between the two fan RPMs that generally produces a poor noise quality. The H220 helix fans are for the most part fairly smooth and the sound quality blended in well with my other fans. That’s not to say you couldn’t replace the kit fans with something better, but in spirit of testing kit vs kit, the H220 is king of silence (minimum noise level) and better at like performance levels by 1-4 dBA over the H100i.