While mentioned occasionally on some cooler reviews about some manufacturers brute forcing additional cooling performance via significantly faster fans leading to better cooling numbers compared to some other more efficient coolers. It raises the question, what if all of he coolers had that option?

While most people would not want their PC to be very noisy all of the time, For quite a long time now, modern systems have had pretty detailed control over custom fan curves. This means that it would not be hard to take a 3000 RPM 140mm fan, and have it gradually increase fan speed to 1500 or probably 1700RPM as the CPU approaches 80C, and then have the remainder of the curve from 81-90C raise the fan speed linearly to 3000RPM, thus allowing for quiet performance for most common workloads, but allowing it to get noisy when you hit every core with an AVX workload when you are exporting a video.
Most people will not mind extra noise while their CPU spends a few minutes at its max power consumption.

I’ll add that change on the next test.

Enjoyed the article Kent. The third and fourth charts could do with an extra row at the top to make it obvious which one was ‘noise normalised’ and ‘100% fan speed’. I like the fact that I can estimate temps for my 5950x at eco, normal and PBO etc at a glance.

The first statement should have read “began to notice”. You are correct on the second. I believe they were in the word document and just got lost somewhere along the way, but perhaps I accidentally edited them out in my final draft. I’ll have to check later.

Other than the editting issues, what is your opinion on the new testing? Ì’d like to hear some feedback.

a little further down “One other note is that I would like to address the point 42 dB may seem high” – did you mean to say “One other note I would like to address is that 42 dB may seem high” ?