Hi doyll
At first, let me explain what I wanted to achieve with these tests. In a good % of cases, when I read a fan or cooler review, I ask myself a lot of questions, for example:
- what is the thermal load procedure ?
- how is the thermal load measurement ?
- is the system, and the cooler, right for the purpose ?
- how loud is this fan at such RPM ?
- thermal paste ?
- will I find the ''ticking'' issue in the fan motor ?
- is the thermal analysis platform the same as the other tests ?
- what sensor it's used ?
- what's the error and the tolerance of the instruments ?
In short, I wanted to create a standard platform, with a very difficult heatsink that could have shown little discrepancies from different fans, even at very low RPM. I've set up a dedicated 1366 platform, with a Thermalright True Copper cooler, that allows me to mount fans with 120/140 and 150mm diameter. This is a very special cooler, because of it's restrictive (so the worst case scenario), and because of it's made of copper, so it can really show what a fan is capable of. These are the answers to my old questions, applied to my reviews:
- thermal load procedure: constant, re-checked every time, same protocol and testing equipment. I use a delta method, to 22.3 C° celsius (the original temperature that I used in the initial tests. I've verified that the delta analysis it's equal, without errors)
- I've decided to use coretemp and take the HIGHEST temp per processor, with 30/45mins of Prime95 "InPlaceLargeFFTs" (the better way to make thermal analysis with software)
- it's special, studied for the fans and very effective due the results. I'm very satisfied of the result
- I've decided to integrate the dBA measurements right in the Thermal load results bar, so you can have an idea of the noise, with the temperatures. The idle measurement it's not interesting, and useless (it even takes time, and effort, to add)
- thermal paste, I use a MX-4 with a standard application procedure. Very thin layer, so maximum performance. I want to underline a thing: I always re-check and re-seat the heatsink, to be sure that the pression of the mounting mechanism it's the same
- I explain every time, in the text, how is the RPM regulation of the interested fan. Due the absence of a frequency spectrum analysis (I have this possibility but unfortunately I haven't got a anecoic chamber) it's not possible to achieve
- yes, it is. What you see it's made on a standard platform, that I won't change until it fails
- FC5V2 thermal sensor, so it's a standard. i've also made some measurements about the delta with different models of fan controllers. If I find the link I'll show you.
- the one of FC5V2 sensor, which it's one of the best, if not the best, fan controllers that your money can buy. I use a degree of error, to approximate the 1°C differencies due the methodology.
- platform and settings, I've got a standard platform with standard settings
page 3/24:
http://www.xtremehardware.com/cooling/air/xtremehardware-presenta-iv-roundup-ventole-120-140mm-giugno-2013-201306178781/?start=2
Here you can find a brief description of the dBA testing procedure, how to and what needs to be avoided.
page 20/24:
http://www.xtremehardware.com/cooling/air/xtremehardware-presenta-iv-roundup-ventole-120-140mm-giugno-2013-201306178781/?start=19
Here you can find some other info related to methodology
Feel free to share, comment or even suggest things. I'm open minded about it