So fan swaps are something a lot of people have written about. There's warnings about CFM and inaccurate descriptions of what static pressure measurements for a fan mean and the like. I decided to do the only sensible thing, ignore everyone, buy a bunch of fans, swap them into a collection of O2 PSUs and test things myself.
Overall notes:
I'll start by saying this is in no way properly scientific but may be of interest!
This was done on my 400mhz R12k O2 w/ a full 1gig of ram for max heat, and a Fujitsu MBA3300NC 15k rpm drive.
The noise readings were just taken with a crappy iPhone app, but still show relative changes.
dBA reading with system off was between 41.8-42.2 dBA
The O2 sits on a desk about 2.5ft from me / where the samples were taken
I took these temp measurements in series, then went back and re-took them again to account for changes in ambient temperature but they were within 1 degree. I used the lowest reading for each. Temperatures were measured taking the drive temperature using scsimon - not total system temperatures or IR readings or the like. Also, to prove how unscientific this all was, I actually used 3 different PSUs rather than swapping the fans between a single unit. This allowed me to swap more quickly but means differences in the PSUs themselves may have some impact.
Testing measure was:
Power on system, let it run for 30 minutes to get up to temp. These tests were done with IRC running (to keep from boredom)
and I'd hammer the drive with a couple `dd if=/dev/random of=testfile bs=512 count=1000000` to ensure something was happening on the drive since that's where the temp was coming from. I would then shutdown the system swap to another PSU and immediately fire it back up to keep things warm.
PHP Code:
fan: stock
52c/125f
49.6-50.2 dBA
fan: ebm-papst 8412n
54c/129f
48.7-49.3 dBA
notes: slightly quieter than stock, but enough to be noticeable
fan: cooler master blade master 80
55c/131f
46.6-47.2 dBA
notes: substantially quieter than stock and noticeably quieter than the papst
fan: noctua nf-a8 flx
56c/133f
notes: this measurement was taken on a different day and ambient temperature may have been notably different
temperature was continuing to rise, may have gotten hotter when I shut the system down
I no longer have this fan in a PSU to take dBA measurements, however it was effectively silent
