sound analyse iMac G5 20"

Here some screenshots of sound records of my new iMacG5, 20",
recorded with SignalScope of Faber Acoustical.
[warning: sometimes the iMacG5 (sound drivers?) reacts buggy when opening and after quitting SignalScope.
When you get a sine-wave like curve in SoundScope after starting it, open "Input Settings" and change the "Sampling Rate" to 48000,
afterwards go back to the main window and click on "Pause" and afterwards on "Start".
When i.e. Quicktime is not running proberly after quitting SignalScope, to run a file in GarageBand might be a fix.
It is not sure whether it is a SignalScope or a driver/iMacG5 bug.
Seems to be solved with OS X 10.3.6]
To indentify and/or to get some insight about tone and frequency you can use a tool like AudioTest by Hidetomo Katsura.

disclaimer: the charts can only be read in comparison, not as an absolute output.
Environment, location, temperature, quality and location* of the microphone heavily influence the messurements.

*When putting your ear nearby the bottom right, where the microphone is located, you will hear a different sound than in front of the iMac.
High frequency noise will be less audible, recorded by the internal mic than it is.

For the first "tests" only SignalScope and "activity" running (appr. 30% processor capacity).
Fans running as fast or only a bit faster than in idle mode.

Table1: recorded with the internal mic

Recorded with the internal mic it looks like this. Vibrations of the iMac case (harddisc) seem to disturb the record a bit (120Hz peak).
You see a peak at appr. 1200Hz.

***

I got a second midplane assembly today.
It is much more silent than the original one.
Here once more some records with SignalScope.

Table 2: midplane2, recorded with the internal mic

The peaks are half as intense as with the first midplane.
The noise is still audible (peak at 1200Hz). Some ringing at appr. 2800Hz.

Some days later .

Table 3: midplane2, recorded with the internal mic

Table 4: midplane2, recorded with the internal mic (this time with dB at Vertical Scale)

Table 5: midplane2, recorded with the internal mic, 10 minutes several processor intense programs

***

changing sound experiences, 27.10.2004
(table 5 - 9)

Table 6: midplane2, recorded with the internal mic, 27.10.2004, after start, I got a new sound in the morning.

Table 7: midplane2, recorded with the internal mic, 27.10.2004, 5 minutes later

Table 8: midplane2, recorded with the internal mic, 27.10.2004, an hour later, running several processor intense programs

* unfortunately, a louder whine is back

Table 9: the same settings like in table 8, but the sound of a churchbell (on the left, at 250Hz) in the background, the church is 3 houses away (windows closed).

hypothesis:
the peak at 120Hz, vibrations of harddisc>case>microphone (not audible),
the peak at 1000Hz from the processor fan (when whining under load).

***

Table 10: recorded 02.11.2004, green line "idle", red line "load" (with the modifications listed below)

damping
experiments

power filter in front of the iMacG5 (cleaning the power source),
top chambers damped (a bit) with 3mm acoustic foam (cut and loosely inserted in the two chambers/air outlets) - Akasa PAX.mate, acoustic absorption mats (thickness 3mm),
the processor fan damped with a small "felt foot/ felt glider" (used for furnitures) on the top of the "fan box" (diameter 22mm, thickness 3-4mm) (reduces the whine, does not remove it).

Table 11: felt glider on the processor fan, Akasa PAX.mate in the fan outlets
the cables (black monitor cable in the middle, black SATA cable below the case fan, cables at the SATA drive) looped carefully around the fans.

Table 12: cut pieces of Akasa PAX.mate

IT IS NOT A FIX (the noise is only masked)

***

Table 13: recorded 06.11.2004, comparing noise with harddisc running and sleeping, green line "harddisc running", red line "harddisc sleeping" (modified, damped machine)
obviously the harddisc (vibration + noise) has no influence to the overall noise.

Table 14: recorded 09.11.2004, external Sony condensor microphone>MD recorder>line in, playing iTunes visuals for 5 minutes (without "felt glider" damping).

Table 15: recorded 13.11.2004, external Sony condensor microphone>MD recorder>line in, idle mode
red line - top right, green line - top middle

Table 16: recorded 13.11.2004, external Sony condensor microphone>MD recorder>line in, iTunes visuals for some minutes
red line - top right, green line - top middle

***

some material to acoustics:

the audible frequency range of the human ear

source: http://www.e-smog.ch/kurs_messtechnik/kapitel_vi/1/3.htm

* most sensitive range 500Hz to 5000Hz.
especially frequencies between 2000Hz and 5000Hz have an alert effect (baby cry).

source: http://www.uni-erfurt.de/sprachwissenschaft/personal/lehmann/Fundus/Hoerbereich.html

3dB => double volume

whispering - 30dB(A)
normal conversation - 50dB(A)
passing truck - 90dB(A)
airplane, jet - 125dB(A)

source: http://www.hitechnatur.ch/sinnen/2/einleit.html