Aristoteles
Measurements

Aristoteles 0.61:

Conditions:

  • All measurements performed outdoors and ungated
  • Excitation: periodic noise.

Raw response of the lower mid-range driver.

Raw response of the upper mid-range driver. The driver was not counter-sunk and there is still some wood to be removed in order to get the best possible transfer function.


Aristoteles 0.5:

Conditions:

  • All measurements are ungated with visible floor bounces between 200Hz and 300Hz.
  • Excitation: periodic noise.

The new panel size and shape now rules out the mid-woofer completely as a contributor to blooming.
The absolute shape of the wave form should be disregarded as it was recorded with the now “incorrect” ASP 0.41 transfer function. However, the small hump around the yellow cursor is interesting...

Also here, the improvement of the polar response becomes clearly visible. Even the upper end, now starting around 6KHz, looks better compared to this picture recorded with version 0.4.

 

With the new panel, another small resonance becomes now clearly visible. It was there all the time (douh !) but it was masked in the on-axis response and I completely overlooked it because there were bigger issues. This now explains the strange behavior around 1.7KHz. It is a result of the dipole operation.
Equalization of the bump should further improve the over all polar response.


Aristoteles 0.4 (single driver measurements):

Conditions:

  • The polar response of the mid woofer was taken vertically on axis at a 1.5m distance. The setup allows for a ~7.2ms gate.
    Only the notches for the dipole peak and the membrane resonances were enabled. No HP, LP or SLP.
  • The polar response of the tweeter was taken vertically on axis at a 1.5m distance. The setup allows for a ~6.4ms gate.
    Only a LR4 HP with fg= 1350 Hz was enabled.
  • All polar measurements taken at 5° increments from 0°...+180°.
  • Although windowed, the measurements are not 100% reflection free due to some solid structures in the proximity.
  • Excitation: periodic noise.

 

Polar response of the Mid-woofer:

This is pretty much reassuring. There are only real irregularities at the back side over one octave above the anticipated cut off frequency. This is most likely caused by diffraction at the spider, the back fire tweeter mounting and a reflection at the top of the H-frame (around 1.5KHz). In addition the irregularities at the back may be due to the building of side lobes because the driver is operated beyond the dipole peak. This is not observed at the front because there the beaming of the driver counteracts the building of lobes.
In any case the driver should be operated below 1.6KHz in the given panel, while good dipole behavior is achieved up to
1 KHz.

Polar response of the tweeters:

At 800Hz the response shows nearly perfect dipole behavior
(-1dB at 30°, -3dB at 45°, -6dB at 60°). The wavelength is
43 cm and thus longer than the panel is wide. Front and rear wave can interact and the resulting sound power is at least close to 4.8dB, which is the typical value for an acoustical dipole.
Above 1KHz the response widens ("blooming") to dipole like behavior due to the relatively large panel size compared to the wavelength radiated. Starting around 3KHz, beaming of the tweeter counteracts the blooming and around 6.3KHz beaming and blooming are in balance. Further up in frequency the response narrows down as the beaming is progressing.
At 6.3KHz, the response is no real dipole response because at a wavelength of 5.5 cm the two tweeters are completely separated from each other by the panel. They have become single sound sources radiating into half space. As a result, the sound power rises by 3dB.


Aristoteles 0.3:

Notes:

  • Measurements taken outdoors as shown here.
  • Some information in the graphs is hardly legible. Arta is quite new to me but I know already how to fix that for the next time.
  • Partially cut off x-axis description due to copy and paste problems. Missing info is described in the text.
  • All graphs are 1/3 octave smoothed.
  • Mic aimed between tweeter and mid.
  • All measurements ungated.

 

Complete panel measured in 1m distance at 30°, 45°, 60°, 75° and 80°
No dipole cancellation correction in place (only notches).
The angles have been approximated, which needs a little more care the next time.
Good dipole behavior up to 1KHz. Then the response widens and finally the tweeters start to control the radiation pattern. The dip that starts out slightly above 1KHz and wanders down in frequency with increasing angle is a resonance of the rubber surround of the mid-woofer. Virtually every cone type driver exhibits such behavior to varying degrees.
The rising frequency response is not a “processing” artifact as stated previously (although that apparently exists). It was caused by applying an inverse function of the microphone frequency response (a compensation response). But Arta expects the normal response.

Equalization:

Last updated 27-Mar-2017

Raw response of the 18WU in panel rev. 0.3.
Dipole peak and membrane resonances are visible. The small dip at 240Hz is caused by a reflection from the floor.
X-axis like in picture below.

Dipole peak and membrane resonance equalized.

H-Frame resonance peak measured right at the opening.

Peak equalized.