Apollo
GATTI AUDIO

If you spend too much time thinking about a thing

You''ll never get it done

- Bruce Lee

Measurements & crossover design

 

To guarantee the most accurate results, farfield measurements were taken outdoors. The speaker is elevated on the edge of my verandah 60cm from the ground, and is 1.5m from the wall behind it.

The microphone was placed at a distance of 270cm (9 feet), 150 cm vertically off the ground (90cm up from the base of the speaker), and 10 degrees off-axis. In other words, the microphone is placed at the same listening distance, height and angle relative to the speaker as the listener would be positioned in my listening room.
Software is JustMLS (part of LSPCAD), and the cailbrated Panasonic microphone feeds into a Bamberg Labs preamplifier driving a Dick Smith integrated amplifer, and all is controlled by a Sony Vaio laptop using Windows XP.

 

Measurements

The farfield measurements are merged with nearfield measurements and modelled bass performance. My Lspcad baffle/room simulations indicate that this is a good approximation.

Minimum phase is approximated using LSPCAD's active filter design and juxtaposed the phase onto the measured frequency response plots.

In the impedance measurements, a glitch occurs at 50Hz as a result of mains interference.

 

Seas T29MF001 E0047 Magnum
Scanspeak 12MU8731T00
Scanspeak 18WU8741T00
Note : The above was 1.5dB higher than specs. i.e 87dB sensitivty.
Morel UW1258

Acoustic centre offsets are determined by measuring combined driver outputs and matching the result with the summed minimum-phase responses. The diagram below as an example, shows the good correlation between the measured combination of the 12MU & Magnum, compared to the individual driver responses with simulated minimum phase and acoustic-centre (dz) offset. My initial estimates of acoustic-centres (based on the point where the voice-coil meets the diaphragm) turned out to be fairly accurate (dz=0). Ignore data below 500Hz.

Overlay of driver responses
Crossover

Crossover points occur at approximately 80Hz, 400Hz, and 4kHz. Every driver's passband is within it's pistonic region.

Note the flat phase response and -3dB point of 30Hz. Ripples around 600Hz correlate with the 52cm baffle width. Bumps around 5.5kHz

correlate with the baffle step located 6cm from the tweeter.

Modelled impedance Modelled group delay
The impedance (above) is a benign 4-5 ohm load. It should be easily handled by any reasonable amplifier.
Crossover version 1.0 (July 2010)
Woofer UW1258
Note : 1st order at 60Hz + 5 dB gain applied
Mid-woofer 18WU8741
Midrange 12MU8731
Tweeter T29MF001 E0047 Magnum

Power handling

The Morel UW1258 can handle 200W before reaching it's excursion limit.

However, with 5dB gain it's power input is multiplied by 3, and it therefore reaches it's "system limit" at 200 / 3 = 65W.

 

The power handling figures in the diagram below assume a system sensitivity of 87dB/2.83V.

Tthe spectral content of music falls rapidly below 100Hz (typically 10dB less at 50Hz), so from the diagram below, the overall power handling of the

speaker can be estimated as 200W, (23dBW relative to 87dB), producing 110 dB SPL.

Low-frequency power handling of the 18WU can be increased to 100W by inserting a 1st order highpass filter at 60Hz, but this probably isn't

necessary.

Apollo design
Apollo construction
Gatti Audio