Calix Phoenix Grand Signature loudspeaker Measurements
The phone call from the shipping company was the harbinger of much sweaty work on my part: "Sorry, the local trucking company can't deliver the Calix speakers today. The four crates weight 1300 lbs and they don't have a forklift available." Neither did I! But once a forklift had been found and the crates stowed in my garage, I was impressed by the way the Phoenix's packaging had been designed for easy handling—always an issue with speakers this complex and this massive.
For obvious reasons, it wasn't possible to raise the Calix Phoenix Grand Signature off the floor for the measurements. The farfield acoustic measurements are therefore affected by an early reflection from the floor; while it was possible to window this reflection out of the MLSSA time window in order to calculate the frequency responses, this does degrade midrange resolution.
The speaker's voltage sensitivity was surprisingly low for a horn-loaded design. My estimate was 85.5dB(B)/2.83V/m, which is 1.5dB lower than the average of the 500 or so speakers I have measured over the past 13 years and 2.5dB lower than specified. However, the plot of impedance magnitude and phase against frequency (fig.1) reveals the Phoenix Grand to be an easy load for the partnering amplifier to drive. The impedance remains above 6 ohms almost the entire time, with a benign phase angle.
Fig.1 Calix Phoenix Grand, electrical impedance (solid) and phase (dashed). (2 ohms/vertical div.)
The plot is also free from the wrinkles and discontinuities that would indicate the presence of cabinet resonances. Fig.2 is a cumulative spectral-decay plot calculated from the output of a plastic-tape accelerometer fastened to the front baffle adjacent to the upper-bass drive-unit. A small number of resonant modes can be seen, the strongest of which lies at 254Hz. However, given the fact that these modes are all low in level and that the radiating area of the baffle is small, their subjective effect should be minimal. Repeating the measurement on the massive subwoofer box revealed a low-level, high-Q mode at 275Hz present on all surfaces (not shown), but this should be inconsequential.
Fig.2 Calix Phoenix Grand, cumulative spectral-decay plot calculated from the output of an accelerometer fastened to the front baffle level with the woofer. (MLS driving voltage to speaker, 7.55V; measurement bandwidth, 2kHz.)
The saddle at 21Hz in the fig.1 magnitude trace indicates the tuning frequency of the large rectangular slot at the front of the subwoofer box, while that at 56Hz reveals the tuning of the port mounted at the back of the upper-bass enclosure. The traces to the left of fig.3 show that while the subwoofer's minimum-motion point does coincide with the impedance minimum, the tuning of the port is complex, its output featuring two humps rather than the one typical of a reflex alignment. The subwoofer driver itself rolls off rapidly above 60Hz, handing over to the upper-bass driver-port combination from the midbass up. The trace in fig.3 representing the latter is the complex sum of both outputs, taking phase and physical separation into account; it appears to peak a little before beginning its 24dB/octave high-pass rollout.
Higher in frequency, the upper-bass driver (black trace) extends surprisingly high in frequency, handing over to the tweeter at around 3kHz. But, of course, there is a fourth drive-unit to be considered, the horn-loaded midrange dome, the response of which is shown in red in fig.3. The horn unit can be seen basically to act as a "fill-in" driver, reinforcing the upper-midrange region where the upper-bass unit's output is suppressed. All things being equal, this should result in an even distribution of energy across the entire audioband.
Fig.3 Calix Phoenix Grand, acoustic crossover on tweeter axis at 50", corrected for microphone response, of woofer-tweeter module (black) and horn midrange (red), with the nearfield subwoofer and port responses plotted below 1kHz and 625Hz, respectively, and the complex sum of the upper woofer and port responses plotted below 450Hz (black).
However, all things are not equal, as the horn unit's acoustic center is more than a foot behind that of the upper woofer and tweeter. The horn was aimed precisely at the microphone for the measurements, using the precision screw on the horn support pillar. Even so, the resultant time delay, coupled with the overlap between the woofer and the horn over almost all the latter's passband, results in acoustic interference that manifests itself as severe comb-filtering in the farfield response (fig.4). (The suckouts due to this interference explain the discrepancy between the specified sensitivity and what I measured.)
Fig.4 Calix Phoenix Grand, anechoic response on tweeter axis at 50", averaged across 30 degrees horizontal window and corrected for microphone response, with the complex sum of the nearfield upper woofer and port responses plotted below 400Hz and the complex sum of the nearfield subwoofer and port responses plotted below 1kHz.
Putting this to one side for a moment, the treble is otherwise evenly balanced, the large suckout at 9.5kHz disappearing on either side of the central tweeter axis. The upper bass is plateau'd a little on the high side before crossing over to the subwoofer, but this will be due in part to the nearfield measurement technique, which assumes a 2pi (hemispherical) acoustic environment. The subwoofer also peaks very sharply, before rolling out quite steeply below 40Hz. However, in a typical room, reinforcement from the boundaries will result in extension to 20Hz. There are a couple of peaks apparent in the subwoofer's midrange output; I could hear these on pink noise when the subwoofer was driven by itself, but they were inaudible when the main speaker was connected.
Returning to the comb-filtering apparent in fig.4: Because they depend on the times of arrival at the listener's ears, the exact frequencies of the interference dips will be very dependent on the listener position. This can be seen in fig.5, which show how the Phoenix's response changes with measurement axis. The pattern of dips and peaks is different at each microphone position; I suspect that the aiming of the horn arranges for the comb filtering to "fall between the gaps" in Western music tuning. Laterally, the response didn't change much over the 15 degrees angle to either side of the tweeter axis that I was able to measure, other than the filling-in of the tweeter suckout at 9.5kHz mentioned earlier. But the quite narrow directivity of the horn's output will not help with the woofer's beaming at the top of its passband. (The rear tweeter, by the way, is connected with the same positive polarity as the front tweeter and covers the three octaves above 3kHz.)
Fig.5 Calix Phoenix Grand, vertical response family at 50", from back to front: responses 20 degrees-5 degrees above tweeter axis, response on tweeter axis 42" from floor, response 5 degrees below tweeter axis.
In the time domain, the Phoenix's impulse response (not shown) is nothing unusual, other than a second arrival a millisecond after the tweeter. As mentioned earlier and as revealed by the step responses in fig.6, this late arrival is the horn's output (red trace), connected, as it appears, in inverted acoustic polarity to the tweeter and woofer (blue). As a result, the overall step response (black) is significantly disturbed, resulting in the comb-filtering mentioned earlier. The speaker is certainly not time-coherent.
Fig.6 Calix Phoenix Grand, step response on tweeter axis at 50" (black), and of tweeter-woofer module (blue) and midrange horn (red). (5ms time window, 30kHz bandwidth.)
The double arrival complicates the interpretation of the Calix's cumulative spectral-decay plot (fig.7). It is perhaps more meaningful, therefore, to look at the individual waterfall plots of the woofer-tweeter array (fig.8) and of the horn (fig.9). The former is still disturbed by some hash in the treble, presumably due to breakup modes in the woofer's cone. (Note that this graph extends to only just over 2ms, due to the aggressive windowing I had use to to eliminate a tweeter reflection from the horn.) The horn's output does have a sharply defined ridge of delayed energy present at 8.5kHz, but this is well-suppressed by the crossover.
Fig.7 Calix Phoenix Grand, cumulative spectral-decay plot at 50" (0.15ms risetime).
Fig.8 Calix Phoenix Grand, tweeter-woofer module, cumulative spectral-decay plot at 50" (0.15ms risetime).
Fig.9 Calix Phoenix Grand, horn midrange, cumulative spectral-decay plot at 50" (0.15ms risetime).
Summing up these measurements is difficult, as the the comb filtering's degradation of the speaker's tonal balance will be ameliorated in-room by the integrating effect of the reverberant field. Certainly, when I auditioned the Calix Phoenix Grands in Paul Bolin's room, I heard nothing untoward. And I was impressed both by the speaker's presentation of recorded detail and its powerful if rather muddy-sounding low frequencies. (See PB's explanation in his auditioning comments of how he fixed the latter.) But overall, the Calix Phoenix Grand Signature, with its low sensitivity, wide overlap between disparate drive-units, and comb-filtered on-axis response left me scratching my head.—John Atkinson