Snell Acoustics XA Reference Tower loudspeaker Measurements
The sheer bulk of the Snell XA Reference—a box more than 6' tall weighing 210 lbs—proved to be an obstacle in obtaining a complete set of measurements. I could not lift it off the ground for the frequency-response measurements, or even onto my Outline speaker turntable to do a set of horizontal off-axis measurements. So bear those omissions in mind as you read on. Note also that I measured it only with the long reflex ports, as used by Larry Greenhill in his auditioning.
With all the tone-control switches set to their central positions, the XA Reference's sensitivity was 88.5dB(B)/2.83V/m, a little higher than specified. Its impedance depended on the switch settings, and was not, surprisingly, lowest overall when the switches were set to their maximum positions and with the rear tweeter at its maximum level (fig.1). Even so, the impedance dropped below 4 ohms only in the midrange and above 10kHz. In itself, this shouldn't lead to any drive difficulty for the partnering amplifier. However, you can see from fig.1 that the electrical phase angle becomes quite extreme in the middle bass region, reaching a maximum value of -72 degrees at 47Hz. The impedance magnitude at this frequency is just above 9 ohms, which will ameliorate the effects of the phase angle. However, half an octave higher in frequency, at 70Hz, the combination of 4.7 ohms magnitude and 52 degrees capacitive phase angle will mean the speaker needs to be used with a good, beefy power amplifier.
Fig.1 Snell XA Reference, all switches set to "+2dB" and rear tweeter on, electrical impedance (solid) and phase (dashed). (2 ohms/vertical div.)
Fig.2 shows a similar plot taken with the switches set to their minimum positions and with the rear tweeter turned off. The impedance magnitude now hardly drops below 5 ohms and the demanding combination of phase and magnitude in the midbass has improved somewhat, to 5.3 ohms and -41.5 degrees. Even so, I wouldn't suggest driving this big speaker with anything other than a true high-end amplifier.
Fig.2 Snell XA Reference, all switches set to "-2dB" and rear tweeter off, electrical impedance (solid) and phase (dashed). (2 ohms/vertical div.)
The traces in figs.1 and 2 have a couple of wrinkles that indicate the presence of resonances of various kinds. The one at 26.3kHz will be due to the tweeter's primary dome resonance, but those at 190Hz, 290Hz, and 530Hz might be associated with cabinet problems. Nevertheless, the XA Reference's enclosure felt solid as a rock to the "knuckle-rap" test, and the output of a simple plastic-tape accelerometer fastened to the various surfaces revealed very few vibrations. Fig.3, for example, is a cumulative spectral-decay or "waterfall" plot calculated from the output of the accelerometer when it was fastened to the center of a side panel about 12" from the base. A very-low-level mode can be seen at 290Hz, but the plot is otherwise clean. Snell's use of damping and bracing in the construction of the XA Reference's cabinet is obviously very effective.
Fig.3 Snell XA Reference, cumulative spectral-decay plot of accelerometer output fastened to side panel 12" from base. (MLS driving voltage to speaker, 7.55V; measurement bandwidth, 2kHz.)
The saddle at 23Hz in the magnitude traces of figs.1 and 2 indicates the tuning frequency of the long ports and coincides with the notch in the four woofers' outputs, shown to the left of fig.4. The output of the twin ports is the bandpass curve centered on the region between 18Hz and 50Hz, and suggests excellent low-frequency extension, as indeed LG noted. The crossover between the woofers and the midrange units seems from this graph to be set approximately between 200Hz and 300Hz, with smooth out-of-band rolloffs evident, these featuring symmetrical 18dB/octave slopes.
Fig.4 Snell XA Reference, all switches set to "0," anechoic response on tweeter axis at 50", averaged across 30 degrees horizontal window and corrected for microphone response, with the nearfield midrange, woofer, and port responses plotted below 500Hz, 750Hz, and 250Hz, respectively.
To the right of fig.4 is the Snell's farfield response, averaged across a 30 degrees horizontal window on the tweeter axis and taken with all the switches set to their central, "0" positions. This is flat overall, but a very slight energy excess can be seen in the upper midrange, and equally slight depressions are evident at 4kHz and in the top octave. The tweeter's dome resonance gives rise to a 20dB peak at 26.3kHz. While this should not have any subjective consequences with signals limited to the audioband, such as from CDs, the resonance might be excited, with unpredictable subjective results, by the wider-band signals from SACD and DVA-Audio, or by vinyl playback using a phono cartridge with an ultrasonic but undamped tip-mass resonance in the same region.
As I couldn't lift the speaker off the floor for my acoustic measurements, I wasn't able to look at the action of the lower-frequency tone-control switches. (The reflection from the floor will obscure the response curve in the midrange, where these switches have their effect.) However, I could look at the upper-frequency switches, set to their boost and cut positions (fig.5). In each case, the curves are normalized to the on-axis response with the switches set to "0." The red traces show the effect of the Upper Treble switch. Basically, it shelves the response up 1dB or down 2dB above a nominal hinge frequency of 5kHz, the former being just about right to compensate for the tweeter's reduced top-octave dispersion in a large room. The blue traces are the Lower Treble switch, giving a useful ±1dB of boost/cut centered on the presence region.
Fig.5 Snell XA Reference, effects of Upper Treble switch (red) and Lower Treble switch (blue) set to their maximum and minimum positions (1dB/vertical div.).
As I said earlier, the XA Reference's bulk made it impractical for me to examine its lateral dispersion in detail. However, it did look well-balanced off-axis (not shown). And in the vertical plane (fig.6), as Snell's David Smith predicted and LG found, the XA Reference does offer superbly wide dispersion for a design with an array of widely spaced drive-units. The tweeter-axis response is maintained over a wide vertical window—just as well, considering that the tweeter is a high 42.5" from the floor.
Fig.6 Snell XA Reference, vertical response family at 50", from back to front: differences in response 20 degrees-5 degrees above tweeter axis, reference response, differences in response 5 degrees-20 degrees below tweeter axis.
In the time domain, the Snell's step response (fig.7) indicates that the tweeter and the midrange units are connected with the same, negative acoustic polarity. Although it is obscured in this graph by the inevitable reflection from the floor (the glitch just after the 7ms mark), the slow rise of the woofers' output begins just after the return to the time axis of the midrange units' output step, and is positive-going. Other than the very-high-amplitude ridge of delayed energy just above 26kHz, due to the tweeter's dome resonance, the XA Reference's waterfall plot (fig.8) shows a very clean decay throughout the midrange and treble, something that always correlates with a smooth, grain-free presentation.
Fig.7 Snell XA Reference, step response on tweeter axis at 50" (5ms time window, 30kHz bandwidth).
Fig.8 Snell XA Reference, cumulative spectral-decay plot at 50" (0.15ms risetime).
It used to be said that "big speakers have equally big faults." But as far as its measured performance is concerned, Snell's big XA Reference appears to be almost beyond reproach.—John Atkinson