Avalon Eclipse loudspeaker Measurements part 2

Looking at the manner in which the fig.6 impulse response decays gives the "waterfall" plot in fig.9. With the exception of some nonsense between 1 and 2kHz, the decay is exceptionally clean in the treble, which undoubtedly contributes to the speaker's transparency and impressive sense of image space. Repeating the measurement for the woofer alone with 15Hz resolution, however, shows that the woofer does contribute some resonant hash in the midrange (fig.10), perhaps due to the cabinet, which might contribute to the feeling of hardness I felt the speaker's sound to acquire at high levels.

Fig.9 Avalon Eclipse, cumulative spectral-decay plot at 55" (0.15ms risetime).

Fig.10 Avalon Eclipse woofer, nearfield cumulative spectral-decay plot (note 15ms time axis).

I also felt it worth looking at the Eclipse's distortion performance at the lower end of the tweeter's passband. The impedance plot (fig.2) implies a low crossover frequency for the 1" dome unit. A representative of MB, the unit's manufacturer, recently mentioned to me that using this tweeter to lower than 2kHz or so was "unwise." In the main, the tweeter appeared to be a low-distortion device. Even at a high 93dB level (measured at 1m), for example, the total distortion imposed on a 2kHz tone by the tweeter lay at a low 0.3%. And two octaves lower, the distortion added by the woofer to a 499Hz tone at the same spl was very low, at around 0.1%. But as can be seen from the spectrum of distortion products in fig.11, changing the frequency to 1kHz and driving the tweeter at this level results in a more than 10-fold increase in distortion. The second harmonic measures just 37dB down, or 1.4%, though the third remains 54dB down at 0.2%. These figures are more typical of a two-way minimonitor.

Fig.11 Avalon Eclipse tweeter, distortion spectrum for 1kHz at 93dB at 1m (linear frequency scale).

The woofer will be making about the same contribution to the sound at 1kHz, so I also looked at its distortion spectrum at the identical drive level (around 7V RMS). Though the second harmonic is below -60dB, 0.1%, the third is on the edge of audibility at -47dB, or 0.45%. Looking at the overall distortion at 1kHz with both drive-units driven at this level (fig.12), it can be seen that the second harmonic from the tweeter and the third from the woofer are the major distortion products, both appearing at levels which will be audible. Dropping the spl to a moderate 86dB results in the tweeter's second harmonic reducing to a more-reasonable -43dB, or 0.7%.

Fig.12 Avalon Eclipse full-range, distortion spectrum for 1kHz at 96dB at 1m (linear frequency scale).

As neither the Classic 60 nor the VTL 160 amplifiers were near clipping at this level, I have to conclude from these measurements that these relatively high levels of distortion from signals around the Eclipse's crossover region, coupled with its treble-forward in-room balance, probably contribute to the hardness and stridency I noted at high playback levels, particularly as the distortion products fall in the exact region where the ear is most sensitive.—John Atkinson