Celestion SL700 loudspeaker 1988 Measurements

Sidebar 2: 1988 Measurements

The frequency response of the Celestion was measured in the listening window—spatially averaged to minimize room standing-wave problems—using a 1/3-octave warble-tone generator; in addition, the nearfield low-frequency response of each speaker was measured with a sinewave sweep to get an idea of the true bass extension relative to the level at 100Hz. The change of impedance with frequency and the voltage sensitivity (using 1/3-octave pink noise centered on 1kHz) were also measured.

Fig.1 shows the manner in which the SL700's impedance varies with frequency. [This measurement was made before Stereophile acquired its Audio Precision System One measurement system, so doesn't show phase.—Ed.] The box resonance can be seen at 62Hz, with a minimum of magnitude 3.7 ohms above it at 135Hz. Comparing it with the impedance plot for the SL600 (fig.2) reveals that, unlike the earlier speaker, it does not have the "trap" or notch filter to compensate for the metal-dome tweeter's "oil-can" resonant mode, when the middle of the dome is going backward while the outer edges of the dome are going forward. This can be seen very clearly on the SL600's plot (the very sharp impedance peak at 22kHz). Despite the dips to 4 ohms in the bass, the SL700 should be a much easier speaker to drive than the SL600, even though its measured sensitivity is identical at 82dB/W/m.

Fig.1 Celestion SL700, electrical impedance magnitude (2 ohms/vertical div.).

Fig.2 Celestion SL600, electrical impedance magnitude (solid) and phase (dashed) (2 ohms/vertical div.).

Low-frequency extension, measured in the nearfield, was identical to that of the Acoustic Energy AE1 at 42Hz, -6dB. However, a glance at fig.3, the SL700's spatially averaged room response, shows that, unlike the reflex AE1, the sealed-box Celestion rolls out much more gradually in the bass, there being useful output in-room down to just below 40Hz. (Even the 32Hz 1/3-octave band is down by just 8dB in-room, compared with the level at 1kHz.)

Fig.3 Celestion SL700, spatially averaged, 1/3-octave response in JA's room.

Overall, the SL700 room response is smooth, and it is instructive to compare it with that of the SL600 (fig.4), which was taken under identical conditions in the same room with the speakers in identical positions. Low-bass extension is virtually the same (which is what you would expect from an identically dimensioned enclosure with an almost identical drive-unit). The '600's upper bass is a little less well-controlled, however, while the midrange is slightly less even. The main difference lies in the treble, where the more sensitive aluminum-dome tweeter of the '700 gives 2-3dB more output. (The shape of the curve in the treble will also be affected by the tweeter's wide dispersion in the low treble.) As I performed the series of tests individually on each speaker, I got an estimate of the closeness of the response matching between the pair. In this case, the pair-matching was about the best I have seen, there generally being less than 1dB difference in-room between the two from the lower midrange upward. This care taken in quality control should be reflected in superbly precise lateral stereo imaging.

Fig.4 Celestion SL600, spatially averaged, 1/3-octave response in JA's room.

Though not shown in fig.5, the above-axis measurements indicated that the SL700 should not be listened to above the tweeter axis, there then being a slight suckout at the top of the woofer's passband, and a peak developing above crossover. The moral: use the supplied stands, which place the listener's ears at tweeter level.—John Atkinson

89 Doug Brown Way
Holliston, MA 01746
(732) 683-2356