Bowers & Wilkins 804 Diamond loudspeaker Measurements
Sidebar 3: Measurements
I used DRA Labs' MLSSA system and a calibrated DPA 4006 microphone to measure the Bowers & Wilkins 804 Diamond's frequency response in the farfield, and an Earthworks QTC-40 for the nearfield responses, the latter's ¼" capsule offering no significant acoustic obstacle to the outputs of the small diaphragms. Before performing any measurements, I removed the plastic disc from the rear of the tube that loads the midrange unit, as instructed in the manual.
My estimate of the 804 Diamond's voltage sensitivity was 89.3dB(B)/2.83V/m. While this is slightly lower than the specified 90dB, it is still usefully higher than average. The plot of the 804 Diamond's impedance magnitude and phase against frequency is shown in fig.1. The speaker is a moderately difficult load for the partnering amplifier to drive. Not only does its impedance reach a minimum value of 3 ohms at 108Hz, there is a combination of 4.5 ohms and a 53° capacitive phase angle at 72Hz, and the impedance remains below 4 ohms for much of the midrange and the top octave.
The traces in fig.1 are free from the small discontinuities that would suggest the presence of resonances in the speaker cabinet's walls. Nevertheless, investigating the enclosure's behavior with a simple plastic-tape accelerometer, I found a strong resonance at 309Hz on the side walls level with the upper woofer (fig.2), with a lower-level mode at a slightly lower frequency. However, as Kal Rubinson didn't comment on any congestion in the midrange that might be laid at the feet of this resonance, it's likely that the affected area, hence the audibility, is relatively small. The sidewall in the vicinity of the midrange unit was dead as a doornail.
The saddle centered on 31Hz in the impedance-magnitude trace in fig.1 suggests that this is the tuning frequency of the flared port on the front baffle. Indeed, the port's output, measured in the nearfield, peaks between 22 and 44Hz, with a smooth rolloff above that region unbroken by any midrange resonances (fig.3, red trace). To my surprise, however, the two woofers behaved differently in the bass: while the upper woofer (green trace) had a minimum-motion notch at 30Hz, the lower woofer's (blue) lay at 26Hz. (For clarity, the levels of the two woofers are each raised by 6dB in this graph.) The crossover to the midrange driver (black trace) appears to lie just below 400Hz, and the crossover seems to feature asymmetrical slopes: 24dB/octave high-pass but 18dB/octave low-pass. The woofers' upper-frequency behavior is identical.
Higher in frequency in fig.3, the 804 Diamond's treble response is uneven, with a suckout evident between 1.5 and 3kHz and a peak indicated between 7 and 16kHz. Puzzled by this, I checked the response using a different microphone (a QTC-40) and a different measurement system (Fuzzmeasure 3.0). The result was identical. I then looked at the review my colleague Thomas J. Norton had written for Home Theater magazine: While the measured response of the 804 Diamond in Tom's review (the red trace in the graph here) differed in the fine details, the overall measured response was broadly similar to mine.
Fig.4 shows how these individual responses sum in the farfield. The bump in the upper bass is primarily an artifact of the nearfield measurement technique; as KR found in his listening, the 804 Diamond offers excellent low-frequency extension, the output lying 6dB down at 28Hz. At the other end of the audioband, the combination of presence-region suckout and top-octave peak persists, though the overall trend is basically flat. The effect of the suckout would be to make the speaker sound somewhat laid-back, though the upside is that the 804 Diamond would be forgiving of too-bright recordings, which are common. The peak is a little too high in frequency to render the sound "steely" or "wiry"; instead, it might just emphasize the airiness of the recording.
The B&W's lateral dispersion, normalized to the response on the tweeter axis, is shown in fig.5. Although the suckout at the top of the midrange unit's passband deepens to the speaker's sides, the radiation pattern is even and uniform overallsomething that always correlates with stable and accurate stereo imaging. In the vertical plane (fig.6), a suckout at the upper crossover frequency of 3.9kHz develops 10° above and 15° below the tweeter axis. It also looks as if the flattest treble response occurs 5° below the tweeter axis.
This is confirmed by the 804 Diamond's step response on the tweeter axis (fig.7), which is 40" above the floor: the sharp up/down spike that represents the tweeter's output doesn't quite smoothly blend with the start of the midrange unit's slower-rising output. Moving the microphone down by 5° would bring the midrange unit's output slightly forward in time and eliminate the slight discontinuity in this graph. Fig.7 also indicates that all four drive-units are connected in the same, positive acoustic polarity, this confirmed by examining the step responses of the individual units (not shown). Finally, the 804 Diamond's cumulative spectral-decay plot on the tweeter axis (fig.8) is generally clean in the treble.
In 2004, I visited Bowers & Wilkins' Research Center, in the village of Steyning, West Sussex, nestling in the shadow of England's South Downs, north of Worthing. I was impressed by both the depth and the breadth of the engineering talent and resources I found there. There is no doubt in my mind that B&W's engineers can design a loudspeaker to have any response they desire. That the 804 Diamond does not have a flat on-axis response is thus a mystery. That suckout in the presence region in fig.4, for example, appears from fig.5 to be due to the large-diameter midrange driver narrowing its radiation pattern in the top octave of its passband, despite the FST technology that is intended to prevent that from happening. But overall, the B&W 804 Diamond's measured performance is quite respectable.John Atkinson