Sidebar 4: Measurements
I measured a different sample of the Franco Serblin Accordo Goldberg loudspeaker to those auditioned by Martin Colloms. Mine had the serial number 326. I used DRA Labs' MLSSA system, a calibrated DPA 4006 microphone, and an Earthworks microphone preamplifier to measure the quasi-anechoic frequency- and time-domain behavior in the farfield. I used an Earthworks QTC-40 microphone, which has a ¼" capsule, for the nearfield responses.
The saddle centered on 40Hz in the impedance magnitude trace suggests that this is the tuning frequency of the port on the speaker's rear panel. The blue trace in fig.3 shows the woofer's response measured in the nearfield; it has the expected notch at the port tuning frequency. The port's response, plotted in the ratio of the square root of its radiating diameter to that of the woofer (footnote 2), is shown as the red trace in fig.3. Its output peaks almost an octave higher than its tuning frequency, before rolling off cleanly, other than a low-level peak between 400Hz and 600Hz.
Footnote 1: EPDR is the resistive load that gives rise to the same peak dissipation in an amplifier's output devices as the loudspeaker. See "Audio Power Amplifiers for Loudspeaker Loads," JAES, Vol.42 No.9, September 1994, and stereophile.com/reference/707heavy/index.html. Footnote 2: See stereophile.com/content/measuring-loudspeakers-part-three-page-6. Footnote 3: A woofer is omnidirectional at low frequencies, due to the fact that the wavelengths of the sound are much larger than the size of the baffle in which it is mounted. As the frequency rises to where the wavelength is of the same order as the baffle's dimensions, the radiation pattern becomes more directional. As more energy is now being projected in the forward direction, the on-axis response rises, which results in the upper-midrange output being higher than that in the lower midrange. Footnote 4: My measured response is very similar to that measured by Paul Miller for Stereophile's sister magazine Hi-Fi News.
Fig.1 Franco Serblin Accordo Goldberg, electrical impedance (solid) and phase (dashed) (2 ohms/vertical div.).
The Accordo Goldberg's voltage sensitivity is specified as 87dB/2.83V/m. My B-weighted estimate, measured on the tweeter axis, was usefully higher, at 88.6dB(B)/2.83V/m. The Accordo Goldberg's nominal impedance is specified as 7 ohms, with a minimum value of 3.8 ohms at 5.5kHz. The impedance magnitude (fig.1, solid trace), examined with Dayton Audio's DATS V2 system, remained above 6 ohms from the low bass through the mid-treble region, though it dropped below 4 ohms between 4kHz and 8kHz. The minimum impedance was 3.8 ohms at 5kHz. The electrical phase angle (fig.1, dotted trace) is occasionally high, which means that the effective resistance, or EPDR (footnote 1), drops below 4 ohms in several regions in the bass and below 3 ohms between 2.3kHz and 5.6Hz and above 7kHz. The minimum EPDR values are 2.84 ohms at 109Hz, 1.81 ohms at 3.4kHz, and 2.1 ohms at 12.5kHz. The Accordo Goldberg is a relatively demanding load for the partnering amplifier.
Fig.2 Franco Serblin Accordo Goldberg, cumulative spectral-decay plot calculated from output of accelerometer fastened to the center of the longer sidewall (MLS driving voltage to speaker, 7.55V; measurement bandwidth, 2kHz).
Some very small discontinuities in the impedance traces imply the presence of cabinet resonances of some kind. When I investigated the enclosure's resonant modes with a plastic-tape accelerometer, I found a low-level mode at 375Hz on the longer convex sidewall (fig.2) and an even lower-level mode at 656Hz on the shorter, concave sidewall. It is extremely unlikely that this behavior will have audible consequences.
Fig.3 Franco Serblin Accordo Goldberg, anechoic response on tweeter axis at 50", averaged across 30° horizontal window and corrected for microphone response, with the nearfield responses of the woofer (blue), port (red), and their complex sum (black) respectively plotted below 350Hz, 500Hz, and 310Hz.
The black trace below 310Hz in fig.3 shows the complex sum of the woofer's and port's nearfield responses, taking into account both acoustic phase and the fact that the port is on the enclosure's rear. The peak between 70Hz and 140Hz will be due to the nearfield measurement technique, which assumes that the baffle extends to infinity in both horizontal and vertical planes. The Franco Serblin speaker's woofer tuning is maximally flat, but the low frequencies will sound somewhat lightweight without boundary reinforcement.
The sum of the Accordo Goldberg's nearfield responses in fig.3 is spliced at 310Hz to its farfield response, this averaged across a 30° horizontal window centered on the tweeter axis. There is a lack of energy in the lower midrange, which might be due to insufficient baffle-step compensation in the crossover (footnote 3). There is then a slight lack of energy in the presence region and a slight excess between 4kHz and 11kHz. Predicting the effect of this behavior on perceived sound quality is difficult, as this will depend both on which frequency region the listener takes as a reference, which in turn will depend on the music being played. If the levels in the upper midrange and mid-treble are perceived as being correct, the loudspeaker will sound lean and polite. If the lower mids and the presence region are heard as correct, the upper mids will sound exaggerated, though this may also be interpreted as added detail.
Fig.4 Franco Serblin Accordo Goldberg, lateral response family at 50", normalized to response on tweeter axis, from back to front: differences in response 90–5° off axis, reference response, differences in response 5–90° off axis.
Fig.5 Franco Serblin Accordo Goldberg, vertical response family at 50", normalized to response on tweeter axis, from back to front: differences in response 45–5° above axis, reference response, differences in response 5–45° below axis.
The perceived balance will also depend on the speaker's horizontal radiation pattern. This is shown in fig.4, normalized to the response on the tweeter axis, which therefore appears as a straight line. The tweeter's dispersion narrows sharply in its top octave, but the mid-treble region evens out to the Accordo Goldberg's sides. (The manual suggests that the tweeter "intensity" can be lowered by increasing the angle of toe-in so that the tweeter axes cross in front of the listener.) The even spacing of the contour lines below 5kHz will correlate with stable, accurate stereo imaging. Fig.5 shows the Accordo Goldberg's dispersion in the vertical plane, again normalized to the response on the tweeter axis. A suckout between 2kHz and 3kHz develops more than 10° above and 20° below the tweeter axis, which suggests that the crossover frequency is in this region. However, the lack of presence-region energy is accentuated more than 10° below that axis.
Fig.6 Franco Serblin Accordo Goldberg, step response on tweeter axis at 50" (5ms time window, 30kHz bandwidth).
Fig.7 Franco Serblin Accordo Goldberg, cumulative spectral-decay plot on tweeter axis at 50" (0.15ms risetime).
In the time domain, the Serblin speaker's step response (fig.6) indicates that the tweeter and woofer are both connected in positive acoustic polarity. The tweeter's output arrives first at the microphone, and the decay of its step cleanly merges with the positive-going start of the woofer's step, which indicates an optimal crossover topology. Other than some low-level hash in the mid-treble region, the Accordo Goldberg's cumulative spectral-decay, or waterfall, plot (fig.7) is superbly clean.
There are positive aspects of the Franco Serblin Accordo Goldberg's measured performance, such as that impressive waterfall plot. But its departure from a flat frequency response (footnote 4) suggests that its sonic character will depend on system matching and setup.—John Atkinson
Footnote 1: EPDR is the resistive load that gives rise to the same peak dissipation in an amplifier's output devices as the loudspeaker. See "Audio Power Amplifiers for Loudspeaker Loads," JAES, Vol.42 No.9, September 1994, and stereophile.com/reference/707heavy/index.html. Footnote 2: See stereophile.com/content/measuring-loudspeakers-part-three-page-6. Footnote 3: A woofer is omnidirectional at low frequencies, due to the fact that the wavelengths of the sound are much larger than the size of the baffle in which it is mounted. As the frequency rises to where the wavelength is of the same order as the baffle's dimensions, the radiation pattern becomes more directional. As more energy is now being projected in the forward direction, the on-axis response rises, which results in the upper-midrange output being higher than that in the lower midrange. Footnote 4: My measured response is very similar to that measured by Paul Miller for Stereophile's sister magazine Hi-Fi News.
