Klipsch Palladium P-39F loudspeaker Measurements
Measuring a speaker as bulky and heavy as the Klipsch Palladium P-39F is always problematic: the speaker can't be raised high enough off the ground to eliminate the bounce of its sound off the floor between the speaker and the measuring microphone. I therefore had to use more aggressive time-windowing than usual to generate the frequency-response graphs, which in turn means that these graphs' resolution in the midrange is more limited than usual.
Looking first at the Palladium's voltage sensitivity, my estimate was 94.8dB(B)/2.83V/m, which is within experimental error of the specified 95dB. The P-39F is thus one of the most sensitive models I have encountered in 20 years of measuring loudspeakersit will play very loud with very few watts. However, the speaker is not as efficient at low frequencies as you might expect; its impedance drops between 3 and 4 ohms in the lower midrange (fig.1), with a nasty combination of 4 ohms magnitude and 50° electrical phase angle at 80Hz. The three woofers do demand a significant amount of current to keep up with the horn-loaded midrange and treble drive-units. The impedance is significantly higher in the region covered by the latter two drivers, averaging 10 ohms, which means that the speaker will sound tilted-up at high frequencies when used with a tube amplifier having a high source impedance.
Fig.1 Klipsch Palladium P-39F, electrical impedance (solid) and phase (dashed). (2 ohms/vertical div.)
The traces in fig.1 are free from the small wrinkles in the midrange that would suggest the presence of cabinet resonances. Nevertheless, investigating the vibrational behavior of the curved panels with a simple plastic-tape accelerometer did uncover a strong resonant mode at 414Hz (fig.2) and another at 520Hz. I note that Wes Phillips mentioned no midrange congestion that might have resulted from this behavior; these resonances may indeed be high enough in frequency to mitigate the resonances' subjective effect.
Fig.2 Klipsch Palladium P-39F, cumulative spectral-decay plot calculated from output of accelerometer fastened to center of side panel (MLS driving voltage to speaker, 7.55V; measurement bandwidth, 2kHz).
Turning to the P-39F's acoustical output, its three woofers behaved identically, as did its three ports. I show in fig.3, therefore, the summed outputs of each array as the blue and green traces, respectively. The ports' output peaks between 28 and 34Hz, as expected from the impedance graph, where the saddle between the two magnitude peaks in the bass occurred at 32Hz. The level of the ports' output is suppressed a little in absolute terms, implying a somewhat overdamped reflex alignment. This is always a good idea with large speakers, because it preserves midbass definition while letting the usual "room gain" at low frequencies give a flat response. There is a small peak in the ports' output at 130Hz, but otherwise the rolloff is smooth. Higher in frequency, the woofers hand off to the horn-loaded midrange unit at 450Hz with what appear to be symmetrical fourth-order filter slopes. The midrange unit looks well behaved in its passband before crossing over to the horn-loaded tweeter at 3.2kHz. Again, the filter slopes appear to be fourth-order or higher. The tweeter also appears well behaved in its passband, but is slightly hot in its top-octave response. The dome resonance occurs just below 30kHz, above which the output drops rapidly.
Fig.3 Klipsch Palladium P-39F, acoustic crossover on midrange axis at 50", corrected for microphone response, with summed nearfield responses of woofers (blue) and ports (green), scaled in the ratio of the square roots of their radiating areas.
Fig.4 shows how these individual outputs sum in the farfield. Klipsch recommends that the listener sit 5° below the tweeter axis, which, at the microphone distance I use for my measurements, is close to the midrange axis. I therefore performed all of my farfield measurements on that axis for reasons of repeatability. The apparent rise in response in the upper bass is mainly due to the nearfield measurement techniquein anechoic conditions, the Palladium P-39F will actually be pretty flat in the bass region. The speaker doesn't go quite as low in the bass as its size might lead one to expect, but this must be balanced against its extremely high sensitivity and the overdamped alignment. In all but very large rooms, the P-39F should give almost the full measure of low frequencies. Other than a slightly elevated top octave, the Klipsch's midrange and treble are impressively flat.
Fig.4 Klipsch Palladium P-39F, anechoic response on midrange axis at 50", averaged across 30° horizontal window and corrected for microphone response, with complex sum of nearfield woofers' and ports' responses plotted below 300Hz.
Whether or not this flat on-axis response translates into a neutral tonal balance will depend on the speaker's dispersion. Fig.5 shows how the Klipsch's response on the midrange axis changes to its sides (only the differences in response are shown). The slight flare at the cursor position is actually due to the on-axis suckout at 4.3kHz filling in to the speaker's sides. The contour lines are otherwise evenly spaced in this graph, which is what matters, but close inspection reveals that the speaker is a little more directional than a typical direct-radiating model in the region covered by its horn-loaded drivers. The tweeter also becomes significantly more directional above 13kHz, which will ameliorate the slight boost in its on-axis output in the same region (for those whose hearing extends that high, of course). I note that Wes didn't toe the Palladiums all the way in to his listening seat, which suggests he gave a high priority to getting a smooth treble balance, while the speaker's well-controlled dispersion over the first 20° to its sides meant he could still get excellent imaging precision and a smooth transition between the drive-units.
Fig.5 Klipsch Palladium P-39F, lateral response family at 50", normalized to response on midrange axis, from back to front: differences in response 905° off axis, reference response, differences in response 590° off axis.
In the vertical plane (fig.6), a suckout develops in the upper crossover region more than 10° below and 15° above the midrange axis, which is 45" from the floor. Fortunately, Klipsch supplies longer spikes for the rear of the baseplate to tilt the P-39F forward a little, to ensure that the ears of a seated listener are on the intended axis.
Fig.6 Klipsch Palladium P-39F, vertical response family at 50", normalized to response on midrange axis, from back to front: differences in response 205° above axis, reference response, differences in response 510° below axis.
The Palladiums' spatially averaged in-room response (fig.7), calculated by taking 20 separate 1/6-octave response measurements for the left and right speakers individually in a vertical rectangular grid centered on the position of WP's head in his listening chair, indicates good integration of the speakers' output with the room acoustics, if not quite up to the standard set by the YG Anat Reference II Professional in March 2009 (fig.7) and the Thiel CS3.7 in December 2008 (fig.8). The Klipsch's middle and high trebles are slightly shelved down compared with the other two speakers (especially the Thiel), even though the Palladiums were being driven by the tubed VTL amplifiers, which will tilt up the treble region. The peaks and dips below 300Hz are due to room acoustics effects that have not been eliminated by the spatial averaging. Even so, the Palladiums maintain their in-room low-frequency output to 25Hz or so.
Fig.7 Klipsch Palladium P-39F, spatially averaged, 1/6-octave response in WP's listening room.
Fig.8 Klipsch Palladium P-39F, step response on midrange axis at 50" (5ms time window, 30kHz bandwidth).
In the time domain, the Klipsch P-39F's step response on the midrange axis (fig.8) looks more complicated than the norm, due to the fact that the tweeter's output leads that of the midrange unit by 0.5 millisecond, due to the latter's setback to permit horn loading. Both tweeter and midrange unit appear to be connected in inverted acoustic polaritythe tweeter's step is the sharp negative-going spike just before the 4ms marker, followed by a lazier, negative-going spike from the midrange unit approximately 500µs later. The woofers are all connected in positive acoustic polarity, and their slow rise away from the timeline in this graph overlays the positive-going overshoot of the midrange unit's step. That each drive-unit's step smoothly blends into that of the next lower in frequency correlates with their good frequency-domain integration seen in fig.4. The Palladium P-39F's cumulative spectral-decay plot (fig.9) is basically clean, but with a slight amount of delayed energy apparent at the top of the midrange unit's passband. Peculiarly, this graph also suggests that the woofer/midrange integration is not as good as indicated in fig.4.
Fig.9 Klipsch Palladium P-39F, cumulative spectral-decay plot on midrange axis at 50" (0.15ms risetime).
Overall, this is impressive measured performance, suggesting that Klipsch has some excellent speaker engineers on staff. However, Wes did mention the fact that the P-39F's three pairs of input terminals, positioned in a recess under the speaker's baseplate, elicited much swearing from me. Indeed they did, as this placement made hooking up and dressing speaker cables for the measurements far more difficult than it needed to be. Had Klipsch continued the terminal panel recess out to the rear of the baseplate so that cables could be routed straight back, that would have been a significant improvement in practicality. As it stands, dressing the thick speaker cables Wes likes to use was awkward: even with the P-39Fs sitting on their spikes, there was still not quite enough clearance between the bottom of the baseplate and the floor to avoid trapping the cables.John Atkinson