Sidebar 3: Measurements
I measured the BorderPatrol P21 EXD using my Audio Precision SYS2722 system (see the January 2008 "As We See It"). I installed the numbered 300B tubes in the correct sockets, as identified in the manual, but I didn't adjust the bias for the two pairs of 300B output tubes, feeling that I should measure the amplifier in the condition it was auditioned by Ken Micallef.
The P21 EXD's voltage gain at 1kHz into 8 ohms was lower than the norm, at 17dB from the 8 ohm tap into 8 ohms, 15.9dB from the 4 ohm tap into 8 ohms, and 14.5dB from the 4 ohm tap into 4 ohms. While performing this test, I could hear a relatively loud 1kHz tone coming from the amplifier. Investigating with a stethoscope revealed that this tone was loudest on the metal top panel. In use, this noise will be masked by the sound coming from the loudspeakers, but it is unusual. Both sets of output transformer taps preserved absolute polarity (ie, were non-inverting). The input impedance was lower than specified but still usefully high, measuring 87k ohms at 20Hz, 75k at 1kHz, and 53k ohms at 20kHz.
The output impedance was high, at 3.3 ohms at 20Hz and 1kHz from the 8 ohm tap, and 1.9 ohms from the 4 ohm tap at those frequencies. At 20kHz, these impedances respectively increased to 3.7 and 2.5 ohms. As a result, the modulation of the BorderPatrol's frequency response with the 8 ohm tap driving our standard simulated loudspeaker was a very audible ±1.9dB (fig.1, gray trace). From the 4 ohm tap, the response variations were still ±1.4dB (fig.2). The BorderPatrol's frequency response rolled off rapidly above the audioband. The reproduction of a 1kHz waveform from both taps was still suitably square (fig.3), though with a small amount of overshoot on the leading edges. Looking at a 10kHz squarewave (fig.4), that overshoot is critically damped, with no ringing evident.
While its low distortion at powers of a few hundred milliwatts will mean it will work best with very sensitive speakers, overall, I was disappointed by the BorderPatrol P21 EXD's measured performance, especially considering that, as reviewed, it costs $14,450. Modern tube amplifiers can do better than this, or perform about the same for a lower price. The question is: Did KM like the sound of the P21 EXD because of its measured performance or despite it? I don't know the answer to that question.—John Atkinson
Fig.1 BorderPatrol P21 EXD, 8 ohm tap, frequency response at 2.83V into: simulated loudspeaker load (gray), 8 ohms (left channel blue, right red), 4 ohms (left cyan, right magenta), 2 ohms (green) (1dB/vertical div.).
Fig.2 BorderPatrol P21 EXD, 4 ohm tap, frequency response at 2.83V into: simulated loudspeaker load (gray), 8 ohms (left channel blue, right red), 4 ohms (left cyan, right magenta), 2 ohms (green) (1dB/vertical div.).
Fig.3 BorderPatrol P21 EXD, 8 ohm tap, small-signal 1kHz squarewave into 8 ohms.
Fig.4 BorderPatrol P21 EXD, 8 ohm tap, small-signal 10kHz squarewave into 8 ohms.
Channel separation was good, at 90dB in both directions at 1kHz. However, the unweighted, wideband signal/noise ratio, ref. 1W into 8 ohms and taken from the 8 ohm outputs and with the inputs shorted to ground, was disappointing, at 60.6dB left and 56.8dB right. These S/N ratios improved to 76.7 and 73.1dB, respectively, when an A-weighting filter was switched into circuit. The ratios from the 4 ohm tap were consistently 2dB greater, and spectral analysis of the low-frequency noise floor (fig.5) indicated that the primary source of noise was AC-supply–related spikes at 120 and 240Hz, though the 60Hz and odd-order harmonics are much lower in level. The frequencies of these spikes, which are related to the full-wave–rectified supply frequency, suggest that they are due to suboptimal internal grounding. I experimented with the grounding between the Audio Precision and the P21, and repeated some of the measurements with the EXD power supply's ground lifted with the rear-panel switch. The published results were taken with the lowest noise I could achieve.
Fig.5 BorderPatrol P21 EXD, 8 ohm tap, spectrum of 1kHz sinewave, DC–1kHz, at 1W into 8 ohms (left channel blue, right red; linear frequency scale).
BorderPatrol specifies the P21 EXD as delivering up to 20Wpc (13dBW), presumably with the load matched to the nominal output transformer tap. With clipping defined as when the THD+noise in the output reaches 1%, I measured a clipping power of 19.5Wpc into 8 ohms from the 8 ohm tap (12.9dBW, fig.6) and into 4 ohms from the 4 ohm tap (9.9dBW, fig.7), with both channels driven. With the 8 ohm tap driving 4 ohms (fig.8), the BorderPatrol clipped at 7.1Wpc (5.5dBW). It did better when the load impedance was higher than the tap value, the 4 ohm tap clipping at 15Wpc into 8 ohms (11.8dBW, fig.9).
Fig.6 BorderPatrol P21 EXD, 8 ohm tap, distortion (%) vs 1kHz continuous output power into 8 ohms.
Fig.7 BorderPatrol P21 EXD, 4 ohm tap, distortion (%) vs 1kHz continuous output power into 8 ohms.
Fig.8 BorderPatrol P21 EXD, 8 ohm tap, distortion (%) vs 1kHz continuous output power into 4 ohms.
Fig.9 BorderPatrol P21 EXD, 4 ohm tap, distortion (%) vs 1kHz continuous output power into 4 ohms.
I examined how the P21's percentage of THD+N varied with frequency at a relatively low level (2.83V, equivalent to 1W into 8 ohms and 2W into 4 ohms). Even so, and from the lower-midband-distortion 4 ohm tap, while the THD+N was respectably low in the midrange into 8 ohms (fig.10, blue and red traces) and 4 ohms (cyan, magenta), there were large rises in both the treble and bass. The higher treble distortion is due to the decrease in open-loop voltage gain as the frequency increases; the higher bass distortion is most likely due to the increase in saturation of the output transformer's core as the frequency decreases. This is poor performance. Note how the right channel (red and magenta traces) has lower midrange distortion than the left (blue, cyan). This will be due to the bias for the left channel being not quite optimal. (I assume it must have drifted during the auditioning.)
Fig.10 BorderPatrol P21 EXD, 4 ohm tap, THD+N (%) vs frequency at 2.83V into: 8 ohms (left channel blue, right red), 4 ohms (left cyan, right magenta).
The THD+N waveform with a 1kHz signal at 1Wpc from the 8 ohm tap into 8 ohms (fig.11) indicates that the distortion mostly comprises the relatively benign second and third harmonics, though every harmonic is surrounded by power-supply–related spuriae (fig.12). The second harmonic is highest in level even at low frequencies, though higher-order products are also present at lower levels (fig.13). Tested with an equal mix of 19 and 20kHz tones, the BorderPatrol produced high levels of intermodulation distortion (fig.14), the difference product at 1kHz lying at –32dB (2.5%), despite the low power level at which I took this measurement.
Fig.11 BorderPatrol P21 EXD, 8 ohm tap, 1kHz waveform at 1W into 8 ohms, 0.11% THD+N (blue); distortion and noise waveform with fundamental notched out (red, not to scale).
Fig.12 BorderPatrol P21 EXD, 8 ohm tap, spectrum of 1kHz sinewave, DC–1kHz, at 1W into 8 ohms (left channel blue, right red; linear frequency scale).
Fig.13 BorderPatrol P21 EXD, 8 ohm tap, spectrum of 50Hz sinewave, DC–1kHz, at 1W into 8 ohms (left channel blue, right red; linear frequency scale).
Fig.14 BorderPatrol P21 EXD, 4 ohm tap, HF intermodulation spectrum, DC–30kHz, 19+20kHz at 1W peak into 8 ohms (left channel blue, right red; linear frequency scale).
