Sidebar 3: Measurements
I measured the Pass Labs XP-27 phono preamplifier with my Audio Precision SYS2722 system, repeating some tests with the magazine's higher-resolution APx500 analyzer. To minimize noise, I connected a wire from the grounding terminal on the preamplifier chassis's rear panel to the analyzer ground connections. The XP-27 gets warm. The temperature of the top panel after an hour of testing was 99.5°F (37.5°C).
The inputs preserved absolute polarity from both the balanced and single-ended outputs. The gain from the balanced outputs was fairly close to the numeric values: 57.85dB with the gain set to "53," 65.5dB with it set to "66," and 77.5dB with it set to "76." The gains from the single-ended outputs were exactly 6dB lower than these values.
The input impedances were close to the values set with the rotary control on the front panel: 30 ohms with it set to "30," 101 ohms with it set to "100," 166 ohms with it set to "160," and 348 ohms with it set to "320," for example, with all these values consistent from 20Hz to 20kHz. With the impedance set to "1k," the measured values were 1138 ohms at 20Hz, 980 ohms at 1kHz, and 966 ohms at 20kHz. With the control set to "47k," I measured 45k ohms at 20Hz and 1kHz, but 22.2k ohms at 20kHz. (The capacitance was set to 100pF for these measurements.) The source impedance is specified as 150 ohms from both types of output; I measured 200 ohms from the single-ended outputs, 540 ohms from the balanced outputs. Though these impedances are higher than the specified values, they are still low in absolute terms.
Fig.1 Pass Labs XP-27, response with RIAA correction into 100k ohms (left channel blue, right red) and with the high-pass filter engaged (left green, right gray) (1dB/vertical div.).
The error in the XP-27's RIAA equalization (fig.1, blue and red traces) was extremely low, though there was a very slight boost in the lower midrange and bass, reaching +0.25dB at 100Hz and +0.36dB at 20Hz. The output rolls off above the audioband, and with the high-pass filter selected with the front-panel button, the response was down by 2dB at 10Hz (green and gray traces). The channel matching was excellent overall. Channel separation (not shown) was >80dB in both directions below 8kHz and still 72dB at the top of the audioband. (I always look at the undriven channel's output on the oscilloscope screen while examining the channel separation; the XP-27's reading was true crosstalk rather than noise.)
Fig.2 Pass Labs XP-27, gain "53," spectrum, DC–1kHz, of output ref. 5mV input (left channel blue, right red, linear frequency scale, 20dB/vertical div.).
With the gain set to "53," the Pass's unweighted, wideband S/N ratio, measured with the input shorted to ground and ref. 1kHz at 5mV, was a very good 80.5dB in the left channel, 74.dB in the right. Restricting the measurement bandwidth to 22Hz– 22kHz increased the ratios to 84dB in the left channel, 75.5dB in the right, and the A-weighted ratios were a superb 94.5dB, left, and 93.1dB, right. (These ratios were measured at the balanced outputs; they were identical at the unbalanced outputs.) As expected, the ratios were lower at the highest gain setting, at 64.2dB/56.7dB, left/right, unweighted, 65dB/57dB when restricted to the audioband, and 81.7dB/77dB when A-weighted. Spectral analysis of the XP-27's low-frequency noisefloor with the gain set to "53" and ref. 1kHz at 5mV (fig.2) revealed that the lower ratios in the right channel were due to higher levels of AC supply–related and other spuriae in that channel (red trace) than in the left (blue trace). Even so, the XP-27 is a quiet preamplifier.
The XP-27's overload margin with the gain set to "76," ref. 1kHz at the standard MC level of 500μV, was an excellent 18.7dB in both channels from 20Hz to 20kHz. It was even greater at 20Hz and 1kHz with the gain set to "66," at 30.6dB and 32.5dB, respectively, though it was the same 18.7dB at 20kHz. Setting the gain to "53" gave overload margins of 18.3dB at 20Hz and 1kHz ref. 1kHz at the standard MM level of 5mV, but the margin dropped to –1dB at 20kHz. This gain setting should only be used with low-output moving magnet cartridges.
Fig.3 Pass Labs XP-27, gain "53," spectrum of 1kHz sinewave, DC–10kHz, into 100k ohms for 5mV input (left channel blue, right red, linear frequency scale).
The XP-27 offered very low distortion. Fig.3 shows the spectrum of the preamplifier's output reproducing an input signal of 1kHz at 5mV with the gain set to "53." The only distortion harmonics that can be seen above the noisefloor are the second, at –76dB (0.015%) and the third, at –104dB (0.0006%). These very low levels didn't increase when I reduced the load impedance to the current-demanding 600 ohms and were similar at lower frequencies and with the two higher gain settings.
Fig.4 Pass Labs XP-27, gain "53," spectrum of 20kHz sinewave, DC–80kHz, into 100k ohms for 35mV input (left channel blue, right red, linear frequency scale).
Fig.5 Pass Labs XP-27, gain "76," HF intermodulation spectrum, DC–30kHz, 19+20kHz into 100k ohms for 10mV peak input (left channel blue, right red, linear frequency scale).
Fig.4 shows a wideband spectrum of the XP-27's output with the gain set to "53" while it reproduced a 20kHz tone at 35mV input, which is 2dB below the point where the XP-27 overloads at this frequency. Both the second and third harmonics lie close to –60dB (0.1%). With this reduced linearity at high frequencies and levels, the XP-27 offered higher levels of intermodulation distortion with an equal mix of 19kHz and 20kHz tones than I was expecting. Fig.5 shows the spectrum of the XP-27's output with the gain set to "76" and the peak input level 13dB below the clipping voltage at 20kHz. While the higher-order products were at a respectably low –90dB (0.003%), the difference product at 1kHz lay at –54dB (0.2%). This was despite the levels of the second and third harmonics with a 20kHz tone at the same peak level and the same gain setting respectively lying at –76dB (0.02%) and –81dB (0.01%).
Other than the limited overload margin at the top of the audioband at the lowest gain setting and the somewhat disappointing high-frequency intermodulation results, the Pass Labs XP-27 offers accurate RIAA equalization, very low noise, subjectively innocuous harmonic distortion, and an output stage that handles low impedances with aplomb.—John Atkinson
