Sidebar 3: Measurements
When I performed the measurements of the original Rogue Sphinx integrated amplifier to accompany Herb Reichert's review in the August 2014 issue of Stereophile, I was impressed by what I found. "Even without taking into account its affordable price, Rogue Audio's Sphinx offers excellent measured performance with little sign of the usual compromises made in class-D designs," I concluded, adding that it "also has an excellent, moving-magnet–compatible phono stage." Herb reports in this issue on the sound of the new V3 version of the Sphinx, which features a new phono stage that can be switched between moving-coil and moving-magnet operation.
Editor Jim Austin suggested that I should examine the V3's measured performance, so I hooked up Herb's sample (serial number 3057) to my Audio Precision SYS2722 system (see the January 2008 "As We See It"). As the Sphinx V3, like its predecessor, has a class-D output stage—the amplifier uses a pair of Hypex UcD-180 modules—I used a precision passive low-pass filter from Audio Precision to eliminate ultrasonic noise that would otherwise drive the analyzer's input stage into slew-rate limiting.
Examined first at its line inputs, the V3 offered slightly higher maximum gain than the original Sphinx, 32.7dB into 8 ohms vs 31.16dB, but still with excellent matching between the channels. The input impedance was similarly high, at >100k ohms, and the amplifier preserved absolute polarity at all four sets of outputs—loudspeaker, headphone, and variable and fixed-level preamplifier. The loudspeaker terminals' output impedance was still very low, at 0.04 ohms at 20Hz and 1kHz, rising only slightly to 0.1 ohm at 20kHz.
I didn't look at the performance of the Sphinx as a headphone amplifier in 2014, as Herb hadn't auditioned it with headphones. The 2020 Sphinx's headphone output offered a maximum gain of –1.8dB—ie, a line input level of 1V gave an output at the headphone jack of 814mV—sourced from a very low impedance of 1 ohm. The V3 will thus have no problem driving low-impedance headphones. While the line-output frequency responses of both the original Sphinx and the V3 rolled off in the low bass, reaching –2dB at 10Hz (fig.1, blue and red traces), the headphone output on the V3 was flat to 10Hz (fig.1, green and gray traces). It also extended slightly higher at the other end of the spectrum, reaching –3dB at 50kHz (right) and 60kHz (left). Distortion was low at the headphone output, all the harmonics lying below –106dB (0.0005%) ref. 1V into 300 ohms, with the third harmonic slightly higher in level than the second.
Measured at the fixed preamplifier outputs, so I could accurately capture the response above the headphone amplifier's ultrasonic rolloff, the V3's RIAA correction was not quite up to the standard set by the 2014 Sphinx, with a mismatch of up to 0.9dB in the treble (fig.5). The bass rolls off early, conforming to the IEC modification of the RIAA specification, though it reaches –3dB at 30Hz compared with the original's 11Hz. Channel separation via the phono input was also not quite to the standard set by the original Sphinx, at 50dB right-to-left and 67dB left-to-right at 1kHz. The phono input's noise performance in MC mode was good, with unweighted audioband signal/noise ratios (ref. 1kHz at 5mV input signal) of 52.5dB (average of both channels). The ratios improved by 6dB when A-weighted.
Fig.1 Rogue Sphinx, volume control set to maximum, frequency response at 1V from loudspeaker output into 8 ohms (left channel blue, right red) and from headphone output into 300 ohms (left green, right gray) (2dB/vertical div.).
Returning to the V3's performance at the loudspeaker outputs, channel separation (not shown) was the same as before, at 60dB in both directions below 1kHz and around 40dB at the top of the audioband. The V3's unweighted signal/noise ratio in the audioband, taken with the line input shorted but the volume control at its maximum, was also the same at 67.7dB left and 66.1dB right, both ref. 2.83V into 8 ohms. This was primarily due to full-wave–rectified, supply-related spuriae (fig.2), though the random noise floor components were also relatively high in level.
Fig.2 Rogue Sphinx, spectrum of 1kHz sinewave, DC–1kHz, at 1W into 8 ohms with Audio Precision low-pass filter (linear frequency scale).
While the original Sphinx met its maximum power specification of 100Wpc into 8 ohms (20dBW) at 1% THD+noise, the V3 fell just short, delivering 96Wpc into 8 ohms (19.82dBW, fig.3). Into 4 ohms, the new amplifier clipped at 150Wpc (18.75dBW, fig.4) compared with the original's 155W (18.9dBW). The distortion was very low at powers up to a few tens of watts into either load, however, and was predominantly the subjectively innocuous second harmonic.
Fig.3 Rogue Sphinx, distortion (%) vs 1kHz continuous output power into 8 ohms with Audio Precision low-pass filter.
Fig.4 Rogue Sphinx, distortion (%) vs 1kHz continuous output power into 4 ohms with Audio Precision low-pass filter.
I measured the behavior of the original Sphinx's phono input at the fixed-output jacks with the volume control set to its minimum so that I could keep the output stage quiescent. (The fixed-level output operates at unity gain for line-level inputs, though the variable preamplifier output offered a small maximum gain of 1dB.) However, I ran into trouble with the V3, as there was a fairly high level of ultrasonic noise at the fixed and variable preamplifier outputs, even with the amplifier's grounding post connected to the analyzer ground. (Peculiarly, this noise was much lower in level when I changed from moving-magnet to moving-coil mode with the internal switches.) I therefore examined the phono input's behavior primarily at the headphone output.
Set to MM, the phono input preserved absolute polarity and offered a maximum gain of 43.3dB at the headphone output, though the maximum gain at the loudspeaker terminals was a high 77.7dB. With the internal switches set to MC, the maximum gain was 12dB higher, which is appropriate for all but low-output moving-coil cartridges. With the phono stage's input impedance set to 47k ohms, I measured 45k ohms at 20Hz, 47k ohms at 1kHz, and 42k ohms at 20kHz. With the impedance set to 300 ohms, I measured 299 ohms from 20Hz to 20kHz. All the other input impedances also closely matched the nominal values.
Fig.5 Rogue Sphinx V3, MC mode response with RIAA correction into 100k ohms (left channel blue, right red) (0.5dB/vertical div.).
As with the first version of the Sphinx, the V3's phono input offered superb overload margins, close to 29dB across the band in MM mode and 20dB in MC mode. Distortion at typical recorded levels was low (fig.6). Though some high-order harmonics are present in this graph, these all lie at or below –84dB (0.006%). Intermodulation distortion via the phono input was also low (fig.7), though low-level, supply-related components can be seen.
Fig.6 Rogue Sphinx V3, MC mode, spectrum, DC–1kHz, of output ref. 1mV input (left channel blue, right red, linear frequency scale, 20dB/vertical div.).
Fig.7 Rogue Sphinx V3, MC mode, HF intermodulation spectrum, DC–30kHz, 19+20kHz into 100k ohms for 400mV peak input (left channel blue, right red, linear frequency scale).
As a line-level integrated amplifier and headphone amplifier, the Sphinx V3 continues the high standard set by the original. The addition of an MC mode to the phono stage is welcome, though I was puzzled by the noise problem I found at the preamplifier outputs with the MM mode.—John Atkinson
