Sidebar 3: Measurements
I measured the Questyle M12 with my Audio Precision SYS2722 system (see the January 2008 As We See It). I performed a full set of tests with WAV and AIFF test-tone files played with Pure Music 3.0 and Adobe Audition on my MacBook Pro running on battery power, and on my iPad mini using the Lightning-to-USB-A–to–USB-C cable combination.
Apple's USB Prober utility identified the M12 as "M12" from "Questyle" with the serial number "0." The USB port operated in the optimal isochronous asynchronous mode, and Apple's AudioMIDI utility revealed that the M12 accepted 16-, 24-, and 32-bit integer data sampled at all rates from 44.1 to 768kHz rather than the 384kHz specified by Questyle. A Drok USB tester indicated that with no signal being passed, the M12 drew 40mA of current from the host's USB port. Driving a 1kHz tone at 0dBFS into 100k ohms, the M12's current draw increased to 90mA, while with the full-scale signal into the AudioQuest NightHawk headphones' 23 ohms, the current draw was 150mA. (All measurements are ±5mA.)
The M12's maximum gain is specified as adjusting according to the load impedance—high gain with loads above 70 ohms, low gain with loads below 70 ohms. With a load higher than 600 ohms, the M12 will switch automatically to a specified maximum output level of 2V. With the DAC connected to the Audio Precision analyzer's high 100k ohm load, the maximum output level was 1.937V and the Gain LED on the top of the DAC illuminated red. However, when I switched the load impedance from 100k ohms to 30 ohms without unplugging the output jack plug, the LED didn't change from red to green, as it had done with the low-impedance headphones I used for my auditioning. I realized that the M12 detects the load impedance when its output is disconnected then reconnected. However, when I did this with a 30 ohm load connected to both output channels, the gain LED still lit up red. A paradox.
The maximum output level into 30 ohms only changed slightly, but more significantly, the output waveform was now clipping on the negative half-cycles. (I suspect the charge pump used to generate the output stage's negative voltage rail from the single-ended 5V USB supply was running out of juice into this load in the high-gain mode.) Increasing the load impedance to 60 ohms eliminated the clipping. I continued my testing primarily with the 100k ohm load.
The M12 preserved absolute polarity (ie, was noninverting) and offered a very low output impedance of 0.5 ohm at 20Hz and 1kHz, 0.8 ohm at 20kHz. Fig.1 shows the M12's impulse response with 44.1kHz data. The M12's impulse response has a small amount of ringing before the single high sample and more ringing after it. This is typical of the "Hybrid" filter for PCM data offered by the ESS Sabre DAC chips (footnote 1). The filter's ultrasonic rolloff (fig.2, magenta and red traces) starts just below 20kHz and reaches full stop-band attenuation at exactly half the sample rate (the vertical green line at 22.05kHz). The aliased image at 25kHz of a full-scale tone at 19.1kHz (cyan, blue) was not suppressed as much as I was expecting, lying at –93dB. Other than the third at –93dB (0.002%), the distortion harmonics associated with the 19.1kHz tone all lay below –100dB (0.001%), however.
Fig.3 shows the M12's frequency response with data sampled at 44.1, 96kHz, and 192kHz. The response follows the same basic shape at all three sample rates, with a steep rolloff just below half of each rate. Channel separation at 1kHz was good in both directions, at 62dB. An increase in bit depth from 16 to 24, with dithered data representing a 1kHz tone at –90dBFS, dropped the M12's noisefloor by 6dB (fig.4). This suggests a resolution of 17 bits, not quite as good as the AudioQuest DragonFly Cobalt's 18 bits (footnote 2), but still respectable for a USB bus-powered DAC. When I played undithered data representing a tone at exactly –90.31dBFS, the waveform was symmetrical, with a negligible DC offset, but the three DC voltage levels described by the data were obscured by high-frequency noise both with 16-bit data (fig.5) and with 24-bit data (not shown).
Footnote 1: See, for example, fig.5 here. Footnote2: See fig.4 here.
Fig.1 Questyle M12, impulse response (one sample at 0dBFS, 44.1kHz sampling, 4ms time window).
Fig.2 Questyle M12, wideband spectrum of white noise at –4dBFS (left channel red, right magenta) and 19.1kHz tone at 0dBFS (left blue, right cyan) into 100k ohms with data sampled at 44.1kHz (20dB/vertical div.).
Fig.3 Questyle M12, frequency response at –12dBFS into 100k ohms with data sampled at: 44.1kHz (left channel blue, right green), 96kHz (left cyan, right magenta), and 192kHz (left blue, right red) (1dB/vertical div.).
Fig.4 Questyle M12, spectrum with noise and spuriae of dithered 1kHz tone at –90dBFS with: 16-bit data (left channel cyan, right magenta), 24-bit data (left blue, right red) (20dB/vertical div.).
Fig.5 Questyle M12, waveform of undithered 1kHz sinewave at –90.31dBFS, 16-bit data (left channel blue, right red).
The Questyle produced low levels of harmonic distortion with the high 100k ohm load, even at full level (fig.6), with the third harmonic the highest in level at just –110dB (0.0003%). As I mentioned earlier, replacing the 100k load with 30 ohms, the M12's output clipped. However, when I reduced the signal level by 3dB to 1.4V, the clipping vanished, and the distortion was as low as it had been with 100k ohms (fig.7).
Fig.6 Questyle M12, 24-bit data, spectrum of 50Hz sinewave, DC–1kHz, at 0dBFS into 100k ohms (left channel blue, right red; linear frequency scale).
Fig.7 Questyle M12, 24-bit data, spectrum of 50Hz sinewave, DC–1kHz, at –3dBFS into 30 ohms (left channel red; linear frequency scale).
The M12 offered virtually no intermodulation distortion when reproducing an equal mix of 19 and 20kHz tones at 0dBFS into 100k ohms. Even when I reduced the load impedance to 300 ohms, the difference tone at 1kHz was absent (fig.8), and while it made an appearance at the same level into 30 ohms (fig.9), it lay at just –120dB (0.0001%).
Fig.8 Questyle M12, 24-bit data, HF intermodulation spectrum, DC–30kHz, 19+20kHz at 0dBFS into 300 ohms, 44.1kHz data (left channel blue, right red; linear frequency scale).
Fig.9 Questyle M12, 24-bit data, HF intermodulation spectrum, DC–30kHz, 19+20kHz at 0dBFS into 30 ohms, 44.1kHz data (left channel blue, right red; linear frequency scale).
The M12 offered excellent rejection of word-clock jitter. Fig.10 shows the spectrum of the M12's output when it was fed high-level 16-bit J-Test data. All the odd-order harmonics of the undithered low-frequency, LSB-level squarewave lie at the correct levels, and there are no other sideband pairs visible. With 24-bit J-Test data, the spectrum was clean (fig.11).
Fig.10 Questyle M12, high-resolution jitter spectrum of analog output signal, 11.025kHz at –6dBFS, sampled at 44.1kHz with LSB toggled at 229Hz: 16-bit USB data sourced from MacBook Pro (left channel blue, right red). Center frequency of trace, 11.025kHz; frequency range, ±3.5kHz.
Fig.11 Questyle M12, high-resolution jitter spectrum of analog output signal, 11.025kHz at –6dBFS, sampled at 44.1kHz with LSB toggled at 229Hz: 24-bit USB data sourced from MacBook Pro (left channel blue, right red). Center frequency of trace, 11.025kHz; frequency range, ±3.5kHz.
Questyle's M12 performed well on the test bench, in particular offering vanishingly low levels of both harmonic and intermodulation distortion.—John Atkinson
Footnote 1: See, for example, fig.5 here. Footnote2: See fig.4 here.
