Digital Audio Labs CardDeluxe PC soundcard Measurements part 3
Fig.7 Digital Audio Labs CardDeluxe, D/A spectrum of 50Hz sinewave, DC-1kHz, at 0dBFS into 600 ohms (linear frequency scale).
Fig.8 Digital Audio Labs CardDeluxe, D/A HF intermodulation spectrum, DC-22kHz, 19+20kHz at 0dBFS into 100k ohms (linear frequency scale).
So far, the CardDeluxe measured as well as the $11,000 Accuphase DP-75V CD player, reviewed in the July issue. But despite such excellent readings, the DAL's jitter performance was disappointing. I assess jitter by driving the device under test with data representing a high-level tone at 11.025kHz, with the least significant bit toggling at 229Hz. I use the Miller Audio Research analyzer to average sixty-four 32k FFTs on the device's analog output. The analyzer software then searches the FFT bins for symmetrical sidebands on either side of the 11.025kHz tone.
The resultant narrow-band spectrum for the CardDeluxe's analog output is shown in fig.9. The actual jitter level is respectably low at 552 picoseconds p-p, with the highest-level sidebands at ±260Hz (marked with a purple "3" in this graph), which is suspiciously close to the frequency of the sidebands seen in fig.2. Data-related jitter (indicated with red numeric markers) is actually quite low, the ±229Hz sidebands contributing just 47ps to the total. But the CardDeluxe's noise floor features a lot of narrow spectral spikes (blue markers) that cannot be due to jitter, as they are not pairs of sidebands.
Fig.9 Digital Audio Labs CardDeluxe, 44.1kHz sampling, high-resolution jitter spectrum of analog output signal (11.025kHz at -6dBFS with LSB toggled at 229Hz). Center frequency of trace, 11.025kHz; frequency range, ±3.5kHz.
To check that these discrete noise components were not due to the fact that the analytical signal was being read from the computer's hard disk, I repeated the measurement using a PS Lambda transport to drive the CardDeluxe's S/PDIF input. The jitter level rose to 846ps, mainly due to an increase in data-related sidebands (due to the S/PDIF link's limited bandwidth) and the appearance of a strong pair of sidebands at ±15Hz (these, in my experience, related to the digital data receiver). But the noise components were still apparent. (Because they are so narrow, these peaks will be averaged out by the wider-band measurements.)
The Miller Analyzer has only a single-ended input, so I performed these measurements by looking at only one phase of the card's analog output. As it is possible that these noise components might be common-mode and will cancel in the balanced output, I repeated the measurement, feeding the balanced signal to the balanced input of a Mackie mixer, and the Mackie's single-ended tape output to the analyzer. The results were the same.—John Atkinson