dCS 972 D/D converter Measurements part 3

I repeated these tests using the Audio Precision System One's AES/EBU output to drive the Elgar, then redigitizing the analog signal at a 24-bit word length with the dCS 904 ADC to drive the 972 at both 96kHz and 44.1kHz rates. The results with a 1kHz tone were comparable, though things did look better with a more complex signal.

Fig.8, for example, shows the DScope-derived spectrum of a signal consisting of 19kHz and 20kHz tones at equal levels, sampled at 44.1kHz and passed through the 972. The 1kHz, 2kHz, 18kHz, and 19kHz intermodulation products are all at very low levels, and the noise floor appears random. Changing the sample rate to 96kHz but with the 972 still outputting 44..1kHz data gave the spectrum shown in fig.9. There's no appreciable difference with this torture-test signal. Note also that the noise floor, having passed through both the Elgar and the dCS 904, both operating at 24 bits, is around the 20-bit level. This is actually an impressive tribute to both components, considering that D/A and A/D conversions have taken place.

Fig.8 dCS 972, HF intermodulation spectrum, DC-22.05kHz, 19+20kHz, 24-bit word length, 44.1kHz sample frequency (linear frequency scale, 20dB/vertical div.).

Fig.9 dCS 972, HF intermodulation spectrum, DC-22.05kHz, 19+20kHz, 24-bit word length, downsampled from 96kHz to 44.1kHz (linear frequency scale, 20dB/vertical div.).

I must say that, given the huge amount of number-crunching taking place in the 972 while it performs sample-rate conversion to and from 44.1kHz, these measurements indicate excellent performance.

Finally, I used Miller Audio Research's National Instruments-based Analyzer to look at the Elgar's and 972's jitter performances. This uses a special test tone: an 11.025kHz sinewave at -6dBFS, overlaid with the LSB toggling on and off at a 229Hz rate, and stored on a CD-R with inherently low jitter, or more correctly, time-base error.

Fig.10 shows a high-resolution spectrum of the Elgar's analog output, with the CD-R played on a Meridian 500 transport and connected to the Elgar via 6' of Canare 100 ohm AES/EBU cable. The jitter level was very low at 142 picoseconds peak-peak, and the noise floor was one of the lowest I have ever measured. The main jitter components are data-related and indicated with red numeric markers. There are also some infrasonic sidebands, these indicated with purple markers. The brown "4" shows a pair of power-supply-related sidebands at ±180Hz, these very low in level.. The components marked with a blue "55" and a blue "46" are very-low-level discrete tones that seem to be characteristic of the Elgar.

Fig.10 dCS Elgar, high-resolution jitter spectrum of analog output signal, 44.1kHz sample frequency, driven by Meridian 500 CD transport via 6' Canare 110-ohm AES/EBU cable (11kHz at -6dBFS with LSB toggled at 229Hz). Center frequency of trace, 11kHz; frequency range, ±3.5kHz. (Grayed-out trace is the Meridian 508.24.)