Author: Jonathan Borden MD

The DAC provides a unique situation to measure a proxy for end to end phase error using nothing more than a high resolution spectrum analyzer.
A high resolution digital encoding of a sine wave at a specified frequency is fed into the DAC. The pure sine wave is generated and fed at either PCM or DSD e.g. 24/192 or 24/354 or DSD256, DSD512 in a format specific for the DAC Under Test (DUT). This may, for example, go through the DUT USB interface or Ethernet interface.
The analogue output is measured.
Close-in phase noise in the DUT will result in widening of the measured FFT peak. By measuring the width of the peak, the degree of close-in error is approximated. This is analogous to the "linewidth" measurement used in laser spectroscopy.
The frequency interval of the FFT needs to be small enough in order to allow accurate measurement e.g. < 0.01Hz
This close-in phase error is related to "flicker noise". Slightly further out phase error related to "shot noise" forms the skirt of the peak. The baseline is formed by "white" noise.