External Clocks do they make a difference
Dec 30, 2023 14:33:46 GMT
stellabagpuss and orange55 like this
Post by MartinT on Dec 30, 2023 14:33:46 GMT
I don't really understand the point you're making in the 'need to be clear' statement?
Let's set a baseline: all digital data is represented by an analogue waveform. A perfect transmitted squarewave is impossible to achieve. Ergo, that waveform is susceptible to jitter (timebase distortion) and phase noise (phase distortion). This noise is superimposed on the data.
In terms of data integrity, there is no issue as we are well able to ensure it, otherwise our computers wouldn't work. In terms of analogue conversion, that noise will enter the DAC together with the data and impact its conversion accuracy, affecting the quality of the derived analogue signal.
Where the noise has the worst effect is closest to the DAC, that should be clear from the above. However, it is very hard to fully remove it so the more stages are involved in reducing the noise, the better the conversion accuracy. Stages involving high accuracy, low noise clocks (and their power supplies), and 'moats' (like EtherREGEN, U18, fibre optic spans) and good ground plane grounding (grounding boxes) will all help to remove some of the noise.
Whether those stages are in the ethernet transmission, USB, S/PDIF doesn't really matter as long as you keep taking out noise and use components with low inherent noise generation.
I believe the above explains it all. Implementation is where it's at, and a lot more difficult to achieve than the description.
Let's set a baseline: all digital data is represented by an analogue waveform. A perfect transmitted squarewave is impossible to achieve. Ergo, that waveform is susceptible to jitter (timebase distortion) and phase noise (phase distortion). This noise is superimposed on the data.
In terms of data integrity, there is no issue as we are well able to ensure it, otherwise our computers wouldn't work. In terms of analogue conversion, that noise will enter the DAC together with the data and impact its conversion accuracy, affecting the quality of the derived analogue signal.
Where the noise has the worst effect is closest to the DAC, that should be clear from the above. However, it is very hard to fully remove it so the more stages are involved in reducing the noise, the better the conversion accuracy. Stages involving high accuracy, low noise clocks (and their power supplies), and 'moats' (like EtherREGEN, U18, fibre optic spans) and good ground plane grounding (grounding boxes) will all help to remove some of the noise.
Whether those stages are in the ethernet transmission, USB, S/PDIF doesn't really matter as long as you keep taking out noise and use components with low inherent noise generation.
I believe the above explains it all. Implementation is where it's at, and a lot more difficult to achieve than the description.
