Digital coaxial audio cable suggestions.


Suggestions for a reasonably flexible 1m long digital coaxial cable for around $200 that is well shielded from RF and sounds good? Also, where can I reliably buy the suggested cable from?
mike60

Showing 2 responses by almarg

One meter is a length that will often not be optimal. See this paper. While the paper recommends 1.5 meters, IMO a very short length (e.g., 6 inches or so, or perhaps as much as 12 inches) may also be a good choice, if that is physically practicable.

There will be exceptions to these guidelines, of course, depending on various technical characteristics of the components that are being connected.

Regards,
-- Al
06-13-12: Mike60
What happens when you have reflections? What do you hear?
There would be an increase in jitter. See this paper for further background on jitter. The degree to which the jitter would have audible consequences is dependent on the particular components. In some cases there may be little or no consequence, depending on the jitter rejection capability of the DAC and other factors.

Typically, though, there would be a general loss of clarity and focus, its exact character depending on the spectral characteristics (frequency content) of the jitter, which in turn are dependent in unpredictable ways on the particular components and setup.
06-13-12: Flashunlock
Would that also apply for a word clock cable. I have a Esoteric player that accepts a 10mhz signal and I'm running a 3m length BNC-BNC 75ohm digital cable from the external word clock?.
I suspect that the risetime and falltime of the 10 MHz signal that is being sent out by whatever is generating your word clock are significantly faster than the risetime and falltime of the S/PDIF output of a typical transport or other S/PDIF source, so it is a fundamentally different situation.

Also, the problem that can occur for S/PDIF with a 1 meter length will be avoided by lengths that are longer than 1.5 meters, as long as the length is not so long that reflections corresponding to one signal edge arrive near the mid-point of the NEXT signal edge, or some other subsequent edge. 3 meters is too short to cause a problem on the "next" edge of a 10 MHz square wave, and is very unlikely to affect subsequent edges, in part because those edges would only be affected by reflections making multiple round-trips (i.e., reflecting and re-reflecting multiple times between the two components), which would cause them to be very small in amplitude.

That said, when possible it is preferable to avoid longer lengths like that, because it will reduce the possibility that jitter may occur as a result of ground loop-related noise or rfi/emi pickup.

Regards,
-- Al