Digital coaxial audio cable suggestions.

Listen to Al. Either go very short to avoid reflections on impedance boundaries or go 1.5m with typical transport transition time that is about 25ns. Assuming threshold in the half of that reflection from the end of the cable (started by the beginning of the transition) has to come later than 12.5ns. Assuming 2/3 of the speed of light signal travels thru 1.5m forth and back (total of 3m) in 15ns just missing threshold point. 1m cable would be much worse bringing reflection in about 10ns - just before signal crossing the threshold. Any disturbance of transition at the threshold will convert to time jitter (equivalent to noise in frequency domain).

Digital cable behaves like transmission line (reflections) when one way propagation time to point of first possible reflection (end of cable) is shorter than 1/8 of transition time (rule of thumb) - about 3ns in our case making for 0.6m. 0.5m cable should be safe but I would go no longer than 12" (as Al stated) since driver is in few inch distance from the connector.

What happens when you have reflections? What do you hear? Ideal length for me would be maybe 0.7m I think, but need to measure. I have a 1m cheap cable currently working.

1.5 will work ok, but i will have to coil it up at one end. So what to buy and from where? $200 or so, good quality and sound, flexible and well shielded from RF.

Kijanki:
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?.

Thanks.

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