Dedicated digital COAX cable vs. stereo RCA

[One and a half]

It's a little more complicated. The simple formula above is only for pure sine waves. An S/PDIF signal is not a pure sine wave, it is a digital signal which contains sharp ramp times, these "edges" contain harmonics which go up much higher than than the fundamental. These harmonics ARE high enough in frequency to be affected by transmission line characteristics of digital interconnects.

 

These transmission line effects can cause significant changes in the shape of the edges of an S/PDIF signal. These edge differences can cause significant behavior in the PLLs which are used to extract the clock in many S/PDIF receivers.

 

It's not just simple edge rate changes, impedance mismatches WILL cause reflections, even for S/PDIF signals. If you feed an S/PDIF signal into a TDR (time domain reflectometer) you CAN see reflections from a real S/PDIF signal. The amplitude of those reflections depend on the impedance mismatch and the ramp time of the signal. Slower ramp times cause less intense reflections. BUT slower ramp times can cause many PLLS to generate more jitter. It's a complex tradeoff made by the designer.

 

The length of the interconnect has several different effects on this system. Short interconnects cause the transmission line effects to occur at higher frequency which generally results in better quality edges and less intense reflections, BUT it can cause the reflections to be far more disruptive to a receiver.

 

Even though there is a standard for impedance, very few transmitters and receivers actually meet this, complicating things significantly. A cable which causes a good match with one component may be very wrong with another.

 

The over all result is that the situation is far too complex to make generalizations. The adage "try it yourself and see" is REALLY appropriate here. One cable can be great with one set of components, not great with another, and another set of components may be quite insensitive to the difference between two specific cables.

 

The impedance mismatch caused by RCA jacks over say BNC jacks (as long as they are real 75 ohm BNCs) DOES make a difference in the signal. Exactly what the difference is can vary depending on the cable and the equipment on both ends. How the receiver responds to these differences also varies.

 

So going from RCA to BNC MAY make a difference in a particular situation and it may also not make a difference. Its very system dependent.

 

John S.

 

Thanks John for that piece, S/PDIF depends on the implementation and more head scratching...

 

Does the use of an S/PDIF reclocker work to avoid the reflections in the transmission? I imagine internally it could prove beneficial, but does it work at the receiver. S'pose they are all PLL to a degree and reject. If AES3 is used, are the reflections assymetrical and rejected? AES3 is all 110 Ohm, so a different set of impedances again.