28-09-2016, 09:44 PM
At audio frequencies, it won't matter if you use 75Ω coax and 50Ω connectors, working into 600Ω load, 10kΩ load, or whatever - unless the cable is a significant fraction of a wavelength long. Which means, even for 20kHz, 2 miles-ish.
At high frequencies/long lengths, yes it is usual to match the receiving end to the cable so that no reflections occur. If this is done properly, the sending end won't matter - it could be a voltage source low source impedance) or a current source (high source impedance). Since the cable+load looks just like a pure resistor, there'll be no problems either way. It's only if the load is NOT properly matched to the cable, that a reflection will occur, in which case if the source impedance is not equal to the cable impedance, a second reflection will happen and the same signal will set off to the load, reaching it at two cable delays after the original.
Which really is just expanding on what Mark Hennesey and Gryphon have already said...
At high frequencies/long lengths, yes it is usual to match the receiving end to the cable so that no reflections occur. If this is done properly, the sending end won't matter - it could be a voltage source low source impedance) or a current source (high source impedance). Since the cable+load looks just like a pure resistor, there'll be no problems either way. It's only if the load is NOT properly matched to the cable, that a reflection will occur, in which case if the source impedance is not equal to the cable impedance, a second reflection will happen and the same signal will set off to the load, reaching it at two cable delays after the original.
Which really is just expanding on what Mark Hennesey and Gryphon have already said...
