On Monday, November 12, 2018 at 5:01:25 PM UTC-5, Don Pearce wrote:
On Mon, 12 Nov 2018 13:54:13 -0800 (PST), wrote:
On Monday, November 12, 2018 at 4:22:25 PM UTC-5, Mike Rivers wrote:
On 11/12/2018 2:44 PM, Don Pearce wrote:
Wire is just wire when source impedance, load impedance and wire
impedance are the same. Any other condition, the length of the wire
will introduce frequency response errors.
Even at RF frequencies, only ONE end (either end) needs to be properly matched to Zo and the reflection will be absorbed and there will be no frequency response issues (other than the inherent loss of the cable increasing with frequency).
I agree, at audio frequencies and practical lengths, the R is the only issue.
mark
It has to be the load end that is matched otherwise there will be
standing waves. A mismatch at the source end alone will result in
overall level error. Actually it is still in theory a standing wave,
but it is one from a zero length line, so it just looks like a
broadband incorrect impedance.
It is the common MISconception that the load must be terminated.
If the load end is mismatched, but the source is properly terminated, the reflection will be absorbed by the source. The load will not see any reflections so the response AT THE LOAD will be flat. This is done all the time in digital PWB traces. The signal along the line has standing waves and looks ugly, because there is both the forward and reflected wave present along the line, but AT THE LOAD, where it counts, there is only the incident signal. The reflection is not re-reflected because it is absorbed by the matched source.
Mark