[time-nuts] GPS Antenna Feed Line Decision
csteinmetz at yandex.com
Tue Sep 5 17:19:08 EDT 2017
> I tried to see if there was any difference in the signal quality reported by Motorola UT+ and U-Blox Neo 6M (not timing grade but has a good 1PPS) over 30m of 2 different cable types.
> I only had RG58(75 Ohm) and RG174(50 Ohm).
RG58 is 50 ohm. RG59 is the "equivalent" 75-ohm cable.
Also, there has been some discussion of internal reflections with mixed
source/cable/load impedances. Yeah, there is some, BUT: The
reflections are generated at the *interfaces* between impedances, not
down the length of any one cable. "F" connectors are 75 ohms, so if any
part of the system uses them, they will generate approximately the same
reflections at the interfaces to 50 ohm components (cable impedance and
impedance of source and/or load) as would using 75 ohm cable.
To eliminate reflections "entirely" (in reality, "mostly"), you need to
use 50 ohm [or 75, or other] components exclusively (source and load Z,
all connectors, and cable).
Yeah, I could pull my Tbolts apart and replace the F connectors with
SMA, BNC, TNC, N, UF, whatever. But any degradation caused by using 75
ohm cable and connectors is completely swamped by other errors in the
system. One of my Tbolts runs with about 30 feet of RG6 quad, the other
with about 150 feet of the same, and all timing and positional solutions
are indistinguishable from a test mule I cobbled together that used only
an adapter (about an inch and a half long) between the antenna and the
Tbolt. In that case, the "cable" delay was so low that any reflections
were time-displaced from the direct signal by such a small amount that
it was meaningless for all practical purposes.
More information about the time-nuts