On 11/21/16 3:11 PM, Attila Kinali wrote:
On Mon, 21 Nov 2016 14:36:49 -0500
Bob Camp <kb...@n1k.org> wrote:
The reflection issue ahead of the antenna is a reflection of the signal from
a single satellite. The multipath
reflection makes that satellite appear to be further away than it really is.
In the case that the reflected
signal is *stronger* than the desired signal, the multipath reflection
“captures” the receiver and the net
solution is messed up.
Even a weaker reflected signal can cause significant change of the
correlation peak and thus of the apparent distance of the satellite.
Only if the multipath is less than a chip away.. if it's more than a
chip away, it doesn't change the timing of the correlation peak for the
primary path.
Actually, since the "timing"of the recovered code is averaged over many
chips/code periods, "close by" multipath might have an effect because it
affects the "shape" of the peak - it's not nice and triangular.
1 chip = 1 microsecond = 300 meters.
So multipath <300 meters away (probably a specular reflection off
something) - here the geometry helps as you average over multiple code
periods - the multipath timing, relative to the true path, is not fixed
- the reflected path may just disappear and reappear (specular
reflections from something "far-ish" away - 10 meters)
That's why, if you can get your receiver antenna up high above the
surroundings, multipath is less of a problem: the antenna has poor gain
below the horizon, and the enforced distance between antenna and
"nearest possible reflector" is greater, which makes the temporal
movement of the spurious signal that much bigger.
A pathological case would be an antenna next to a vertical wall. There
will likely be two signals with identical strength and very small
differential time that doesn't change very much.
The multipath error envolope diagrams are usually for multipath to
direct path ratios of between 1:2 to 1:10 (mostly depending on what
the author wants to show or how much he wants to cheat).
_______________________________________________
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
