[time-nuts] Quality of timing mode GPS vs survey accuracy

[Ole Petter Ronningen]

Just to add one datapoint to the discussion, I am presently looking at a
professional grade receiver ($20k+++), tracking GPS/GLONASS/GALIELO/BEIDOU,
44 sats total as I type this. It reports standard deviation of the XYZ
coordinates of 0.76m, 0.44m and 1.3m respectively. It is not crystal clear
how this std is calculated, but I can only guess it is the std off x number
of positions calculated on a per-epoch basis.

Granted, the antenna I am using is *highly* suspicious, but it seems to
track most signals with a decent SNR.

The point of this is that even a *great* receiver can only get you so far.
Some form of averaging must take place, either in the receiver itself (as
is done with survey mode and "position hold") or externally (as is done
with PPP).

As I understand it. in the former case with a fixed position, the std
quoted above will give an indication of the short term noise in the time
pulse - which we already know to be horrible, and which the GPSDO is rigged
to filter out. Multipath is a bigger issue, since the "moving average" of
the receiver can/will skew the position appreciably in a 24 hour period.

(And then ofcouse ionosphere and related cans of worms.)

Ole

On Sat, May 9, 2020 at 1:55 AM jimlux <jimlux at earthlink.net> wrote:

> On 5/8/20 3:38 PM, Charles Steinmetz wrote:
> > Jim wrote:
> >
> >> I'm attaching a screen shot of 4 receivers, taken at the same time
> >> (within seconds,not time-nuts same time) All sitting on the table
> >> together.
> >
> > One might expect closer agreement if all of the receivers were fed from
> > one antenna through a splitter, but even then each rx will choose its
> > own series of constellations and will switch constellations at different
> > times.  There will also be phase dispersion among the rx input filters.
> >   And lots of other small effects.
> >
>
> one might so expect. But I'm not too sure.
>
> There's >10 meters difference in the positions (in the X coordinate of
> the ECEF position alone).
>
> They're all tracking the same satellites (except b203)
>
> but the S/N is noticeably different among them.
>
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe, go to
> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.
>