Re: my total miserable failed experiment with PPP and rtklib

[Gary E. Miller]

Yo Hans!

On Mon, 24 Jul 2023 22:26:36 +0200
Hans Mayer <gpsd@ma.yer.at> wrote:

> I was fascinated about this article 
> https://gpsd.gitlab.io/gpsd/ppp-howto.html Therefore I tried it.
> I own a ZED-F9P from U-blox with a fix mounted antenna on the roof of
> my house.

Cool.

> I run 6 times a 24 hour job with the config after reset of my u-blox 
> ZED-F9P ( defined as config1 for my me and my tests )
> 
> I calculated the distance between any of these results and it was not 
> bigger than 37.2 cm.

That sounds about right.  But no need to calculate, just use gpsprof.

> the next test was to use “dynModel 2”. ( 2 = stationary )  (  I
> called it config4 )
> 
> I expected better results. But this wasn’t the case.

Note that "dynModel" standds for Dynamic Model.  When you average over
24 hours there are no dynamics.

> The maximum distance between these results was 22.2 cm. Between all
> 12 measurements it is 40 cm. But the distance for these averaged
> values for config1 and config4 is only 2.4 cm.

As above, best to just post your gpsprof png.

> Therefore I expected the real position of the fixed mounted antenna
> on the roof of my house will be the average value within a circle of
> about 20 cm but maybe even less.

Maybe, maybe not.  Not all GPS errors cancel each other out over time.

> Later on I stumbled over rtklib. It offers a lot of tools and one of
> the programs is rtkrcv. It gives the possibility of a more precision 
> position for a "rover" with the correction from the "base" station.

ftklib has not had many updates in a decade.  Better to use a PPP
service.  Much better.

> Currently I don't have two stations. Only one, the fix mounted as 
> "rover". So I registered myself for these public available ntrip 
> streams. As base station I selected one near my position.

A good base should be less than a km from you.  How far is your base?

> And this was the surprising result: the distance to the average 
> position, which I was calculating previously, was more than 1.1 meter.

Normal, if your base is not within <100 m.

> Probably I have done something wrong.

See above.

> Or is my assumption is wrong
> that 12 times 24 hours PPP measurement gives an accurate value ?

GPS orbits repeat almost exactly every 12 hours, so any multiple of 12 hours
is good.  Unless you are also using other constellations, then it gets
trickier.

> Is there anywhere a forum to discuss rtklib with others ?   Didn't
> find anything till now.

There may be better places, but some people here still use it.   You'll get
better results with a PPP service.

RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
        gem@rellim.com  Tel:+1 541 382 8588

            Veritas liberabit vos. -- Quid est veritas?
    "If you can't measure it, you can't improve it." - Lord Kelvin

pgpF2ttXgkP6g.pgp
Description: OpenPGP digital signature