Hi again,

My final question (yes I hope!) is in relation to the repeatability of raw double difference carrier phase residuals for L1.

This does not relate to DDBase itself (I'm sure now this is stable for repeatability), but to RDDs as a whole (in theory). After hearing from you all in previous enquiries I thought you might have an idea.

I understand double differences are highly sensitive to any "minute movements" or "temperature variation" which may cause 0.10-0.005mm changes (Axelrad) in residuals. However I'm finding this problem an ongoing concern, as I can't appear to get raw double difference experiments 100% correct repeated - it seems like an impossible goal.

My experiment consists of observing 2 consecutive sidereal sessions (day1,day2) of raw double difference carrier phase residuals, at permanent high precision survey grade CORS sites here in my town. I am using either IGS rapid or final ephemeris for both days.

However I'm finding variation or 'wandering' in the plotted RDD series, which match periodically, diverge and then may reconverge later. Typically divergences last for 500-1000 seconds, or indefinately. Sometimes the series begins seperaterd and then matches after 1 hour or so.

If there was something wrong with the sites, the series would not match perfectly at all, at any time, and would continue offset. However in all cases the are combinations of matching, offset, divergence and convergence. Making it impossible to obtain 'matching' RDD series (for possible adaptive filtering).

This appears to be a systematic effect.

In some ways the problem is similar to that seen using lower precision ground coordinates for the sites (as alerted to by Brian in a previous question). However I have since rectified this problem (which has worked a treat, thanks again Brian).

The CORS sites are permanent monumentations using higher precision 'geodetic grade' receivers with optimal site design, ie. tops of large buildings using ground planes. Assumptions for many multipath mitigation techniques require unchanged environments and this has been a key goal to maintain throughout.

I should note that all post-fit residuals or DDRs match perfectly, absolutely ideal for the whole observation periods. However unfortunately cannot be used in this experiment as I'm wishing to correct the raw DDs before estimation.

Any possible ideas for this variation in raw double differences?

I have plenty of screengrabs which illustrate this problem.

Thanking you kindly,


-- MichaelRigby - 16 Nov 2007


ALERT! If you answer a question - or have a question you asked answered by someone - please remember to edit the page and set the status to answered. The status is in a drop-down list below the edit box.

I am not sure I understand the problem. If you have data from two sites on two consecutive days, with fixed positions and nothing changed, and you compute DDs for the same two satellites, then I am not sure that you would see exactly the same result on the two days. The crucial point is the amplitude of this oscillation, which I don't see.... If it is of order the baseline length, then I would say the site(s) have moved; if it is much smaller then I would suspect multipath; I suppose there would be very small variations due to a number of things....

-- BrianTolman - 26 Nov 2007 No such template def TMPL:DEF{PROMPT:supportquery}

Topic revision: r3 - 27 Nov 2007, RickMach

This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Foswiki? Send feedback