Post-Processed RTK

A summation of a recent survey utilizing RTK and Static methods.

Here is the equipment/software used:
Trim 4700 w/Microcentered L1-L2 (base), Trim 4800 (rover), TSC2 with Survey Pro 5.0.3, TBC 2.6 and Star*Net Pro 6.0.

OK, here's what I did:
Base collecting data at 5-sec. intervals, Rover collecting data at 5-sec. intervals.

Perform survey; Calibrate base to some other control from previous projects in the area. Noticed an odd scale factor and manually fixed to appropriate combined factor in DC. Collect 2-min. of data on each point in an RTK survey (about 35 prop. corners). Collect 1 hour of data each on 2 of the property corners at the extremities (while base is still running), straight up static at this point, but still 5-sec. intervals.

Post-Process; Import static data including 5 CGPS RINEX files with broadcast ephemeris into TBC with processed RTK vectors via the raw file out of the DC. Process. Nothing alarming going on yet. Export vectors.

Adjustment; Minimally constrained adjustment warps out a 14.00 error factor (probably the screwy scale factor).

Back to processor; I blitz out the RTK vecs and re-run the static only, minimally constrained adjusts just fine. Now I pull in the DAT file from the RTK rover, +/- 35 point, each 2-min. sessions. TBC provided a fixed solution for every basline. Good RMS etc. The CGPS are not that close either.

Back to adjustment; minimally constrained network now returns an error factor of 0.985. end up fixing X, Y and Z (NAD83 2011.00 with no update to epoch) from all of the CGPS except the Z on one that is always problematic (4 PBO and 1 NGS, guess which one is not like the other). Error factor before solving transformations and applying geoid and vertical deflec modeling comes in at 1.26. Final adjustment comes in at 1.008. Standard errors for the RTK points hovering in the 0.008' to 0.020' range and error ellipses around 0.03' (errors are not scaled).

I thought this was way to good to be real, so I revisited TBC and found that I had 2 baselines for each observation; 1 static baseline and "PP Stop & Go". Did the data double-up or something? What is happening here? I imported 10 files; 3 DAT from static obs, 1 DAT from RTK obs, 5 CGPS RINEX files and 1 ephemeris. From what I can tell on the import, the files are bringing in a single point for each RTK shot.

I deleted the project, repeated the process and got the same result. Are my suspected repeat data lulling Star*Net into a false sense of security? All of the adjustment settings are plain vanilla 95% confidence region stuff. What is going on here. My data can't be that good...