Recently, I tested the RTK PointPerfect solution with the ZED-F9P module. I collected data once per second using the NAV-PVT message type and ran the u-blox module continuously for 24 hours. The antenna was positioned under nearly open skies with a proper ground plane. All data was stored in a database for analysis.
Throughout the test, the reported horizontal and vertical precision never exceeded 0.14m, which was impressive. However, when I reviewed the minimum and maximum values of latitude and longitude, I noticed a significant difference—over 1.4m!
This discrepancy seems puzzling, considering the precision metrics. Has anyone else observed such large deviations in their data, or could there be an explanation for this behavior? Any insights would be greatly appreciated!
The last key roll-over was Nov 17/18, 2024
To start diagnosing what exactly was occuring, I think you’d need to collecting UBX-NAV-PVT,SIG,SAT and UBX-RXM-SPARTN,COR
I’d recommend using HPG 1.50, or 1.51 as these have addressed some issue related to roll-over, and the 3 hour delta between GPS and GLO time
Yeah, 1.4m still does seem excessive, I’d probably dwell on the antenna situation, because a static antenna, even in RTK Float should be a lot tighter than that.
Perhaps look at a Sky Map View in uCenter Classic to better understand any obstructions.
The receiver doesn’t really have any gauge of absolute accuracy.
Is there any indication that it’s not working? They have a 30-day free trial, which you could test in uCenter Classic with a ZED-F9P and an internet connection. An L-Band modem configuration might run you $130-$200 USD
You’re in a location that falls within the coverage area of the L-Band satellite service. The concern would then be line-of-sight visibility due to the satellite being rather low in the southern sky.
Yes I have seen jumps like this also and have talked with UBOX support about it. They say thats why the zed 9r has dead reckoning to correct stuff like this. Also, the really odd thing is that upping the number of nav reading to the number of nav solutions maked the problem worse.
