I have two GPS-RTK2 modules and want to get a very basic base and rover setup working.
From what I understand so far, each ZED-F9P has two UARTS (UART1 and UART2).
I can view output on both UART ports using PuTTY and a couple of RS-232 converter boards (this works).
By default, UART1 (TX/MISO and RX/MOSI on GPS-RTK2) outputs NMEA data at 38400 and I can view this working.
UART2 by default is configured to input and output RTCM messages at 38400. When I connect to TX2 I see nothing over PuTTy.
The ZED-F9P Integration Manual (page 20 of 114) lists steps to configure ZED-F9P as a base station. I followed these steps and I do not see the RTCM messages as shown on page 24 of 114 in Figure 12. I also do not see the Fix Mode change to TIME as shown on Figure 13. Any idea why these steps do not work as written? These steps seem to be configuring UART1 to output RTCM messages. Why would they configure UART1 to output RTCM data when UART2 does this by default?
Assuming I can get the steps shown for base station configuration working, would I connect UART1 TX on the base station to UART1 RX on the rover? Why not just connect UART2 TX on the base station to UART2 RX on the rover and skip the steps listed in the Integration Manual for base station setup?
I’m confused. Thanks.
Unfortunately, we have not tested your exact setup. I would advise trying to look at our most recent guides if possible:
https://www.sparkfun.com/news/3430
OK. I got the two modules to work as a base and rover.
Answer: you can directly connect the TX on the base to the RX on the rover. Configuring the two was the issue.
My new question is about accuracy.
The base and rover are supposed to show 1cm level accuracy. Not seeing that…
Although the U-Center software displays accuracy levels around 0.01m, looking at the NMEA data says something else.
I have a simple test with the base located about 2 feet away from the rover. Looking at the results of rover (rover in fixed position) readings over a period of 1 hour (1 second readings) indicates variation of about +/- 1.5 inches (typically cyclical). This is a measure of precision and not accuracy - but the accuracy can’t be better than the precision so no way this testing shows anything close to 1cm (0.4 inches) accuracy.
I also tried moving the rover a series of fixed distances - 2 inches per step (where each position was held for 5 minutes). Errors were in the 1 to 2 inch ballpark.
Are there any issues with how close the base station can be with respect to rover?
I used “survey-in” setting on base station. Any issues with this setting?
Thanks in advance for any input on this…
Are you using an antenna ground plate? https://www.sparkfun.com/products/15004
Those accuracy measurements, provided by the U-blox datasheets, are usually tested under ideal conditions. Results may vary.
Yes, using steel plates under the antennas.
Just curious what kind of results you are getting with your base-rover setups?
With rover in fixed position do you have a scatter plot of results you could share?
Just trying to figure out how far to pursue this …
I know oversampling and low pass filtering will improve some things but it is best to look at unprocessed results first.
The guides were written by an engineer spurred by his own GPS curiosities. What ever values or results they share are what we currently have in test setups. I apologize for the inconvenience. Any progress you’re making is helping make our community better, for that we sincerely say thank you.
My ZED-F9P rover reports a “0.0141 meter accuracy” when it has a RTK fix solution. I need to read the SparkFun library code to figure out exactly what Ublox parameter is being reported and then read the Unix documentation to full understand the meaning and significance.
To help you out, where in the ucenter application would I get such a scatter plot or the data to make one? I can post the results I am getting with a pair of F9P’s
