I am having problems on Windows 10 with RTKLIB and an NTRIP correction source (NYSNET). I can get the board working fine with the same correction data using U-center so I know the hardware is working.
Has anyone got this working with RTKLIB?
- I setup this device as my rover and NTRIP source as correction with nothing in the base selection.
- Both lights blink green but I never get my solution plot (or the middle light blinking).
- My NTRIP server is sending RTCM3 messages and I have tried the rover as RTCM3, RTCM2, and ublox format without success.
- I also downloaded a version of RTKLIB which is supposed to support this new ublox chip from http://rtkexplorer.com/downloads/rtklib-code/ without any different behavior.
- I wrote logfiles for both the inputs and they look ok (rover mixture of NMEA ascii and binary UBX, corrections all binary — I assume this is RTCM3).
Here are some of my RTKLIB settings.
Thanks for any help here.
Hi, what are you trying to achive ?
Are you trying to feed the ntrip sender to your u-blox for it to compute a real time solution (onboard u-blox algo), if so, you should use STRSVR utility (from rtklib) with input as ntrip client connected to your ntrip correction source, and output to the serial port of your rover.
If you are trying to achieve a position computed by RTKLIB algorithms from u-blox raw data, you’ll to use RTKNAVI and need to output the raw data measurements from the rover (message UBX-RXM-RAWX) or the observables as RTCMV3 stream (as if the rover was a base) from the serial port of the rover conencted to your PC.
The second input stream of RTKNAVI should be the base station one as a ntrip client, and not the correction one. This stream needs to contain the ephemerids message because unless wrong, you cannot output ephemerids data from the F9P so those message needs to come from elsewhere in order for rtknavi to compute a position.
When going post-processing, those ephemerids are the brdcNNN.19g or .19n (in 2019) that you can download for a given GPS day NNN from various servers ( server list included in RTKGET)
Hi Matioupi, Thanks a lot for your help.
I just want to get close to subcentimeter accuracy in whatever way is easiest. I used to use RTKNAVI with a SkyTraq successfully but didn’t get enough fixes fast enough for my lawnmower application. I was using a garmin antenna which wasn’t up to the task and am in a wooded area. I was not aware that the ublox chip was smart enough to apply corrections and that seems like a better approach so I just tried STRSVR and fed the correction data to the serial port and immediately got an RTK fix as indicated by the LEDs on the simpleRTK2B!
1. how to I get my corrected positions out of the simpleRTK2B? The serial port is busy with STRSVR and won’t open from another program.
2. Assuming I can read from the GPS while it is getting a correction stream, how can I tell it is in RTK mode?
Hi, from your answer, I think what you want to do is :
– feed the rtcm correction stream to the F9P
– get the PVT output stream from the F9P with all the indicators about position type and accuracy.
and not to actually have rtklib computing the rtk solution (which is another good exercise and quality check for F9P solution quality)
The easiest way is to use u-center as this software will allow easily to feed the rtcm thought the serial port (PC to F9P) and on the same serial port (probably the USB one in your case) send the internal F9P computations results and show them in u-center. My prefered settings for this are to disable all NMEA outputs and enable the following Ublox messages a (on the USB port) :
With those messages enabled, the u-center will clearly display 3D/DGNSS/FIX to tell you that you are in fixed ambiguitie solution mode, and you’ll get the accuracies down to few cm (depending on the baseline length and sky conditions).
If you don’t want to use u-center (but you would usually want to do this in a fully integrated system environment), you need a software to feed the ntrip correction to your receiver. STRSVR is one of them easy to use (and the command line version of it can be easily be compiled on nice dev platforms such as raspberry pi). On PC there is also GNSS Surfer (https://igs.bkg.bund.de/ntrip/download with more features like ascii decoding of the RTCM stream) that can do this task.
As for parsing/displaying the output, if not using your own, you need a software that is able to understand the PVT output stream format out of the u-blox so either NMEA or UBX depending how you configure it. e.g. RTKplot can handle a connection to a NMEA stream, but do not parse the UBX-NAV-* messages so you would need to configure your F9P with NMEA type outputs, eg GPGGA. Idon’t know other off the shelf softwares that understand the UBX-NAV-* messages. If someone have other names than u-center, i’d be glad to learn about other ones.
As for the issue you raised that you cannot open the same serial port as an input and output from 2 different softwares (STR2SVR & RTKPLOT) I remember I could workaround this in the past with some virtual serial port splitter (software name was VSPE at time, but that was a while ago and you can probably find something else newer. Still the simplisest way would be to wire one af the free uarts on the board itself and use a TTL to USB / RS232 converter.
I am going to take a look at the reconfigure the GPS. I also want to use RTKLIB the old way to look at the position plots and see if they are drifting at all. Also, the USB port contention is not very important because I hope to use the PixHawk interface and try Arudurover and AgOpenGPS in the long term.
You have been very helpful. Thanks a lot.
I finally got this working this morning. I downloaded this file from a Tim (from RTKEXPORER Blog) at http://rtkexplorer.com/downloads/gps-data/ from the first post. The I extracted only the .conf file from that zip.
Then I got rid of his output and log settings which were serial outputs and changed the base to point to my NYSNET source. Now I am getting a fix in RTKLIB for the first time! I still need to compare configurations to see what really fixed the problem.
maybe you can have a look into the RTKNAVI RTK Monitor (the little small rectangle above the Start Button)…and there navigate to the ‘Error/Warning’ page. Maybe you will find there any hint to solve the problem.
m4tk. Thanks a lot for your help.
I did look at the RTKNAVI RTK Monitor after first correcting my setup to use my Ntrip source as my base station instead of my correction. Then I put in the actual location of the state provided correction tower into the (options->position->basestation). Then I was seeing satellites for both the rover and the “base station” but most of the rover satellites were not green. Also, I went into the monitor and was indeed getting error messages. I was receiving “no double-differenced residual”.