Diagnosing injected rtcm messages from MAVProxy to Ardurover

I’m working on a new MAVProxy module to receive GPS corrections from an NTRIP server and send them to Ardurover on my Pixhawk4 (for further forwarding to the Ardusimple GPS which is on the Rover).

My code is based on the existing MAVProxy module “DGPS” including verbatim copy-and-paste of the method send_rtcm_msg(self, data) (here).

MAVProxy is running on a Raspberry Pi connected to Telem2 on the Pixhawk4 (and this connection works: I can send and receive messages from MAVProxy to Ardurover just fine). RPi has a cellular modem for internet connection, also working fine.

For what it’s worth, NTRIP corrections from Mission Planner, sent over RF, connected to Telem1 (i.e. not using MAVProxy), work just fine. Ardusimple GPS switches to RTK mode.

My MAVProxy module is connecting to the NTRIP server, receiving RTCM messages, and seems to be sending them to the Pixhawk4. However my Ardusimple is not switching over to an RTK mode. The messages are either (a) not getting delivered to the Ardusimple, or (b) have incorrect contents in them somewhere…

So, question is, how should I diagnose, on the Pixhawk/Ardurover, whether (a) it is receiving the messages coming in from the MAVProxy on Telem2, and (b) whether it is forwarding them to the Ardusimple. There has got to be a debug log or setting screen somewhere where I can check whether these messages are arriving, and whether they are being forwarded, but for the life of me I can’t find it.

Thanks in advance. When I get mavproxy_NTRIP.py module working I’ll submit it for inclusion in a future release of MAVProxy…

Chris

@njoubert @tridge @tridge2 @rmackay9 @stephendade @mirkix

1 Like

I guess there must be something different about the messages that MP is sending vs what the RPI is sending. I would probably try and intercept those messages from MP and the RPI and compare them to excruciating detail. I’m not sure how you’d intercept the messages though. I guess MP stores the tlog on the PC and perhaps MAVProxy on the RPI does the same.

Maybe @peterbarker or @tridge have other ideas…

1 Like