RTK using Ublox evaluation kit

[ktrussell (Kenny Trussell)]

I have worked extensively to get the C94-M8P working with Pixhawk and have it working flawlessly now but boy, did I struggle for a while!

A couple of things to note:

1. The C94-M8P base to rover RTK (RTCM3) communication uses the serial transmit on the base’s M8P and serial receive on the rover’s M8P. So, there is no way for the Pixhawk to talk to the M8P. If you connect the Pixhawk transmit to the M8P, you will be clashing with the RTK comm taking place on the same line. So, you cannot connect anything to the C94-M8P J8 pin 9 or pin 14 IF you are using the built-in 915 MHz radios. Another small thing: I am using J8 pin 13, not pin 10, but either should work (refer to schematic of C94-M8P).

2. Baud Rate: The Pixhawk tries to set up the GPS it finds on either the GPS port or Serial 4 by sending it commands. Since the C94-M8P will not recieve those (because of 1 above), we just have to have our board set up the way the Pixhawk WOULD set it up if it could. I hope that makes sense. AND I found, using a PC with a serial-to-USB adapter to snoop on the lines between a more standard GPS (ublox M8N) and the Pixhawk and then between the C94-M8P and the Pixhawk, that even with the baud rate set to 19,200 in the config in Mission Planner for Serial4 on the Pixhawk, where my C94-M8P was connected, the Pixhawk was sending out 115,200 baud. So, it must be expecting the GPS to default to that. So, you have to have that baud rate coming in. I tried a couple of things, but eventually disable the onboard radios of the C94-M8P and went with a LoRa solution.

The draft at Using U-Blox C94-M8P with Ardupilot on a PixHawk Rev 2 gives some info on the above. Also, my YouTube channel (https://www.youtube.com/channel/UCkIJQFyZdzLKh7AR-m3t2Bw) might help you. I am still a newbie and benefiting from this forum so much. I hope I can give back a portion of what I am finding.

You might read this thread as well:

Bad GPS Health when I have RTK Fix

I mistakenly thought that the other GPS I have connected had the same delays, but I was looking at the same graph by mistake! I see now that GPA.DELAY has the expected 200 mSec +/- 20 mSec very perfectly. Only the M8P (above on GPA2.DELAY) had the wild swings. Attached is the well-behaving GPS graph, which is a Ublox M8N - NO problems with it [gpa].

I will be happy to share code, etc. if you decide to go with the Adafruit Feather M0.

I look forward to hearing how your project is going.