GPS yaw let the drone arm in loiter mode when RTK is not fixed but it was fixed before

ArduCopter Copter 4.1
1

Good evening everyone,

We discovered something today running a ground test, we were able to arm the drone in loiter mode when the GPS has not RTK fixed (status 6). Usually, it does not let us arm since when the base F9 GPS is not in status 6 there is no yaw correction. The sequence of steps we followed was:

  1. Let the drone get RTK fixed without arming.
  2. Cover one of the GPS, the base, for making it lose RTK fixed.
  3. Try to arm.

We did it a couple times and we were able to arm in the second try as shown below (CH7 is our arm/disarm):

image
image1649×601 194 KB

It may be clear with Mission planner since APM planner plots the two GPS status as if they were the same type of data, so it looks as switching:

image
image1633×504 43.4 KB

We actually saw this behavior in 4.0.7 version previous to a flight, so we ran the ground test wondering if it also can happen in 4.1 version. It seems to be the case. Please find the log below:

2

@sbaccam,

Thanks for the report. Definitely if we are using GPS-for-yaw and the GPS has never reported a good heading an arming check should trigger. It is a little less clear in the case of a temporary loss of GPS yaw because gyros can be used to keep the yaw valid for quite some time in many cases.

I’ll discuss with @tridge. Txs again.

3

@sbaccam I’ve opened a PR with an arming check for yaw being available within the last 15s:


I’d also note that you have GPS_AUTO_SWITCH=0 and GPS_PRIMARY=0. I suggest changing that GPS_AUTO_SWITCH=1 so the beta6 firmware can use the moving baseline RTK base GPS (2nd GPS) as the primary, which is the right thing to do for beta6

1 Like
4

Thank you @rmackay9 and @tridge for your help!

1 Like