Doing my first APM build (3DR Y6) and got it to a nice stable hover and good vibration tests. I wanted to test the throttle failsafe and the bird immediately flipped. I've been troubleshooting the failsafe behavior on Mission Planner (yes, something I should have done more of first, lol) and need some advice.
I am using a PX4 autopilot and a Spektrum DX8 transmitter. I saw that the PX4 can be fit with a Spektrum satellite receiver, but the website says that's only for medium range applications. It suggested an Orange RX receiver if you want more range, since it is DSMX, but has an S.bus output. I used the S.Bus connector and mixed the Spektrum Ch 1 (throttle) to Ch 3 and did the same for elevator (3 to 2) and aileron (2 to 3) channels. Manual flight mode worked great with minimal roll and pitch adjustments.
What I think I see in the Mission Plannner failsafe tests is that, no matter what position the controls are in when I bind the Orange RX (both with and without pulling the binding plug before turning on TX), it always failsafes to Channel 1 = 0, and Channels 2-4 at last position. This probably put on a big roll in the moment before RTL took over, but I'm not skilled enough at reading the logs to be sure.
I think my best next step is to go back to the Spektrum satellite plugged directly into the PX4, so I don't have in intermediate receiver confusing things on loss-of-signal. All the Orange RX documentation says the failsafes are set at bind position, but either I'm doing it wrong, or Channel 1 is excluded from this programming.
Any other ideas?
Thanks!
Robert
Replies
Robert Perkins, I had exactly the same problem with my setup. The Spektrum satellite failsafes work perfectly, but i was only able to fly about 500m away, so i decided to fit an Orange DSMX receiver. All controls worked well, took it on my first flight, i was about 400m away and the hex fell out the sky. after further research i discovered like you did the failsafe did not work with the Orange receiver.
I have decided to go back to just using the satellite receiver as i think 500m is far enough and my video link struggles with anything beyond that any way so i cannot see much point.
it was just an annoying and expensive learning curve.
Hi Robert,
You have some crazy looking spikes in your logs. Any chance you're using a Mac/Apple? I've heard this can cause these artifacts in the logs.
You do show a spike on your pitch input, which could be the same thing I'm experiencing, but I'm actually not inclined to believe that at this point because they are impossibly big (400,000,000 degrees) and you are also showing them on the pitch and roll values, not just the inputs. Actually, you have the crazy spikes in all the data, so I'm inclined to say the log data is no good.