I just upgraded my pixhawk to the latest firmware 3.2 from 3.15. I have and all 3dr flight control setup including ppm and telemetry and speckrum radio and receiver. I went through the hardware set up including setting up the failsafes. I set the Throttle fail safe to 1018 as the radio power off is 1009 and the the throttle at its lowest when the radio is on reads as 1126. I checked to see if this was working by shutting my radio off with the props off and then me wondering about the apartment with the copter in a flying state. I have 3dr telemetry running and my laptop running mission planer. Every time I shut my radio off I would get the message that my RC connection was lost and that RTL was engaged. The problem is A few times this message was followed by the motors switching off when I was “in flight”. I have attached a log and the problem occurs near the end of the log.
There is a known glitch that looks exactly like what you reported. However, it was reportedly only with Taranis RX. Not sure if it’s a coincidence… I don’t fly Spek now.
[quote=“mabo 2014”]I just upgraded my pixhawk to the latest firmware 3.2 from 3.15. I have and all 3dr flight control setup including ppm and telemetry and speckrum radio and receiver. I went through the hardware set up including setting up the failsafes. I set the Throttle fail safe to 1018 as the radio power off is 1009 and the the throttle at its lowest when the radio is on reads as 1126. I checked to see if this was working by shutting my radio off with the props off and then me wondering about the apartment with the copter in a flying state. I have 3dr telemetry running and my laptop running mission planer. Every time I shut my radio off I would get the message that my RC connection was lost and that RTL was engaged. The problem is A few times this message was followed by the motors switching off when I was “in flight”. I have attached a log and the problem occurs near the end of the log.
Only test that failsafe worked. I don’t think RTL will work if it’s not really in the air. It will sense that it is not moving and think it has landed and shutdown. You should see an event that says Landed.
[quote=“iseries”]Only test that failsafe worked. I don’t think RTL will work if it’s not really in the air. It will sense that it is not moving and think it has landed and shutdown. You should see an event that says Landed.
sorry I just realized the log failed to upload as it’s to large. I’ll try and replicate and produce a smaller log file.
I’m not sure if i was moving every time the condition occurred and do know that it did occur when I was not moving. It does not however usually shut the motors off when it’s not moving and looses contact, it goes into landing mode if not moved at all.
Thanks
I wrote this before seeing bills post,
Hi thanks for the responses. My experiences have been that the copter goes in to RTL if I’ve moved it from where I initialized, otherwise it goes directly into landing mode, not motor shutoff. The motor shutdown happens about 15% of the time. I tried the procedure with telemetry connected to my tablet instead of my laptop and it happened on the first try but I couldn’t get it to repeat. I was hoping that maybe the telemetry connection to the tablet would somehow prevent the condition, it didn’t. The next thing I’m thinking of is calibrate the DX9 throttle channel so that the failsafe generates a value below the default setting of 975 and reseting the flight controller failsafe trigger back to 975.
I just repeated the problem again with the copter definitely moving when the radio was turned off. I turned the radio off then immediately there was a tone from the pixhawk and the motors shut off. I have been able to observe the proper RTL procedure as well with the copter going into RTL, me placing it were it took off from, it then goes into landing mode the motors slowly throttle down a tone sounds and they turn off. I have attached another log with the event at the end.
I cannot see anything strange in the logs. I do think what is happening is that it think it has landed or not even taken off. It doesn’t even go more than 40cm in the air.
If you really want to test the ‘logic’ of these actions the best thing to do is to run it in a simulator. Obviously to avoid smashing up lots of copters this type of testing is done in a program called SITL (Software In the Loop) It runs the exact same flight code that on the autopilot, but software simulates the the sensors. i.e. you command the motors to increase and the sensors reports gains in altitude etc…
For real world tests, you can do this using a ‘beater’ quad and don’t attach an expensive camera.
The other thing besides consistency is the speed at which this occurs, it’s like the power is switch is the disarming switch, usually the automated commands are relatively slow. I tried it again this time arming on a table and then holding it at about eye level and was starting to get comfortable, (I repeated the process about 6 times) but it happened again. I’ll include the log. Do you think I should try to calibrate the DX9 throttle channel so that the failsafe generates a value below the default setting of 975 and resetting the flight controller failsafe trigger back to 975, currently I set the Throttle fail safe to 1018 as the radio power off is 1009 and the the throttle at its lowest when the radio is on reads as 1126. Do you thing that’s worth trying?
I’ve been meaning to set up the simulator, I’ll give that a try but it seems a bit of a project.
You could try raising the failsafe trigger to 1050 if you think that your receiver is not alway getting below 1018. Setting 975 and your TX to be lower will make no difference to the logic.
I think the problem may be related to what mode you are in when the Thr FS occurs. I don’t fully understand it yet, but there’s an issue where the Rx Failsafe PWM should be set to select Flight Mode 5, and then FM 5 should be Loiter or Alt Hold. There is some logic whereby, if the Thr FS occurs in Stabilize, that it will just shut the motors off.
I don’t see any indication in the logs that the copter thinks it is landed. Because normally it logs that as an event I believe.
FYI:
I currently my modes are set to mode 1. Stabilize, mode 2. alt hold, mode 3. Pos Hold, mode 4. RTL, mode 5. RTL and mode 6. RTL. I did this because I have a mixed setting on my remote that sets one 3 position switch to do the 3 modes and one two position that toggles RTL. I set the unused modes to RTL because the value of the two switch in various combinations will trigger mode 5 or 6 or both I’m not sure at the moment. Hope this helps.
