AC 3.2-RC1: Potential problem with SBUS failsafe

[Stefan Gofferje]

Hi,

a forum user reports some problems with failsafe and SBUS here:
http://ardupilot.com/forum/viewtopic.php?f=100&t=7793&p=18240#p18240

Maybe somebody could have a look at that before the 3.2 release? :)

-S

--
(o_ Stefan Gofferje | SCLT, MCP, CCSA
//\ Reg'd Linux User #247167 | VCP #2263
V_/_ Heckler & Koch - the original point and click interface

[Peter Plischka]

Hi

I can confirm that with FrSky x6x and X8R receiver with the TARANIS failsafe does not work with S-Bus connection.

.

regards Peter

[Jesus Alvarez]

Have not tried AC3.2 but with Pixhawk 3.1.4 and Turnigy 9XR with FrSky XJT 2.4Ghz for JR + FrSky X8R 8/16Ch FAILSAFE works OK after configuring receiver.

[Xian06]

Failsafe working but need ARMED to test it. (New feature in 3.2-rc1)

[Peter Plischka]

Then you would have to remove the propeller to the test, or is there a parameter or # define

for testing without arming?

regards Peter

[Randy Mackay]

     So the reason that some people need to arm the copter when testing the throttle/radio failsafe is because a growing number of receivers simply stop sending signals to the flight controller in a failsafe situation.  Older receivers instead pulled the throttle to a low (predefined) value like 960.

     With the older receivers you could quickly check that the radio's failsafe was working by looking at the mission planner and checking that channel 3 dropped very low.  This was never a complete check though because it was only testing the receiver.  If you check out the wiki page you'll see it goes through four other checks including arming the copter to ensure that the failsafe is working.

       Throttle Failsafe | ArduCopter

			Throttle Failsafe | ArduCopter Content Throttle Failsafe When it will trigger and what will happen Receiver and flight controller Set-up Testing for correct operation of Throttle Failsafe How the...
			View on copter.ardupilot.com	Preview by Yahoo

     With the newer receivers you cannot see easily when it's triggered because the signal simply stops but the other four checks on the above page will ensure it's working.

      "Armed" does not necessarily mean "battery plugged in and motors spinning".  Connect the APM/Pixhawk to your computer via USB but simply don't connect the battery.

     The behaviour hasn't changed in AC3.2 by the way.  It's mostly unchanged since AC3.0.1.

     The wiki hasn't been updated with the information on these new receivers that simply stop sending updates but I'll add that sometime between now and when AC3.2 goes out (although it's unrelated to AC3.2).

-Randy

--
You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Peter Plischka]

Hi Randy,

 

yes, it works. (PIXHAWK, X6R and TARANIS)

When I switch to stabilize mode, arm motors and raise throttle to the mid point then turn off the Transmitter the Flight Mode switch to LAND. Fine.

But at the same time MP shows also disarmed.
Would then the motors are switched off?  

 

regards Peter

[Randy Mackay]

Peter,

 

     Thanks very much for testing.

 

     I suspect that it’s switching to LAND first and then disarming.  So once it’s in LAND mode it will disarm the motors as soon as it thinks it’s landed and the throttle is at zero (or below the failsafe level).  The land detector is quite conservative so I’m pretty confident that it will never think that the vehicle is landed when it’s not.  I imagine if you waved the copter around a bit after turning off the transmitter (i.e. simulate that it’s flying) then I think you’ll find that it stays in LAND mode a while before disarming.

 

     By the way, I have an SBUS tx/rx now so I’ll give it a try as well.

 

-Randy

--

[Peter Plischka]

Wow Randy,

so it is. When I move the copter before switching off, it will remain armed.
When I put it on the table, he is disarmed.

That's a trick. :) That is certainly not in the wiki.
Thank you.

regards Peter

[Holger Steinhaus]

Hi Randy, 

     So the reason that some people need to arm the copter when testing the throttle/radio failsafe is because a growing number of receivers simply stop sending signals to the flight controller in a failsafe situation. 

 

Do you have an example for that type of receiver? I have done lots of debugging and analysis with S.Bus and have not seen this behavior yet. 

Perhaps there is still some confusion between "stops to send a signal" and the explicit out-of-band failsafe signalling as implemented by S.Bus using an explicit failsafe flag? Until latest master, ArduPilot unfortunately ignores this flag and relies on a timeout-based approach looking on channel data only. PX4Firmware is delivering all needed data, a fix for ArduPilot (together with a more robust channel count evaluation) is available here: https://github.com/diydrones/ardupilot/pull/966/commits . Flights tests, I guess 50-100 flights,  with real and provoked failsafe situations were positive so far.

BTW, for my personal convenience I use the suggested change together with a small patch that brings back a true throttle failsafe (based on the failsafe flag only) - so I am able to do easy and intuitive failsafe testing using the MP channel screen (can be found in my private repo: https://github.com/hsteinhaus/ardupilot/tree/flight_test4 ). 

Holger