Warning! Don't Try This At Home Kids (radio fail safe test disaster)

I've been building and flying drones since October of last year without any major incidents, at least not until today.  I have a Flamewheel 450 running a pixhawk FMU a Frsky Taranis transmitter and a Frsky Delta 8 receiver - it works great and I have lots of flight hours on it.  I have tested the radio fail safe on the ground several times and it all seems to work as advertised.  So is says to myself, I wonder if the radio fail safe really works in the air?  Seems like a reasonable thing to test.  Not that I would every get far enough away to loose radio contact given the power of the Taranis and the new FAA regulations on VLS flight.  I've been out 500 meters with no problems, so hey, why not give it a try - Bad Idea! 

I understand exactly what happened - now.  In fact, I had one of those "Oh Crap" moments of awareness as by beloved copter was free falling from 20 feet up with the motors off.  Here's what happened. 

First, let me say that the Pixhawk did EXACTLY what it was advertised to do in a radio fail safe situation.  So no problem there.  The Frsky Taranis is an awesome transmitter for sure, but it has this feature on power up where it checks that all the switches are in the down position and the throttle is off before it will turn on the RF.   I was flying it Position Hold mode (middle mode switch position) when I started the test by turning the transmitter off.  Sure enough, the copter went into RTL mode and started to come home.  Imagine my joy.  Everything would have been fine except I tried to regain control by turning the transmitter back on (you know what happens next don't you).  The transmitter started complaining, verbally, that the switches were in the wrong position and the throttle was not down.  Again, exactly what it is advertised to do.  So, I pulled the throttle down and put the mode switch back to it's down position which was Stabilize mode.  That was the fatal move.  The receiver woke back up and sent the new mode to the Pixhawk which forced it out of RTL mode (as advertised) and put it into Stabilize mode (as I commanded) with the throttle off. 

The rest is history.  The bill of materials for my 20 foot free fall was not that bad considering.

  1 3DR telemetry radio; 2 props; 1 motor boom; and 1 GPS mast.  

A bunch of other stuff came apart, but it was designed to do that so no harm.  This would make an interesting study in energy distribution in random events.  Pieces were spread out all over the place with the furthest piece, my RunCam 2, landing a good 15 feet from the crash site.  The good news is it was still running.

Bottom line - DO NOT live test the radio fail safe mode with a Frsky Taranis transmitter - Crap!

  Doug

Views: 1848

Reply to This

Replies to This Discussion

Thanks for sharing i had a good read and thanks to all commenters as well :) I just love this community!

Rich,

I had disabled the switch checks already, but I think I like your suggestion better to leave all the checks on and just hit any button on startup to override the checks.  That makes more sense to me.  I also thought about making Pos Hold the default.  I like that suggestion as well.  Thanks for the good advise.

   Doug

Rich Beckwith said:

Doug, 

You have a couple of options with the Taranis radio:

1.) After turning on the transmitter with the throttle above zero you should be prompted for "Press any key to skip".  This will allow you to continue with whatever switch positions and throttle state you had before turning the transmitter off.

2.) In model setup on page 2, "Throttle State" and "Switch Positions" can be turned off.  I personally prefer to leave these enabled, so I am sure everything is as it should be on transmitter startup - no surprises.  

Consider making your default mode PosHold with all switches in the default startup position.  I use SD as my mode switch (PosHold,Circle, Stabilize), and SA (AltHold, Break) with SF as RTL.

Don't forget that your receiver has a fail-safe mode as well which can be managed from page 2 of your model setup.  I always set "Failsafe mode" to "no pulses" for APM, to make absolutely sure the FC knows it has lost communication and needs to initiate an RTL - or whatever you have setup as failsafe.

Good luck.


Rich

Adding more safety logic, does not always translate into actual safety.

Sometimes having your equipment just do what you tell it to at all times, can be a nice thing to have...

" Bottom line - DO NOT live test the radio fail safe mode with a Frsky Taranis transmitter - Crap!"

I wouldn't go so far as to say that.  I've tested this same fail safe with a Taranis X9E and a FlySky TH9X, but the difference between your testing and mine was I allowed the test to run to completion.  After all the purpose of the test should have been to verify that the entire FAIL SAFE worked which includes two flight mode changes, one to RTL, and another to LAND.

What I would have said is, "Do not abort this test."

Doug it was really great of you to post this.  Lots of people wouldn't be aware of this and a lot certainly wouldn't post about it if they had made the same mistake.  A real education for all as I reckon everyone has had the same thought about trying the RC failsafe.  Thanks!



Clifton Hipsher said:

" Bottom line - DO NOT live test the radio fail safe mode with a Frsky Taranis transmitter - Crap!"

I wouldn't go so far as to say that.  I've tested this same fail safe with a Taranis X9E and a FlySky TH9X, but the difference between your testing and mine was I allowed the test to run to completion.  After all the purpose of the test should have been to verify that the entire FAIL SAFE worked which includes two flight mode changes, one to RTL, and another to LAND.

What I would have said is, "Do not abort this test."

Agreed, in the end, its more of an user error than design error. But practice and experience, it's what makes you better.

I absolutely agree with Clifton.  Had I let the test run to completion (which would be the case if the transmitter really had failed) the copter would have landed and everything would have been fine.  Aborting the test is what created the crash scenario.  All of the hardware worked exactly as advertised which points out the need to read the fine print. 

I also appreciate  Grant's comment.  I'm sure lots of people have contemplated running this test but lacked the confidence to try it.  Now we know it really works, but you have to let it run it's course.

Thanks again for all the good feedback.

  Doug

I agree with Grant too, thank's to post, posting this things prevent others to avoid crashes so your bad experience it's usefull for others, I hope I can prevented many accidents reading post like this.

Reply to Discussion

RSS

Groups

Season Two of the Trust Time Trial (T3) Contest 
A list of all T3 contests is here. The current round, the Vertical Horizontal one, is here

© 2019   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service