Preventing motor stop: pictures of what happens if motors are stopped midair..

During tesing of motor-off recovery at safe altitude (60meters), the quad fell stright down, - propellers were spinning backwards, and motors stalled, not being able to turn in right direction.

At much too low altitude, the quad rolled, making 2 of the motors able to spin up, but it was too late then.

-motors should be kept on at all time when the APM is armed, stopping motors is asking for dangerous situations..

I created a issue here: https://github.com/diydrones/ardupilot/issues/431

Feel free to support it.

Views: 2726

Reply to This

Replies to This Discussion

To recap and make certain we are not misinterpreting your test:

  • The quad was at altitude in Stabilized mode (hovering, ascending, ?)
  • You commanded full throttle off (!!) or did you command DISARM?(same throttle location)
  • The aircraft began to descend..(immediately -- probably pitching,rolling or a combination)
  • ...at some point...you commanded some value of throttle (half, full?) to arrest the descent
  • The quad was unable to spin up all the motors (allegedly) and recover to a level hover and crashed

Based on this experiment, you are petitioning the developers to implement a flight mode that completely compensates for any pilot action of shutting down the motors in flight? In short, you are asking for a 'smart throttle' function.

Interesting concept.

Nylon or other plastic screws may not be a bad idea if they do not get too hot and soften or deform and then twist. In the case where the arm is aluminum, some heat sink action of the motor base and the arm may delay this possibility.

-=Doug

In stabilize mode, throttle stick all the way down means motors stop.

Yes, That's what I did.

The quad fell stright down for ~2 meters before I applied a little , then more throttle.

While falling, I tried to go max throttle several times, but the initial fall made then spin in reverse, denying spin-up.

Being able to stop motors mid air in a multirotor is just dangerous, not useful at all.

The motors should start spinning once armed, and run until disarmed. - maybe 5% throttle should be minimum.

Being able to shut down motors in stabilize adds lots of possible disaster scenarios.

Imagine: flying Loiter or Land mode, throttle stick all way down to descend , then pilot discovers need to do it manually, or otherwise wish manual (stabilize) control.   - motors will stop, at low altitude that will result in a crash.

one were ripped off, not cut, - the rest I cut off to ease disassembly.

The option used to be there to either spin motors when armed or not.  Option seems to have been removed somewhere along the way.

Making the motors always spin when armed creates a whole lot of new problems.  You can't shut them down when things go horribly wrong, to reduce damage (both to copter and to people/property).  You have arming and disarming delay which may confuse people(hey why don't my motors start, oh now that I put my finger in they start), etc.

Furthermore you have the variable pwm value of "stationary spin' for different copters.  An overpowered copter may take off when spinning at a stationary pwm, while a normal one does not.  For some esc's the stationary pwm might be insufficient to make the motors spin.  For esc's without external oscillator the stationary rpm will vary with temperature, etc etc

The issue has been brought up  before on the developpers list.  But my opinion is that a pilot still needs some common sense and that a machine will never be able to replace that.

"Being able to shut down motors in stabilize adds lots of possible disaster scenarios." - like be ing able to turn the ignition of in your car while driving on the highway.

"Imagine: flying Loiter or Land mode, throttle stick all way down to descend , then pilot discovers need to do it manually, or otherwise wish manual (stabilize) control.   - motors will stop, at low altitude that will result in a crash." Unfortunately - been there, done that. Now I always make sure I leave my throttle at hover so if I do "panic out of auto" it does not fall like a stone. 

"In stabilize mode, throttle stick all the way down means motors stop." So when you crash (from something else) you would rather your motors continue to spin and chewing up more props, motors and burning out the ESCs because even though after you hit the deck you slammed the throttle closed but alas, it continues to break things.

There certainly could be something in the way of smart throttle but I think it still needs to yield to the pilot's wish - i.e. close throttle completely when he says so which then cancels out your main desire.

My suggestion would be to try an minimise damage from crashes. I have all motors held on with nylon screws. I cannot count how many motors I saved from that. Just on Saturday I was throwing it around in Acro and I had a spinner let go in a particularly hard maneuver at about 80m. Poor thing flick-flacked all the way down. All I could do was close the throttle (!!!) to save the motors after impact. It hit hard, snapped all 4 motors off the arms. I retrieved about an inch soil sample from the field out of one of the arms. :D 

Damage - 8 new plastic screws needed and a spinner top which I could not find. No props, no motors, no bent shafts. Works every time and I have only once had a screw let go during flight which is because I reused it after another crash. Lesson learned on that one - always new screw.

I think a feature which would be handy and along the same vein as this would be a "rapid descend" where the APM will bring it down as quickly as possible but without stalling the motors (tell-tale sign without feedback is uncommanded pitching or rolling)

I often descend for the fun of it with the motors just about stalling. I bring it down as quick as possible and as and when they stall I give it a bit of throttle. From an APM PoV it would be tough because of the instability caused by falling through your own  prop wash.

not a problem.

Idle PWM should, of course, be set manually.

regarding "shutdown to reduce damage"  - well, IF the multirotor already crashes, an idle speed will not cause more damage, as idle speed would not be enough to overcome any resistance if motor/propeller hits something.

of course the PWM should be set correctly, by the person that choses to enable this feature.

Mikrokopter worked like that for years

Actually always idle motor run when armed. Several other multicopter controllers do that. Works great.

It should just be a new paramater called idle speed. Set it to zero to have it behave like it does now or set it at the maximum speed that doesn't allow it to take off to have an idle speed. Shutting off completely would then be done by down left rudder.

I also heard there is a motor kill switch in the works for 3.1 so maybe will be better to implement along with that. This way we can kill in an instant in case of trouble.

I can see a new wave of fly-aways if the idle throttle is an unchecked user-configurable parameter.

If someone sets that too high there's no coming down without an in-flight disarm, which newbies won't know to do. Maybe it needs to be capped at something like 300, most copters need more than 300 throttle to hover.

Letting the props stop means an almost certain crash, so it should be impossible even in ACRO.

Why not skip the is-flying detection and just spin the props always when armed? That is much simpler, but for the pilot to understand and for the programmer to implement. The most bug free part of a computer program is that which is not there at all.

Reply to Discussion

RSS

© 2019   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service