Runaway phantom

[Tito]

Let it idle in manual mode. I doubt it thinks it is still airborne because in fail safe it can auto shut down after idle.

Anybody who has been holding his phantom at idle knows it's still responding to movement. I think that when it's on the ground, the compas, accelerometers and GPS are still enabled. So if it would lose a satellite or the accelerometer / compas is bugging because there's no input, it would try to compensate and indeed the friction of the skids cause it to tip over.

 

[TeamYankee]

@SpiritsKeeper - you hit the nail on the head..
The Phantom doesn't know it's landed.

On take off the barometric pressure is recorded as the "altitude above ground" - at take off it's obviously 0 feet - in aircraft radio speak it's called QNH.
(The altimeter setting in aviation. QNH is the barometric altimeter setting that causes an altimeter to read airfield elevation above mean sea level when on the airfield. In ISA temperature conditions the altimeter will read altitude above mean sea level in the vicinity of the airfield.)

As the Phantom climbs the pressure drops and this is calculated in a change of altitude in feet, as displayed on IOSD etc.
Since the barometric pressure is constantly changing and the accuracy of the equipment is not "super accurate", when it lands and you leave it idling any change in pressure/wind etc will result in an altitude change, even a tiny one... and it will appear to rise up a little.

When it's down, rule of thumb is to hold down the left stick and it will power down.

 

[BlackTracer]

@SpiritsKeeper - you hit the nail on the head..
The Phantom doesn't know it's landed.

On take off the barometric pressure is recorded as the "altitude above ground" - at take off it's obviously 0 feet - in aircraft radio speak it's called QNH.
(The altimeter setting in aviation. QNH is the barometric altimeter setting that causes an altimeter to read airfield elevation above mean sea level when on the airfield. In ISA temperature conditions the altimeter will read altitude above mean sea level in the vicinity of the airfield.)

As the Phantom climbs the pressure drops and this is calculated in a change of altitude in feet, as displayed on IOSD etc.
Since the barometric pressure is constantly changing and the accuracy of the equipment is not "super accurate", when it lands and you leave it idling any change in pressure/wind etc will result in an altitude change, even a tiny one... and it will appear to rise up a little.

When it's down, rule of thumb is to hold down the left stick and it will power down.

I have to respectfully disagree with the bolded statement above. And feel free to shoot me down if I am way off about this.

The Phantom does know if it has stopped descending while a command to descend is present (The Phantom definition of "landed"). Whether it is on the ground or on top of a roof or in your hand. When it reaches this state we all know it powers down the props to idle speed and allows the left stick fully down to shut off the motors instead of descend at maximum rate. I am sure it uses the barometer to detect this since that is its way of measuring where it is on the z axis. If it didn't know it has landed how would it do all this?

As far as it spinning back up above idle after landing, I could see at least one way of that happening. i.e. If you left it idling then turned off your controller it may try to RTH. I think it says this in the manual as well.

Maybe I have misunderstood what you meant by it not knowing it has landed.

 

[N017RW]

OK, I see it a little bit differently.
Saying the Phantom knows anything is an over-statement of it's logic.

In order to allow shut-down with left stick only it must not sense any changes in at least the 'Z' axis, a simple logic function. It may be Baro., Accel., Gyro or all of them summed. If you want to call this 'knowledge' well then to each his own.

If it knew it landed it would not allow the slow motor ramp-ups that are known to cause tip overs.
Since the a/c is static on the ground the IMU and other inputs to the FC are also static +/- any system errors, noise, drift, etc.

Motors are set to min RPMs when the stick is released (centered) and on the ground cause there's no need for more, they only spin as fast as they need to to maintain position, on the ground it's like a perfect hover. Once sensor drift begins to occur, a/c motion is associated with this so then the FC begins trying to correct with control outputs (motor RPMs)... Eventually causing a tip-over. GPS position error/accuracy may also play a role in this, I'm not sure.

How does the a/c 'know' you've decided to land 'full-stop' and not a touch and go"?

It doesn't cause it doesn't know anything, it's just executing a logic(al) flow chart.

 

[Dalite]

I thought the RAP firmware had been corrected to eliminate unpiloted launch. Here is a discussion where the warning was made by one of the most experienced RC Heli pilots in the hobby.

http://www.phantompilots.com/viewtopic.php?f=4&t=7231

For background only. The warning post was originated by Mike Mas http://www.rotory.com/site/Rotory.html

Glad to hear this only resulted in prop damage.

 

[N017RW]

Well this 'report' is a couple of years old so it may have been fixed for described situation.
I can say my experience was that if you never advanced throttle after arming the motors (i.e. no flight) the motors would stop.
I'll test again ASAP.

We are talking here about landing (post-flight) and allowing motors to idle resulting in a slow gradual RPM increase leading to a tip-over. This has been my experience and that of others too.

 

[Dalite]

Well this 'report' is a couple of years old so it may have been fixed for described situation.
I can say my experience was that if you never advanced throttle after arming the motors (i.e. no flight) the motors would stop.
I'll test again ASAP.

We are talking here about landing (post-flight) and allowing motors to idle resulting in a slow gradual RPM increase leading to a tip-over. This has been my experience and that of others too.

I also thought it had been addressed. However, the NAZA truly doesn't know when it is in flight, and if RAP is enabled and it sees a situation where it thinks it is in flight at a hover (no change in motor speed) and seen no directional commands in a given time frame, it assumes radio comms are lst and starts it's routine.

So, just a thought, but that is what I would try to eliminate first as a cause. Too many similarities to overlook.

 

[N017RW]

Well this 'report' is a couple of years old so it may have been fixed for described situation.
I can say my experience was that if you never advanced throttle after arming the motors (i.e. no flight) the motors would stop.
I'll test again ASAP.

We are talking here about landing (post-flight) and allowing motors to idle resulting in a slow gradual RPM increase leading to a tip-over. This has been my experience and that of others too.

I also thought it had been addressed. However, the NAZA truly doesn't know when it is in flight, and if RAP is enabled and it sees a situation where it thinks it is in flight at a hover (no change in motor speed) and seen no directional commands in a given time frame, it assumes radio comms are lst and starts it's routine.

So, just a thought, but that is what I would try to eliminate first as a cause. Too many similarities to overlook.

I absolutely agree based on new/more info. about this. Again I recall the time-out for no stick input (dead-man) but did not consider that in this scenario. It may indeed be another incarnation of the same [root] cause.

You said: 'However, the NAZA truly doesn't know when it is in flight...'
I agree, and the same for 'landed' as well..

 

[deepeshdotcom]

Wasnt this fixed in latest firmware 3.12..

 

[kirk2579]

It is in the list of features /updates for 3.12

the OP was using 3.08 or earlier as it was prior to 3.12 update.

 

[N017RW]

It is in the list of features /updates for 3.12

the OP was using 3.08 or earlier as it was prior to 3.12 update.

Not sure I can agree.

Here's the notes:
http://download.dji-innovations.com/dow ... tes_en.pdf

Maybe the release notes are vague or ambiguous (imagine that).

We are (I'm) not talking about an 'automatic landing' here.

After a pilot initiated/controlled landing. If left to idle on the ground eventually the motors slowly speed up and a tip-over occurs.

After new info. posted here recently, I'm leaning to the dead-man timer causing the fail-safe to engage and that results in the motor speed-up leading to a tip-over.

 

[kirk2579]

It is in the list of features /updates for 3.12

the OP was using 3.08 or earlier as it was prior to 3.12 update.

Not sure I can agree.

Here's the notes:
http://download.dji-innovations.com/dow ... tes_en.pdf

Maybe the release notes are vague or ambiguous (imagine that).

We are (I'm) not talking about an 'automatic landing' here.

After a pilot initiated/controlled landing. If left to idle on the ground eventually the motors slowly speed up and a tip-over occurs.

After new info. posted here recently, I'm leaning to the dead-man timer causing the fail-safe to engage and that results in the motor speed-up leading to a tip-over.

I agree , The notes do say "automatic" I was missing that word in my recall of features /updates. Hopefully manual as well as auto are included but not noted. Either way I kill motors asap upon landing...