Flying over marina's?

[SteveMann]

There's nothing in the referenced thread to "accept". Just a few anecdotal "it happened to me" posts.
"... cause I did see mine do it"
" ... people lose control of their Phantoms all the time because of their compass."
They are all anecdotal. I want to know *how* can bad compass data cause a flyaway. "Because mine did" offers no information to *how*.

 

[bobmyers]

All the compass does is to tell the Phantom which way it's pointing relative to North. A large metal item near the Phantom is only a problem if you are flying waypoints. (RTH is a waypoint flight). And even then the Phantom will eventually get to where it's going although on a circuitous route (toilet bowling). If your compass calibration is bad, but not bad enough to prevent a flight, you may see some squirreliness in holding course, but no one has produced the technical theory behind the claim that a bad compass calibration can cause a flyaway.

If north became south during calibration due to magnetic interference, then the Phantom could indeed fly south when it was supposed to fly north. In that case it my not ever RTH resulting in a fly away. IMHO

 

[SteveMann]

If north became south during calibration due to magnetic interference, then the Phantom could indeed fly south when it was supposed to fly north. In that case it my not ever RTH resulting in a fly away. IMHO

The controller (from my experience with the APM-2) would quickly detect that the distance to home is increasing and make a heading adjustment to correct for what could be wind. It does this calculation several times a second. It would eventually get home, but the battery would likely go critically low and the quad would land immediately wherever it is.

 

[ianwood]

I explained it, Steve. Vertical stability, 9DOF vs. 6DOF sensor fusion, relative vs. absolute positioning, accuracy of gyro drift / accel noise vs. gravitational vector, gimbal horizon loss during TBE events, yadda, yadda, yadda.

If you don't believe me, take my challenge. Unscrew/unclip your compass and loosen it so it will flop around freely. Grab a video camera and a cardboard box. Set the camera to record. Go flying.

Once you crash, about 30 seconds later, use the box to collect all the pieces of your busted Phantom.

 

[SteveMann]

Flopping the wind is not the same as a calibration error.
You didn't explain anything. You posted a lot of anecdotal "it happened to me, so now it's fact".
Nothing in the description of sensor fusion explains *how* can bad compass data cause a flyaway.

 

[bobmyers]

The controller (from my experience with the APM-2) would quickly detect that the distance to home is increasing and make a heading adjustment to correct for what could be wind. It does this calculation several times a second. It would eventually get home, but the battery would likely go critically low and the quad would land immediately wherever it is.

It is an interesting dilemma-- I can just see it flying back and forth trying to decide where the home point is-- and not ever reaching Home Point. Flying under the Influence of a compass error

 

[Alex_]

We've already been through this:

http://www.phantompilots.com/threads/the-compass-misunderstood-or-easy-scapegoat.39304/

Either you don't accept or you don't understand it. Or both.

Ugh, that string made my head hurt!

 

[SteveMann]

It is an interesting dilemma-- I can just see it flying back and forth trying to decide where the home point is-- and not ever reaching Home Point. Flying under the Influence of a compass error

It knows where the home point is, but without reliable compass data it will hunt for it a long time because it doesn't know which way to point.

 

[Meta4]

It is an interesting dilemma-- I can just see it flying back and forth trying to decide where the home point is-- and not ever reaching Home Point. Flying under the Influence of a compass error

It doesn't end up going back and forth.
What it does is to fly a big curve rather than a straight line.
This is known as toilet bowl effect.

 

[ianwood]

Aaaaaaaaaand, during toilet bowl effect (TBE), you will see that the gimbal can no longer stay level with the horizon. Why? Because it is receiving bad vertical orientation data. Why is the vertical orientation data bad? Because the compass is supplying bad data into the vertical stability algorithm.

If the compass data is bad enough, the Phantom won't just slowly rotate, it will accelerate into larger and larger circular patterns. The Phantom will lean heavily into the turn so much so that it will begin to lose altitude. Why is it leaning so much? Same reason as the gimbal losing the horizon. Bad vertical orientation data.

How can I be so sure? Because of flight controller priorities:

1. Maintain vertical stability using accelerometer, gyro and compass data.
2. Maintain altitude using barometer data.
3. Navigate using GPS and compass data.

If the compass was only used in navigation (priority 3), how could TBE cause sustained acceleration and loss of altitude? If maintaining stability and altitude take priority over navigation, the Phantom should never lean as much as is seen in strong cases of TBE. The higher priorities should prevent navigation from causing a sustained acceleration and loss of altitude.

That is unless the acceleration and loss in altitude are in fact caused by bad vertical stability which is the highest priority.

 

[Alan NZ]

No problems here

 

[skyhighdiver]

There's nothing in the referenced thread to "accept". Just a few anecdotal "it happened to me" posts.
"... cause I did see mine do it"
" ... people lose control of their Phantoms all the time because of their compass."
They are all anecdotal. I want to know *how* can bad compass data cause a flyaway. "Because mine did" offers no information to *how*.

Never a issue fling over any marina

 

[ELHEAPO]

I wish I never asked that question now!, trying to keep up with you guys is hurting my little brain...good discussion though!!

 

[markarch]

Not much to add to the debate other than no issues here as well....

 

[ianwood]

And to tie it all together, there shouldn't be any issue flying over a marina. Magnetic fields tend to be very localized in man made objects and have no measurable impact more than 50ft away. Maybe 100ft for larger objects like multi-story buildings. Electromagnets, large electric motors, or anything else with a magnetic charge would be notable exceptions.

I've posted these charts before. Both were fully manageable flights on a P2 with no issues. This is within 100ft of tall buildings:

This is out in an open field:

 

[JJPetrisko Video]

I know you have to be careful about calibrating around metal objects etc, my question is, is it ok to fly over marinas and boats (obviously at a safe height) as i have read stories about the compass getting interference near large metal structures etc leading to problems. If it is ok, what would you suggest is a good altitude bearing in mind i dont want to be to high away as the nearer you can get the better the video but i also dont want to be reckless and possibly lose my p3.

When I was in Turks and Caicos I flew over a few over Marina's and had no problems. They were about 3k feet away as well. My altitude was around 30 ft so I was able to get pretty close, not experiencing any problems.
Check out some video I shot, it include some Marina's.

 

[SteveMann]

I wish I never asked that question now!, trying to keep up with you guys is hurting my little brain...good discussion though!!

Don't let the grumpy old man here spoil your fun. Please ask anything - you are guaranteed a few hours of entertainment and we might learn something.

 

[Bret Lucas]

No interference at all operating from 50'-350'

 

[JohnK]

I fly near and above the Walkway Bridge here in the Hudson Valley all the time with no issues, and it is a HUGE metal bridge that was once a railroad overpass. Here's an early vid before I ever did any serious work with color correction or serious editing (which I am still learning how to do ... I've only been at this about a month)