ianwood
Taco Wrangler
The answer is simple: there is no GPS flyaway. I have seen no credible theory or evidence to support its existence. Someone prove me wrong.
Flayaway theory
- Thread starter SteveMann
- Start date
- Joined
- Dec 18, 2014
- Messages
- 8
- Reaction score
- 1
I use this procedure to give me some peace of mind. I check the find my Phantom page and make sure it shows it where it's supposed to be. With me right next to it.
I have no proof, just a speculation of how erroneous GPS data could cause a flyaway.ianwood said:
I wish that DJI provided some kind of data logging that could be analyzed. Some flyaways are relocated later and the data could be quite revealing.
- Joined
- Mar 30, 2014
- Messages
- 173
- Reaction score
- 9
The following is not a theory, but an actual account of a fly-a-way.ianwood said:
I believe that taking off in GPS mode without acquiring enough satellites there can be a fly-away. I have witnessed it. Here is the the account of what happened and when.
8/30/2014
mm/ss
0:00 Power up the Phantom 2. Controller in GPS mode. Version 1.08.
0:27 CSC invoked BEFORE satellite acquisition has completed.
0:36 Full power and take off
0:37 With only an up Throttle Up stick used, the Phantom 2 flies away heading East
0:48 Controller's Mode is switched from GPS mode to ATTI mode to begin recovery of fly-a-way Phantom 2.
1:15 Phantom is returned and landed in ATTI mode.
A few minutes later after a restart, the Phantom 2 is flown successfully in GPS mode after the appropriate start-up sequence is followed and CSC is performed after enough satellites are acquired.
How would you describe the cause of this fly-a-way?
Last edited:
Zinnware said:
One possibility: Immediately after takeoff, the quad acquires sats and its xy position is right there where you are. Still has 0,0 as its home point and Its previous position. This is just like grabbing the quad and moving it away from its previous position. When you let go it will fly back to its previously stored position (where you grabbed it). In this case, that position is (0,0). Bye bye phantom. Solution: Switch to atti which no longer uses GPS positions and come home to papa.
EDIT: This is why you always wait for a home point if flying in GPS. If you take off and fly in atti, then you don't necessarily have to wait.
ianwood
Taco Wrangler
SteveMann said:
I do too! There was promise of the iOSD mini recording this data but it never materialized sadly.The OSDII does record CAN messages to an internal SD card that can be downloaded. The files are not easy to read but there are software libraries to decode them. I've written some scripts to decode them. There is an abundance of data, everything from individual cell voltages to motor RPM to individual stick positions. It broadcasts the home point lat/lon/alt and operating mode on an ongoing basis.
A DVR of the FPV/OSD is a good analog "black box" that provides some important data of the flight especially in case of something going wrong or loss of the craft. I use one on all flights. When something unusual happens, I will go back and review it. I find it's usually something the operator (me) is doing that explains the unusual event. Other times, it could be a bad battery or a loss of GPS signal near tall buildings.
Zinnware said:
Everything is a theory until it is proven otherwise.
Zinnware said:
I have done this several times on most versions of the firmware starting with 1.08 without any adverse issue and on more than one Phantom. It only means the home point doesn't exist at take off. The home point, if one is recorded, is somewhere in the flight path after take off. I would also argue that if the Naza had a defect that caused a flyaway when the home point wasn't recorded before take off, we would have seen many many many reports of it.
Zinnware said:
Do you have a video of this that we can see?
A theory, based on fact, is that when your Phantom is out of range of your transmitter, it will try to return to the home point. However, if the battery voltage gets too low, it will descend and land, wherever it is.
Now, this isn't a visible fly away, but it could explain some of the cases. In any case where you are out of visable range, before you attempt such a flight, get a GPS tracker!
OT, but I have been involved in discussions about parachutes. Some reading in MultiRotor Pilot and Rotor Drone magazines, turned up two companies made such an item. One design is for a different drone and comes with it. Another company sells them in varying sizes. The smallest is 36" in diameter (plenty big). and darn it if they didn't use exactly what I had planned!! When a problem is sensed, it kills all rotors, a solenoid releases the compressed spring and out goes the chute, attached to several arms. Bummer...
Now, this isn't a visible fly away, but it could explain some of the cases. In any case where you are out of visable range, before you attempt such a flight, get a GPS tracker!
OT, but I have been involved in discussions about parachutes. Some reading in MultiRotor Pilot and Rotor Drone magazines, turned up two companies made such an item. One design is for a different drone and comes with it. Another company sells them in varying sizes. The smallest is 36" in diameter (plenty big). and darn it if they didn't use exactly what I had planned!! When a problem is sensed, it kills all rotors, a solenoid releases the compressed spring and out goes the chute, attached to several arms. Bummer...
A theory has been extensively tested and is generally accepted, while a hypothesis is a speculative guess that has yet to be tested.ianwood said:
I think you might mean the latter ?
ianwood
Taco Wrangler
Possibly but my goal is really to flag false conclusions. When it comes to flyaways, there are way more conclusions than there is evidence to support them.
ianwood said:
I agree with your sentiment... though in a lot of this conversation it has been far from conclusive - I think the conversation has been a lot of "what could theoretically cause a Phantom to suddenly fly away" - It's all been pretty interesting to read.
I have never experienced a fly away (knock on wood I guess)... but considering the many thousands of these things that are produced, the cutting edge tech, human programming (I am in software development myself) - I imagine there is some possibility that catastrophic failures of some sort are possible (in not exceedingly rare).
I agree that the most likely culprit of a theoretical fly away is some error in home point setting by either pilot error or really rare bug (if it weren't rare - the boards would have exploded with people experiencing the exact same symptoms and a class action lawsuit - because everybody sues
) anyway - I have found the discussion interesting on an intellectual level.
ianwood
Taco Wrangler
ianwood said:
I don't know why this bothers you so much - maybe the word I should be using is "hypothetical"?? - since you really seem to be hung up on the literal version of theory (or the word theoretical).
My entire interest in the matter is an intellectual one.
IF flyaways EVER occur...
what would one look at as a possible systemic culprit?
In a system that is driven by GPS, failsafes and manual piloting I would look at the following systems:
1. Pilot error - human error always being a possibility
2. The GPS system - what can "hypothetically" go wrong with a gps system to cause the craft to fly out of control
3. The fail safe system - what could "hypothetically" go wrong with a system designed to take over control of an aircraft
The reason I have narrowed it down to the three things above is because the hypothetical "fly-aways" I have most heard discussed are situations where the pilot (in perception or reality) loses control of the aircraft and the aircraft travels in a way that doesn't match the expected behavior.
The expected behavior of the aircraft is to return to home at a loss of signal.
Why would an aircraft hypothetically NOT return to home?
One possibility MIGHT be is the pilot did something wrong.
One possibility MIGHT be that when the aircraft entered the return to home routine - it didn't know where home was? OR it didn't know where IT was.
None of this is based in fact (which is why you might be bothered by the word theory) - however - I was really trying to discuss the hypothesis the OP was discussing. Maybe I didn't make that clear.
Hope I am not stating the obvious, but one other theory that some radio experts have offered, is the phantom flies into a space where there is a much stronger 2.4ghz radio signal or even another totally different frequency signal that is so strong, it overwhelms and confuses the phantom, and it then reacts erratically. I have no proof to back it up, but it would make sense, in some flyaway scenarios.
In many of the possible causes though, I would think that many problems that could occur, are known, and could also be detected in more robust coding, and many of these problems could have been avoided or filtered out, in the first place.
In many of the possible causes though, I would think that many problems that could occur, are known, and could also be detected in more robust coding, and many of these problems could have been avoided or filtered out, in the first place.
Yep. I've flown very close to long distance wifi towers without a problem.
I'd be more interested to see the inside of the new compass and I can't find a decent picture that would show if the cables are now shielded. I really doubt they updated it because of moisture or whatever.
I'd be more interested to see the inside of the new compass and I can't find a decent picture that would show if the cables are now shielded. I really doubt they updated it because of moisture or whatever.
