Lost my drone...for 15 minutes

[J0ck0]

Hey I been on this non stop but I just wanted to drop in here and let you know it will be either later or tomorrow until I have something. I can see where and when you selected to GoHome. It looks like the a/c (a/c = aircraft) was given additional commands to return home but the a/c ignored them and continued on. I have not read any part of this thread so I don't know if I'm repeating stuff that you already know or have been told by others. I will say if I had to make a call based on what I have seen so far, it would be all surrounding the compass.

Btw, also just to let you know there is a couple of columns of data that I have seen numerous times but never had to use them for any reason. But I do not know what exactly they are suppose to represent. So I have sent out a email to another Guru to get his thoughts on what they apply to. I have a feeling these may provide some good insight. I have to take a 30 minute break and then I will research these two columns more if the response I'm waiting on hasn't came in at that point.

-Frank

thanks, Frank. i appreciate your help. here is a summary of the events as i have been able to piece them together:

the a/c took off left as soon as it got to about 6' off the ground (it looks like i used auto takeoff according to the logs, but i don't remember). i had read enough in the forums to know that it was going to be uncontrollable at that point. i pushed the left stick to give it some height to make it over some two story houses and anything else in the neighborhood (all other attempts to control the a/c failed), and then i hit RTH in the hopes that it would bring it home. it didn't, but it did stop the a/c from going left, rotated it to face home, and then begin to fly right at a high rate of speed (i believe i saw 57 mph on my screen). it flew almost a mile before i lost contact with it. i thought all was lost, so i jumped in the car and headed to the last known location to gather the scraps. it was at that spot that my controller reconnected with the a/c, and it appeared to begin to head home at that point. i cancelled the RTH and flew it back to me and landed it. after looking at the flight data from the controller, it looks like it drifted a hundred feet or so once i lost connection to it.

there are some pics of the flight path as well as a video of it in this thread somewhere. i don't have them here at work, or i would reupload them to this post.

 

[Youngbill]

Here is a screenshot of the flight. Weird thing is the straight leg at the end of the curl shows as an instant jump from one spot to the next (several hundred feet ) when watching the replay. The arc begins near home because I had hit RTH. It just turned toward home and then flew sideways the rest of the arc while pointing home.View attachment 23929

There's that half moon arc again! Every fly away that I've seen a track of, has that half moon shape. Not sure what it means but someone at DJI must know.

 

[Axidrain]

i seem to see a common theme where flyaways happen after a compass/imu recalibrations here.

 

[bobmyers]

i seem to see a common theme where flyaways happen after a compass/imu recalibrations here.

Could very well be, Always best to thoroughly test the P3 at close range before venturing out. Personally, after any kind of upgrade, recalibration or adjustments, I would test the bird with a couple of full battery flights just to make sure that everything looks right-- at least until all the R&D with accompanying upgrades is completed and proven.

 

[suck it up]

i seem to see a common theme where flyaways happen after a compass/imu recalibrations here.

Or, after NOT doing them...............

 

[2Flush]

I'd suggest using some strong fishing line to make a tether for your next 'flight'. Tie one end to the phantom, and the other end to a stake in the ground. If you can replicate this error I'm sure DJI (and other P3 pilots like us!) will be very interested.

LOL.. Sounds good, but?? Hahah

 

[Axidrain]

Or, after NOT doing them...............

uhh. if you've read around its mostly due to bad calibrations.

 

[Brad Pierce]

uhh. if you've read around its mostly due to bad calibrations.

Do you mean calibrations that return an error? Or calibrations that report ok but may not be...
It would seem that if someone got an "error" after cal'ing, they would prob. not fly and try another cal instead? Or do you think people are flying even tho they know the cal reported an error.

 

[Axidrain]

Do you mean calibrations that return an error? Or calibrations that report ok but may not be...
It would seem that if someone got an "error" after cal'ing, they would prob. not fly and try another cal instead? Or do you think people are flying even tho they know the cal reported an error.

Well thing is. These sensor calibrations are very sensitive to outside forces. non level ground/vibrations during IMU calibrations and an area with a lot of magnetic interference during a compass calibration will do more bad than good. it will show that everything is fine in the app but when you fly it all hell breaks loose lol.

Rule of thumb is, its better to do these calibrations in a controlled environment. Otherwise you might risk introducing anomalies to your already good settings.

 

[Brad Pierce]

Well thing is. These sensor calibrations are very sensitive to outside forces. non level ground/vibrations during IMU calibrations and an area with a lot of magnetic interference during a compass calibration will do more bad than good. it will show that everything is fine in the app but when you fly it all hell breaks loose lol.

Rule of thumb is, its better to do these calibrations in a controlled environment. Otherwise you might risk introducing anomalies to your already good settings.

Would you say that getting a "rc connection is not good due to interference" (paraphrasing) would indicate an IMU cal would not be advised in that space? Are there other indicators of "poor cal conditions" you could use to quantify the space your in wen doing a cal? Seeing that your saying the cal seems to go ok, but it is not. I guess it make me nervous to think the cal report is a "go" but in reality it is not... Im looking for cross checks you might do to get a good indication of the quality of the cal. thx

 

[Axidrain]

Would you say that getting a "rc connection is not good due to interference" (paraphrasing) would indicate an IMU cal would not be advised in that space? Are there other indicators of "poor cal conditions" you could use to quantify the space your in wen doing a cal? Seeing that your saying the cal seems to go ok, but it is not. I guess it make me nervous to think the cal report is a "go" but in reality it is not... Im looking for cross checks you might do to get a good indication of the quality of the cal. thx

Generally it will only show a warning if the readings are off from the baseline that it already has.
So if you calibrate again it will overwrite that good setting with the new one. If the new calibration data is messed up right from the start, everything will still show green after calibration because that's the new baseline.

 

[Brad Pierce]

Generally it will only show a warning if the readings are off from the baseline that it already has.
So if you calibrate again it will overwrite that good setting with the new one. If the new calibration data is messed up right from the start, everything will still show green after calibration because that's the new baseline.

So, by checking the IMU in the RC, and confirming the right numbers in the sensor readout would be the only way to know if your last IMU cal was a success? thx

 

[mark linzenmeyer]

Where exactly in Peoria were you
I'm from here and will avoid this area

 

[Axidrain]

So, by checking the IMU in the RC, and confirming the right numbers in the sensor readout would be the only way to know if your last IMU cal was a success? thx

Pretty much. you mod readings should be at
Gyro Mod: 0.00
Acceleration Mod: 1.00
Compass: 1400 -1600

or somewhere close around that.
Also it helps if you hover for a bit before actually taking off to see if your phantom is showing any weird behaviors.

 

[flyNfrank]

thanks, Frank. i appreciate your help. here is a summary of the events as i have been able to piece them together:

the a/c took off left as soon as it got to about 6' off the ground (it looks like i used auto takeoff according to the logs, but i don't remember). i had read enough in the forums to know that it was going to be uncontrollable at that point. i pushed the left stick to give it some height to make it over some two story houses and anything else in the neighborhood (all other attempts to control the a/c failed), and then i hit RTH in the hopes that it would bring it home. it didn't, but it did stop the a/c from going left, rotated it to face home, and then begin to fly right at a high rate of speed (i believe i saw 57 mph on my screen). it flew almost a mile before i lost contact with it. i thought all was lost, so i jumped in the car and headed to the last known location to gather the scraps. it was at that spot that my controller reconnected with the a/c, and it appeared to begin to head home at that point. i cancelled the RTH and flew it back to me and landed it. after looking at the flight data from the controller, it looks like it drifted a hundred feet or so once i lost connection to it.

there are some pics of the flight path as well as a video of it in this thread somewhere. i don't have them here at work, or i would reupload them to this post.

So are we looking to confirm; what took place, or are we looking for a error of some kind to show what happened should not have happened?

 

[J0ck0]

So are we looking to confirm; what took place, or are we looking for a error of some kind to show what happened should not have happened?

just trying to find out what happened and how i can avoid it going forward, and if i have a bad bird and need to send it in for repair/replacement.

 

[J0ck0]

Where exactly in Peoria were you
I'm from here and will avoid this area

olive and the 101. i had flown in that area 5 or 6 times before with no issue...just a weird night, i guess.

 

[J0ck0]

Pretty much. you mod readings should be at
Gyro Mod: 0.00
Acceleration Mod: 1.00
Compass: 1400 -1600

or somewhere close around that.
Also it helps if you hover for a bit before actually taking off to see if your phantom is showing any weird behaviors.

i always hover for 30 seconds or more before i fly. this one took off once it hit about 6'. nothing i could do to stop it at that point.

 

[J0ck0]

were you able to figure anything out, @flyNfrank ? thanks!

 

[flyNfrank]

were you able to figure anything out, @flyNfrank ? thanks!

Yeah I sure did. You did not calibrate the compass, the gimbal, or the R/C, after the IMU calibration. That's why when the Auto Lift and Land was activated, the a/c did not know where it in order to know what to do. Next time don't watch it fly away. Flip the Mode Switch on Upper Leftside of the R/C to the Middle and fly it back, or set it down if it is windy because there is no gps to help assist with stability.