Welcome to PhantomPilots.com

Sign up for a weekly email of the latest drone news & information

Professional P3P on "Go Home" seemed to be heading off away from home?

Discussion in 'Phantom 3 Help' started by GuernseyZoo, Jul 20, 2016.

  1. GuernseyZoo

    Joined:
    May 21, 2016
    Messages:
    7
    Likes Received:
    3
    Location:
    Guernsey
    Hi All,

    I had a bit of a fright yesterday when flying my P3P. I was taking some photos of boats moored in a local bay and when the battery level was getting low the drone seemed to go into automatic return to home mode, but instead of returning to home it appeared to be flying out to sea. I managed to resume manual control, and just about got it back to land with a bit of a bump in some long grass, one motor was briefly obstructed but I reacted with a CSC fairly quickly, and there doesn't appear to be any serious damage fortunately.

    I would be very interested if anyone could please help shed any light on what I might have done wrong during this flight, as I would be very keen to avoid making the same mistake again.

    If it helps, I have already uploaded the .txt log from the tablet and the .dat file from the drone to dropbox, they are available at these links:

    Dropbox - FLY065.DAT

    https://www.dropbox.com/s/mtifp0 ... 6-41-09%5D.txt?dl=0

    I uploaded the file to HealthyDrones - link: HealthyDrones.com - Innovative flight data analysis that matters

    and Phantom Log Viewer here: Phantom Log Viewer - PhantomHelp.com

    I don't think that I did anything different on this flight. I had just successfully completed two earlier flights with two other batteries in the hour before this flight from the exact same location, but I realise that on the balance of probability it is likely to be caused by something I have missed, and it might help me and others to avoid a similar situation if the reason can be found.

    The start of the flight which I captured on video seemed perfectly normal. I don't have any video of the Go Home section of the flight though as I had been taking photos of the boats on the way back.




    After the crash landing I swapped in my final spare battery (fully charged) and did a quick auto takeoff and auto landing which seemed perfectly fine, which was a huge relief!

    Any help in identifying what went wrong would be very much appreciated, thanks..

    Kind regards,

    Steve
     
  2. GuernseyZoo

    Joined:
    May 21, 2016
    Messages:
    7
    Likes Received:
    3
    Location:
    Guernsey
    I just upgraded my HealthyDrones account to be able to see the wind data for my flight, and I am no longer surprised that the drone flew the course it did.

    Even though it seemed calm to me at ground level, and the nearby airport at 335 feet above sea level was only reporting 8 to 12 knot winds, the wind data shown in my flight logs shows an average of 20 miles per hour, with one gust exceeding 29 miles per hour, and the wind direction was out to sea.

    HealthyDrones.com - Innovative flight data analysis that matters

    I consider myself extremely lucky to have got this drone back to land from this flight, it stands as testament to just how good the P3P is that despite the odds being stacked so heavily against it, the drone managed to make it back to land.

    I hope this might help someone else avoid the same mistakes I made, it has certainly opened my eyes to the dangers of the wind.
     
  3. BudWalker

    Joined:
    Jun 24, 2015
    Messages:
    772
    Likes Received:
    225
    Location:
    Santa Margarita, Ca
    This is a really interesting incident. It would appear that when RTH was in effect the P3 was trying to dead reckon back to the home point. I.e., it didn't do any course correcting as is normally done.
    upload_2016-7-21_6-47-15.png
    At 1007 the waypoint mission initiated a RTH(displayed as the area with the light blue background). The P3 started to climb to it's RTH altitude (blue line). It turned to a heading of 84 degrees (red line), the direction of the home point. However, the directionOfTravel settled in to 55 degrees. The problem is that directionOfTravel remained at 55 degrees until the pilot took over and started providing stick input at 1108.
    upload_2016-7-21_6-52-50.png

    The question I have is why didn't the P3 determine that the directionOfTravel needed to change to 83 degrees and make the necessary motorSpeed changes.
    upload_2016-7-21_7-7-58.png
    The pilot was able to steer the P3 back home. It's not like there was too much wind and it couldn't achieve the requisite motorSpeeds. The P3 didn't even turn into the wind it just kept on the same 84 degree heading while being blown off to a directionOfTravel of 55 degrees.
     
    Meta4 likes this.
  4. GuernseyZoo

    Joined:
    May 21, 2016
    Messages:
    7
    Likes Received:
    3
    Location:
    Guernsey
    Thanks BudWalker for spending the time to look at my flight data and providing your analysis comments.

    I had just accepted the DJI official response that the problem in my RTH was caused solely by the wind / low battery as being the final reason, especially after seeing the wind speed data on the HealthyDrones site.

    I was so relieved to have taken over in time (just) to have got my drone back over the headland, and ultimately within only 20 feet of me.

    Today I carried out a full physical inspection of the P3P, checked the tightness of all the screws, did an IMU calibration, gimbal calibration, compass calibration, fitted a brand new set of props and did a dozen automatic take offs and landings plus some automatic RTH flights whilst being in the safety of the field behind my home. The P3P handled perfectly and I (hopefully) have got away without causing any damage. Admittedly the panic landing was quite soft and slow in the long grass, much better than if I had landed it on the surrounding granite rocks or hard pathways. It was only the one prop that first entered the long grass that stalled a single motor, and my CSC in under a second seems to have worked OK.

    Your comment about the P3P not trying to correct its heading during the RTH stage of the flight was inline with my initial gut feel about the incident. I had imagined that the drone would have used whatever momentum it had to try to head back home, even if that resulted in it basically standing still fighting a headwind but trying to fly in a straight line back home, or course adjusting during the RTH to always try to head back home. Instead it seemed content to fly more or less in a straight line out past the headland, at least until I took back over the control.

    I know I made many mistakes in this flight, including letting the power get so low before the "Go Home" kick in whilst still 2200 feet out to sea, and relying on the "Go Home" for too long before taking over was amongst them.

    I wasn't looking to blame the drone, but to try to understand more about why the drone maintained the constant trajectory during the "Go Home" stage without any course correction, compass and GPS were good both at take off and during the flight, out on the headland and across the bay must have been fairly free from wifi interference, it is only the section of flight where I sweep in along the coast road near the houses that I get some signal strength issues, presumably in part from so many of those homes having wifi.

    If "Go Home" can sometimes (rarely) happen without there being any course correction, it might contribute to some users fly-away experiences and not everyone would be so lucky as I was to get their drone back safely. I will certainly be more wary of the automatic RTH in future.