Discussions

Replies

                • Developer
                  leonardthall > Edgar Scott December 20, 2014 at 6:06am

                  Hi Edgar,

                  Let us know how you go.

                  Good luck.

                  • Edgar Scott > leonardthall December 21, 2014 at 5:05am

                    Leanoard,

                    I did a re-calibrate and it did seemed to help a little

                    I would like to to a compassmot but terminal is not working with 3.2/APM 2.6 (garbled data), i know you guys had to move the logs to the tabs, was compussmot moved too?  If so where is it now

                    [MAV 001:1] ArduCopter V3.2 (c8e0f3e1)

                    Thanks again...like always,

                    Ed

                  • Developer
                    leonardthall > leonardthall December 20, 2014 at 2:57pm

                    Hi Edgar,

                    I would suggest recalibrating your compass and doing a compasmot.

                  • Edgar Scott > leonardthall December 20, 2014 at 2:24pm

                    Leonard, I just raised the Pitch P Rate to match and increased the Stabilize P Rate a little and its better...but I have another issue, in GPS modes after staying in one spot for less than minute it starts to toilet bowl just trying to hover (POS_HOLD)

                    Any ideas on that?

          • Edgar Scott > leonardthall December 16, 2014 at 4:48pm

            I assume you mean of the drone, btw, thanks very much for looking at it, its really appreciated

            3702832277?profile=original

  • Brandon Clifford December 11, 2014 at 6:03pm

    So earlier my quad just crashed during auto tune, I had it in alt hold was doing its thing then all of the sudden goes full throttle, I lose orientation, lower throttle and then hit RTL (atleast i think i did was happening pretty fast) and it crashes into the side of a house. It broke some siding off, (stupid carbon props wish I stuck to plastic)

    Can anyone tell me what happened here?

    It crashed into snow at the end which would have shorted my ESC and my power module, so it may explain if the voltage goes crazy at the end.

    2014-12-11 20-31-20.log

    https://storage.ning.com/topology/rest/1.0/file/get/3701891801?profile=original
    • Developer
      leonardthall > Brandon Clifford December 12, 2014 at 11:02pm

      Hi Brandon,

      You had catastrophically bad GPS at the end of the flight (and to a lesser extent in the middle) which the EKF initially rejected, but after 10 seconds, the EKF gave up dead reckoning based just on the accelerometers and reset the position and velocity states back to the GPS (that had come back again).  Unfortunately this reset is also applied to the the vertical velocity, and the GPs vertical velocity at that time happened to be reading 60 m/s (ie crap).

      So this could turn out to be a bug/poor decision in the code. We are looking into it. I suspect we should be initialising it to a filtered baro rate or something.

      It is very hard to make these decisions reliably because each sensor falls over in different ways. GPS is great in the horizontal when it is working and ok in the vertical but when it fails, it fails spectacularly. The Accelerometers are pretty consistent all the time but suffer from noise and offset integration making them good for about 10 seconds without correction and baddly effected by attitude estimation errors. Then the baro can move up and down by a couple of meters from various aerodynamic effects and has a big delay.

      There is a good reason a high percentage of the dev's are bald!!!

      • JoeBob > leonardthall December 19, 2014 at 5:47am

         "GPS is great in the horizontal when it is working and ok in the vertical but when it fails, it fails spectacularly. The Accelerometers are pretty consistent all the time but suffer from noise and offset integration making them good for about 10 seconds without correction and baddly effected by attitude estimation errors. Then the baro can move up and down by a couple of meters from various aerodynamic effects and has a big delay."

        Thank you for the clear explanation!  And working on a solution.

        I've had no success using Sonar, moderate improvements using a Neo 8 GPS upgrade (Double the satellites, better triangulation from low position inputs), and hope to get a Lidar-Lite eventually.

        I want to master high speed terrain following at one meter (angle the sensor forward and down 45 degrees below horizon.)

        It is good to see people smarter than I am working on the same thing.

      • Rob Ratcliff > leonardthall December 13, 2014 at 8:24am

        With the issues with the barometer and GPS, especially close to the ground, it seems like lidar or radar would be a more reliable solution for altitude measurements. I noticed that 3DR has a lidar coming that looks promising: https://store.3drobotics.com/products/lidar-lite 

      • Brandon Clifford > leonardthall December 12, 2014 at 11:30pm

        Hm, I have been having bad luck with my GPS and that was actually a brand new unit in order to prevent any issues. I think it may have been because I had it to close to the telemetry radio this time. I shall try moving it to the back of the quad and try again.

        Thank you for getting to the bottom of this, that is something that I was definitely not capable of.

        Hopefully it was a one off thing.

        Was this the same issue that happened when someone flew indoors a month back? Hopefully tomorrow I can give auto tune a try again.

This reply was deleted.

Activity