Shop OBEX P1 Docs P2 Docs Learn Events
ELEV-8 V3 - Flt 4 - crash report — Parallax Forums

ELEV-8 V3 - Flt 4 - crash report

macirishmacirish Posts: 53
in Robotics
4th flight - drone flipped and broke to propellers.

My analysis is that this was a ground effect crash - here's a picture with the details:
(Please comment)

flt4_crash_anal.jpg
flt4_crash_anal.jpg
900 x 900 - 57K

Comments

  • Ken GraceyKen Gracey Posts: 7,395
    Were you in an auto-level mode, or free-flight?

    Ken Gracey
  • macirishmacirish Posts: 53
    Free flight - still working on learning the art of flying these. You can bet I'll stay a little higher from now on.
  • xanaduxanadu Posts: 3,347
    Excellent image, do you work for the NTSB by chance?

    When I get close to the ground I fly in self-leveling mode, even in calm conditions, ground effect can be hard to tame!
  • macirishmacirish Posts: 53
    Nope - no NTSB here, just my anal nature. If I have to draw it and explain it - I'll remember it better in the future.

    This is my first exposure to the aviation world - I have to admit I'm impressed by the wealth of knowledge and 'rules of the road' that I never knew existed about flying. I'm trying to approach learning about it as professionally as I can - I'm sure this isn't news to the rest of this group, but while drones are sold as toys (not referring to Parallax, but to the general retail market) - they are anything but.
  • macirishmacirish Posts: 53
    BTW - anyone have a way to do local data logging on the ELEV-8? Could use it for analyzing this kind of thing. It would have been helpful to know the interplay of commands the FC saw when it flipped.
  • xanaduxanadu Posts: 3,347
    macirish wrote: »
    Nope - no NTSB here, just my anal nature. If I have to draw it and explain it - I'll remember it better in the future.

    This is my first exposure to the aviation world - I have to admit I'm impressed by the wealth of knowledge and 'rules of the road' that I never knew existed about flying. I'm trying to approach learning about it as professionally as I can - I'm sure this isn't news to the rest of this group, but while drones are sold as toys (not referring to Parallax, but to the general retail market) - they are anything but.

    First experience? You listed density altitude in your report. That's spectacular. Welcome to the hobby!

    I can't help with the data logging, I know there's telemetry and ground station software so it's definitely possible. Someone else is sure to chime in about that.

    Happy flying!
  • Kyle M.Kyle M. Posts: 112
    A few questions of clarification:
    1) What was you altitude above ground level?
    2) At what approximate speed (and/or pitch angle) was the ELEV-8 traveling at when the incident occurred?

    I do believe there is (or was) some code in the firmware for data logging, but I don't know how it works or what hardware to use it with. Hopefully Jason can enlighten us.
  • macirishmacirish Posts: 53
    Yes, first experience but FAA study exposed me to the concept of 'density altitude' - very interesting part of flying (or not flying).

    Altitude was front of drone was probably about 12" off the ground while the back was probably about 18". I'm pretty sure I was pushing the throttle up - not a lot but some.

    The slope is more gradual than the drawing - from peak to valley is 24" - and top to flat is about 7 feet.

    The pitch angle was about what the image is - the speed was slow, not really familiar with estimating - but maybe walking speed - 2-3 MPH.

    KenMc
  • JasonDorieJasonDorie Posts: 1,930
    There is some logging on board, but it's not terribly useful unfortunately. The code was written to log to an Atmel-based logging tool that doesn't have enough ram to buffer a full file block while writing one, and it ends up corrupting data quite regularly, which limits the usefulness. In addition, the FC code is pretty tight on timing, so full logging all the data during flight takes too much additional CPU time - you have to pick and choose what variables to log, which again, limits the usefulness somewhat.

    I'd really like to have another Prop on-board to do full logging of all internal state, but that's not a small undertaking.

    I have personally flipped mine exactly once when doing a hard forward run followed by a hard opposite-direction stop. The quad flipped front-to-back and plopped down in the grass upside down. There's a possibility that the control vector estimator incorrectly chose the inverted version of the direction. I'll take a look at the code and see if I can figure out if that's even possible.

    If you were in "manual" control mode it's hard to rule out user error or just overshoot, but if you were flying in one of the two auto-level modes it shouldn't be able to go upside down, ever, unless you were flying in a hurricane or something, or changed direction really hard and then killed the throttle. What you describe doesn't sound like either of those.
  • JasonDorieJasonDorie Posts: 1,930
    Oh - one last question - did you have a GoPro or anything mounted? I have personally had flips caused by WiFi interference from a GoPro camera with the WiFi enabled. Anything that broadcasts a 2.4ghz signal mounted on the quad can cause serious control interference. Just ruling that out.
  • macirishmacirish Posts: 53
    Thanks for the detail.

    No abrupt change in direction - I've only got about an hour with this drone - but I don't overcompensate.

    No, nothing additional mounted.

    Although - come to think of it I was flying right next to a pole that has a weather station that uses WIFI back to its display. An Accurite.

    The pole is aluminum - about 12 feet high and mounted on a steel pipe. Still, I was right next to that when the crash occurred.

  • JasonDorieJasonDorie Posts: 1,930
    That's not definitive, but it's certainly a possibility. Would depend on if it was broadcasting at the time. I had two different quads pitch over close to the ground with a broadcasting GoPro mounted before I realized what was causing it.
  • macirishmacirish Posts: 53
    I was mistaken - (senior moment...really) - the Acurite weather station is not WIFI - it uses 433 MHz.

    I surveyed the WIFI environment close to the spot where the drone flipped with NetSpot.

    There's 5 WAPs with a 2.4GHz signal - my WAP has a -80 dB signal - the rest are negligible. Noise is about -90 dB.
  • The_MasterThe_Master Posts: 200
    JasonDorie,

    Does the ELEV-8 firmware allow flipping more than 180 deg?

    Suppose I am in level flight. Then encounter freak windshear that flips me more than 90 deg. Say it flips me 185 deg. Will the computer know that it's upside down (plus 5 deg)?

  • JasonDorieJasonDorie Posts: 1,930
    Yes, the controller uses quaternions to track the orientation estimate, the user-desired orientation, and the difference between them.. It's done that way specifically so you can do things like fly upside down in "manual" mode, flick the switch to engage auto-level, and have it fully right itself.

    I've done a bunch of testing including throw-tests, flipping in actual flight, and flipping in tethered flight to make sure it all works as expected.



    I'd be surprised if a windshear could actually flip you - Unless you have the PID's turned way down, the quad fights you really hard when you deviate it even 10 degrees.
  • macirishmacirish Posts: 53
    Revised conclusion

    This was not a 'ground effect' flip - it was a battery running out flip.

    I flew today and tried making the flip happen - same spot same motion - not even a slight hick up.

    Looking back at the battery records - and considering that this was a long flight - I'm certain now that the batter was failing - which caused the flip.

    KenMc
Sign In or Register to comment.