Discussions

APM 3.2.1 Very bad Loiter, how to "debug" ? Any hints ?

Posted by Nils Heidorn on May 25, 2015 at 11:40am

Hi !

I have a small (370 mm) Hexa running 3.2.1 on a Crius APM clone with a Ublox 8 GPS / Compass combo.

AFter some stupid mistakes it works ***perfect*** in Alt-Hold and the Heading is spot on as far as i can measure ( i point the Copter towars a known landmark and check in Mission Planner if the Heading-Line gors through that landmark).

However i ***did*** have trouble with calibration but i believe the current MP GZI is buggy, see thread: Compass Calibration trouble Thread

I have no idea if that is relea6ed, but i thought i mention it :-)

Everytime before when i had bad Loiter (toilet bowl) the compass was not working or was a huge angle off.

But this time the compass seems okay from the heading shown in Mission Planner, I did run compassmot but i had very little interference from the start and of course even better after the compassmot ( <2% ).

So i am a bit helpless, is it a bug in 3.2.1 ???

Mission planner cant correctly calibrate on a "known good" Hexa running 3.1.5. (also no dots, no graph rotation, error 99).

Do you have any hints how i can start to debug this or how to find a trace ?

I guess some log channel can shed more light on this but i dont know which ?!

(Something like "loiter desired position = XXX, current position = XXX, moving at Heading XXX at Speed XXX)...

Please give me a hint, or maybe i should downgrade to 3.1.5 and find an old MP ?!

Do i need an EEPROM clear after down/upgrading ?

Greetings,

Nils

You need to be a member of diydrones to add comments!

Join diydrones

Email me when people reply –
Follow

Replies

  • GadgetGeek August 1, 2015 at 10:35am
    Did you ever get any help with this issue? If you did please post your results. It's hard to believe you didn't get any response here. I haven't had this exact issue before but I had experienced the "toilet bowl effect" on my first build. I got away from that once I realized the manometer (compass) was getting DC interference. As you stated you've ran the compassmot test and at 2%, no problem should exist with DC interference. If you haven't got that fixed and need some help I'd be more than happy to help you trouble shoot and run through a few different options.
  • Nils Heidorn June 1, 2015 at 1:09pm

    Hi !

    Please see a log attached, showing how i keep swithcing to loiter and "saving" the Hexa by switching back to alt hold as it wanted to fly away.

    Any help ?

    Greetings,

    Nils

    2015-05-29 21-17-41.7z

    • Klayt Morfoot > Nils Heidorn May 20, 2016 at 7:09pm

      You have several instances of GPS failsafe 1 (loss of lock for 5 seconds) and one GPS failsafe 2 (glitch). Where were you flying? It lost all satellites for a bit.

  • Nils Heidorn May 28, 2015 at 1:08am

    HELP :-)

    (aka "bump")

    Is there no kind soul that can put me on the right tracks to this problem ?

    I cannot find a reason why Loiter is so all-over-the-place.

    I tried to gather facts in the postings above, if there is something missing to help me, please let me know.

    I do not know which stuff to log to make a log file sensible.

    I did try something new: I flew with "simple mode" to check if compass is also good in-flight.

    simple mode works great, i can rotate & move in a direction ("UFO") just fine.

    So compass seems to be okay like indicated on the ground and during the compassmot.

    Also the Map Plots are okay, so i always end up with:

    GPS okay, compass okay, why does it not Loiter okay ????

    Whenever i activate Loiter from ALT Hold it starts to flyaway. changing direction, sometimes slow, sometimes flank speed.

    I don't get it !

    Greetings,

    Nils

  • Nils Heidorn May 25, 2015 at 2:05pm

    Hi again,

    i erased the EEPROM with arduino IDE, flashed 3.1.5 and redid calibrations.

    Same Problem, all looks fin but as soon as i activate Loiter the Hexa tries to "run away".

    Again i looked at the logs: started out with 8 sats, went to 10 sats fast. All positions look fine (within my 5x5m test ground, was too lazy to go to wide open space).

    ANy ideas how i can get a hold of this ?

    Greetings,

    Nils

  • Nils Heidorn May 25, 2015 at 12:48pm

    Hi !

    Meanwhile i checked the log again and can be 100% sure that the GPS Position itself is perfectly precise when i activated Loiter (I am using the APM Analyzer Software). So in my last test flight where i engaged Loiter 9 times (with harsh "wanna-fly-away" and switching back to Alt-hold quickly) i could veryfy that the actual GPS position were fine (startiung at 7 sats at the beginning up to 11 at the end).

    So the GPS is not the problem, and the mag "looks" okay from Mission planner as well.

    The compass is external and has orientation of 180 Roll. Could it be that the current Firmware version does report correct heading to mission planner but does not take in account the "Roll 180" for Loiter ? I cannot find "compass heading" in the log file anywhere.

    See attached a screenshot of all (???) compass related data.

    Greetings,

    Nils

    compass.JPG

This reply was deleted.

Activity