INS13 indoor
« on: June 20, 2017, 12:24:56 am »
Hello,
revo mini FC, in complementary+mag - horizon is steady when sitting on the ground, INS13 indoor -  the horizon is balancing 2-3 degree.
Any idea? Normal behaviour? I'm kind afraid to try a flight. Thanks!

f5soh

  • *****
  • 3695
    • LibrePilot
Re: INS13 indoor
« Reply #1 on: June 20, 2017, 12:40:20 am »
INS13 Indoor is only used for testing purposes and should not be used for flight.

Re: INS13 indoor
« Reply #2 on: June 20, 2017, 03:30:00 am »
In INS13 (both Indoor and Outdoor=Navigation) it is normal for "Flight Data page" horizon to wobble a few degrees.  This is because compass sensor has a lot of noise, and INS13 uses full 3D of compass, including for leveling.

Mateusz

  • *
  • 720
Re: INS13 indoor
« Reply #3 on: June 20, 2017, 04:09:44 am »
Small deviation 1-2 deg is ok I guess, but before flight

  • Make sure you use external magnetometer
  • Make sure external magnetometer behaves the same way as internal one on scopes
  • Check attitude states x,y,z axis on scopes, if they do not oscillate. If you leave INS13 (which is EKF for navigation and should be used) for long time stationary, it may diverge and oscillate, but that's normal and won't happen in flight. After powering the board few seconds you should see nice horizontal lines in attitude states and no oscillation there.
  • Check system health, that all boxes are green all time.

Re: INS13 indoor
« Reply #4 on: June 20, 2017, 08:02:08 am »
Thanks all for answers, system health all green, I'll try re-calibrate everything and see where is going.

Re: INS13 indoor
« Reply #5 on: June 24, 2017, 09:03:24 am »
Hi, no luck for me yet... I attached 3 photos,
1 complementary+mag alg. watch what happen when I switch to INS13 (indoor), 2 the transition between and 3 the INS13 (indoor), look at the attitude scope (right up corner).
Any toughs about the oscillations? Thanks!

PS I forgot to say, in complementary+mag, flies perfect, no drift.... and it's a tricopter (Tcopter :-)
« Last Edit: June 24, 2017, 09:31:34 am by octavvian »

Mateusz

  • *
  • 720
Re: INS13 indoor
« Reply #6 on: June 24, 2017, 03:58:41 pm »
Bottom right scope is not informative.
It is not informative because you just show MagState.x which is internal magnetometer. You don't see "x" from AuxMag here. And you should move copter a bit too see changes in all x,y,z axis for both magnetometers.

Re: INS13 indoor
« Reply #7 on: June 24, 2017, 04:48:35 pm »
OK, so what should I do? The copter is on the ground in both situations. On INS13 attitude scope (up right) looks weird and copter isn't flying. What I'm missing? Thanks!

Re: INS13 indoor
« Reply #8 on: June 24, 2017, 08:05:47 pm »
That oscillation is an EKF divergence that only happens when the model is motionless for many minutes.  It shouldn't happen if you are moving it around or flying it.  It should go away (for many more minutes) if you rotate the model around on all axes.

Look at the flight data page and if the PFD/HUD looks reasonable (only a few degrees of jitter) and health is green it should be OK.

Re: INS13 indoor
« Reply #9 on: June 24, 2017, 08:32:37 pm »
Hi, unfortunately the oscillations appears few dozen seconds  after siting on ground, I try to fly it but, it yaws and not stabe at all.
All green but it look like it have more than few degree of jitter. I'll try making a short movie. Thanks for ideas. 

Re: INS13 indoor
« Reply #10 on: June 24, 2017, 08:40:15 pm »
What does FlightData -> PFD look like, after GPS is green and you rotated it around and then set it on ground?

Another cause of this oscillation is aux mag misalignment by a large amount (say > 20 degrees).

Your first tests with INS13 should be with Attitude mode (Stab1).

Re: INS13 indoor
« Reply #11 on: June 24, 2017, 08:52:02 pm »
Hi, physically the aux mag is optically aligned, so it could be a misalignment by few degrees, I'll try that soon. I always start with stabilized...
Thanks.
« Last Edit: June 24, 2017, 10:04:33 pm by octavvian »

Re: INS13 indoor
« Reply #12 on: June 24, 2017, 10:02:03 pm »
What I'm really frustrated is the fact that the copter fly's very well in complementary+mag alg. and switching to INS13 alg. everything goes haywire ... I do known that is a complex algorithm but never the less this should add security and reliability to it.
I'm sure it's my fault, but don't seems to find the error. To summarize when switching from complementary to INS13 it goes from flat virtual horizon to dancing Mary  ;D.

Re: INS13 indoor
« Reply #13 on: June 25, 2017, 06:39:00 am »
Dancing Mary...  Would you say that the PFD dances just a few degrees (less than 10) or a lot like flipping upside down etc.

Flipping upside down is caused by aug mag rotation problem.  If you are running an I2C aux mag, you probably need the attitude->magnetometer->auxmagrotation (whatever it is called...) to be 0,180,0 or 180,0,180 (they are the same).

Re: INS13 indoor
« Reply #14 on: June 25, 2017, 09:47:17 am »
Hi, is not flipping... The aux mag is a stand alone board so the mag IC position is exactly like on board one.
Thanks.