Greetings All,
Last weekend I maiden-ed my new 8' (2.4m) Super Sky Surfer with Arduplane 2.72 and had a fantastic first flight. Stabilize, FBW-A and Loiter all worked fine right out of the box with no further tuning; although I'm sure some refinement will be necessary for optimal auto flight modes. The tlog labeled 'Maiden SSS 2013-05-04 17-43' shows the Maiden flight for the SSS, the flight actually starts at around 46% of the recording.
Two days later I decided to try to get some more flights in; however I was foiled by higher than anticipated winds. Undaunted I decided to work on some other things. So I set up my antennae tracker and my FPV equipment to test them out and get the bugs worked out so to speak. Almost immediately I realized there was a problem. I had a friend point the nose of the plane South yet the heads up display from the MinimOSD and the compass rose in the Mission Planner showed that I was pointing North. I had my friend march around the field plane in hand pointing in the direction of travel. If you play back the tlog labeled 'Walking 180 2013-05-06 17-42-05' and move forward to 22% of the recording, you will see where my friend starts walking around the field. You will note that the plane is depicted as traveling backwards instead of forwards despite the fact that the nose was pointed in the direction of travel the whole time; indication that the compass reading is off by 180 degrees. The attached file labeled '2013-05-04 17-43-56.param' is a copy of the parameter file used in both instances.
Here are all the Compass Parameters and there current settings:
Line 65: COMPASS_OFS_X 43.09858
Line 66: COMPASS_OFS_Y -38.92876
Line 67: COMPASS_OFS_Z 41.06197
Line 95: COMPASS_MOT_Z 0
Line 169: COMPASS_DEC -0.247
Line 185: COMPASS_MOT_Y 0
Line 224: COMPASS_AUTODEC 0
Line 236: COMPASS_MOTCT 0
Line 238: COMPASS_USE 1
Line 239: COMPASS_LEARN 1
Line 259: COMPASS_MOT_X 0
I'm not sure what the COMPASS_MOT_* parameters are for and the COMPASS_OFS_XYZ values from what I understand don't seem that unusual. One thing to note is that the motor was never turned on during the walking 'flight' so the magnetic field would have been different during that phase.
For the record my setup includes the following:
APM 2.5
Attopilot 90A voltage and current sensor
uBlox LEA-6H GPS
Airspeed Sensor
MinimOSD v1.1
900MHz 3DR Telemetry
Castle Creations Thunderbird 54 AMP ESC (signal and ground only connected to APM)
Castle Creations 10 AMP BEC (powers the board and servos via the output servo rail)
Immersion RC 5.8GHz 600mw VTx (powered by a 3S lipo that also powers the analog side of the MinimOSD)
Futaba T8FG Super with 6208SB 2.4GHz Receiver
Now for the questions:
1) Is there any real advantage to using the compass in a plane? Won't the fact that we are in constant motion enable the GPS to provide the heading? (As opposed to a copter that can stand still)
2) If one does use the compass, should they always leave COMPASS_LEARN on, or should it be turned off after a successful flight thus fixing the offsets at their current value?
3) Given the current situation if I have another flight should this correct itself while in flight with the motor running, or should I manually return the COMPASS_OFS_XYZ values back to zero and start over?
4) If anyone can give me a good analysis of the flight logs with the relevant parameters I would love to see it here?
5) Is there a good way to manually calibrate the compass?
Thanks to everyone in advance I look forward to hearing your input!
Happy Landings,
Nathaniel ~KD2DEY
Replies
It has a built in detection system for large compass errors. If it detects that the compass is way off it will switch to gps nav alone.
I suggest taking your plane for a spin again, just to see how it behaves. For manual stabilize and fbwa modes you should have no issues since compass is not used.
I suggest you try rtl as well but be prepared to switch back to manual. Just see how it navigates.
We'll analyse that log to see if everything is functioning as it should.
(I still suspect your friend was carrying something magnetic close by the airplane. Those magnetometers are absurdly sensitive)
I have graphed below three values for mag_field, heading, and servo3raw. In both graphs what I am really looking at are the mag_field plots as you pointed out Pieter. The other two plots, heading and servo3raw are simply for reference to show the beginning and end of activity.
The first graph is from the maiden flight and you can clearly see the magfield with a peak around 750 then stabilize back to 0 and the flight begins around 950 along the X axis as indicated by the servo3raw trace.
In the next graph I have the same three values plotted for the walking 'flight' but the results are quite different. The magfield is shown in red again, the heading in green and the servo3raw in blue (throttle was never engaged) the scale on the right is the servo3raw and the compass. This time the magfield starts at about 750 and peaks around 2800 stabilizing at around 1950. This obviously explains the compass error, but the question is why?
Regards,
Nathaniel ~KD2DEY
it looks like your compass declination is off.
I am no expert on this but when I look at the map it should be around -14 take a look: http://magnetic-declination.com/
Maybe if you turn compass autodec on you will have better readings.
(but this should not cause your compass to point 180 degrees in the wrong direction)
you also could try to do a manual calibration in mission planner (hardware, compass) when you do that just make sure you're outside and avoid magnetic interference.
then move the plane around all axis to learn the compass the correct offsets.
I leave compass_learn always on. I notice an offset when I apply full throttle, I think the power wires and motor are emitting large magnetic interference but they are at least 3 inches away from my APM. I guess compass is very sensitive.