New Compass Calibration Recommendations (from the v2.0 manual)

[Robert Mitchell]

It still may need calibration for local magnetic fields or random magnetisation it can get. For example, my Mavic's drift was caused by putting a Bluetooth speaker on top of the case where it was stored. Speakers = magnets.

Yeah, no problem doing it, I just didn't before because they advised not to unless prompted. I will now because they are saying to, and after all, they built the thing.

 

[Xtreme Drone Pilot]

I just didn't before because they advised not to unless prompted

I think they advised not to because most users don't pay attention to detail, such as removing all metallic objects from their persons, including (smart)watches and phones, doing it outside, away from metal structures, etc. And bad calibration can result in flyaway/crash.

 

[sar104]

Magnetic declination perhaps. Declination is totally a factor of location.

It does vary with location, but the compass calibration is unable to detect or measure it because the compass can only detect magnetic north and has no directional reference for true north. It could detect magnetic inclination, since that is relative to an absolute direction (vertical) that the IMU can detect via its accelerometers.

But neither of those is the purpose of calibration. During calibration you are rotating the aircraft, which has its own magnetic field, in a fixed external magnetic field ( the earth's). By mapping the magnetic field detected by the magnetometers as a function of orientation, the FC can determine which magnetic field components are static (the earth's) and which rotate with the magnetometers (the aircraft's). It then stores the magnetic field components due to the aircraft and subtracts them from all subsequent readings to leave just the earth's field.

And yes - declination is determined by a global model of declination as a function of location, stored in the firmware.

 

[Dan]

I picked up one of these http://cfixer.com, as I felt when in ATTI this bird should stand still, but no matter what I've done in ATTI with no wind it still wants to drift 1-2 miles per hour.

 

[sar104]

I picked up one of these Fix your compass error, as I felt when in ATTI this bird should stand still, but no matter what I've done in ATTI with no wind it still wants to drift 1-2 miles per hour.

Drifting in ATTI mode has absolutely nothing to do with the compass. In ATTI it is making no attempt at all to hold position, only attitude, so if it is consistently drifting in a particular direction, relative to the aircraft, then it most likely needs an IMU calibration, not a compass degaussing. Alternatively it could be due to a motor angle offset or a distorted prop - anything that leads to an imbalance in lateral thrust when torque is neutralized around the x, y and z axes.

 

[daisy-girl]

DJI recently released v2.0 of the Mavic manual. Check out the changes here.

The most notable change is the new compass calibration recommendations. DJI is now recommending the compass be calibrated when:

• Flying at a location further than 10 km away from the last flight location
• The Mavic has not been flown for 1+ months
• A compass error is displayed in DJI GO and the error remains after changing locations

Thanks for the info. How did you become aware there is an update to the manual?

 

[msinger]

How did you become aware there is an update to the manual?

DJI posted the new manual on the Mavic download page.

 

[Robert Mitchell]

My suggestion is don't fix what's not broken.

I tend to adhere to that philosophy, but being an ex Air Force tech, I pretty much go with the manufacturers suggestions, with regards to flying things.

That’s why I try to understand their reasoning.

 

[Robert Mitchell]

I think they advised not to because most users don't pay attention to detail, such as removing all metallic objects from their persons, including (smart)watches and phones, doing it outside, away from metal structures, etc. And bad calibration can result in flyaway/crash.

I’m wondering if the risk of bad calibrations, by those who don’t fully understand its importance, won’t outweigh the benefits. But I guess they’ve considered the consequences.

 

[msinger]

I’m wondering if the risk of bad calibrations, by those who don’t fully understand its importance, won’t outweigh the benefits.

It would be very difficult (if not impossible) to successfully complete a bad compass calibration.

 

[Dan]

Drifting in ATTI mode has absolutely nothing to do with the compass. In ATTI it is making no attempt at all to hold position, only attitude, so if it is consistently drifting in a particular direction, relative to the aircraft, then it most likely needs an IMU calibration, not a compass degaussing. Alternatively it could be due to a motor angle offset or a distorted prop - anything that leads to an imbalance in lateral thrust when torque is neutralized around the x, y and z axes.

Thanks I did notice that a propeller change did minimize a bit and I even went as far as doing the IMU Calibration on a table that was as perfectly level as possible within 0.1º.

 

[FASTFJR]

Anyone remember calibrating the P2 and the "J" Hook problem?

 

[sar104]

Anyone remember calibrating the P2 and the "J" Hook problem?

I spent a lot of time working with DJI and Ian Wood testing beta software to fix that problem.

 

[PepeLePew]

DJI recently released v2.0 of the Mavic manual. Check out the changes here.

The most notable change is the new compass calibration recommendations. DJI is now recommending the compass be calibrated when:

• Flying at a location further than 10 km away from the last flight location
• The Mavic has not been flown for 1+ months
• A compass error is displayed in DJI GO and the error remains after changing locations

Yeap noticed that this morning. Thanks for posting.
Back in my Chroma days everytime I wanted to fly in areas
I always calibrated the compass.
It only takes a few moments and as paranoid as I am of
watching a bird just fly away I feel DJI made a good call on this.
Most of the areas I hit with the Chroma either had major elevation
changes or electrical interference to some degree.
In most cases I was always 5 or more miles away from
the last home position.

 

[FASTFJR]

I spent a lot of time working with DJI and Ian Wood testing beta software to fix that problem.

That was a real big issue and I remember you guys working hard to solve the problem

 

[sar104]

That was a real big issue and I remember you guys working hard to solve the problem

Yes - it was a strange issue that we thought was due to incorrect declination values - some of the measurements I made by physically rotating the compass module suggested that the sign of the declination values was reversed, leading to significant problems in regions where the value was not close to zero. It was fixed, eventually, in a firmware update but the explanation was never confirmed as far as I'm aware.

 

[FASTFJR]

Yes - it was a strange issue that we thought was due to incorrect declination values - some of the measurements I made by physically rotating the compass module suggested that the sign of the declination values was reversed, leading to significant problems in regions where the value was not close to zero. It was fixed, eventually, in a firmware update but the explanation was never confirmed as far as I'm aware.

Your correct and no I never found out what final explanation was. Really glad it was finally fixed

 

[Top Hat]

Follow your best judgement guys.

 

[Xtreme Drone Pilot]

DJI posted the new manual on the Mavic download page.

Interesting that they bumped the version 1.8->2.0 just for this minor change. Skipping 1.9 altogether.

 

[Xtreme Drone Pilot]

It would be very difficult (if not impossible) to successfully complete a bad compass calibration.

I think during calibration it goes into "I'm accepting reality as is" mode, i.e. it's not comparing data to previous calibration to say "move away from ferromagnetic objects". So, if someone calibrates in a room next to powerful speaker, or in the car, it can be screwed up.