Got this GPS message today

[msinger]

If I did re-calibrate, sitting on the table, I'm sure the calibration would be incorrect since it is in the presence of an external magnetic field, and I would be in for some unpleasant surprises if I flew at that point.

Actually, the calibration attempt will fail unless you somehow manage to attach the table to your Mavic so it spins around with your Mavic as you calibrate the compass.

 

[geigy]

Actually, the calibration attempt will fail unless you somehow manage to attach the table to your Mavic so it spins around with your Mavic as you calibrate the compass.

Just standing near the table and doing the calibration may create a bad calibration, with no error message from the drone.

Now, it's possible, as you say, that the DJI software might detect that the magnetic field varied in a non uniform way, but I think it will have no way of knowing. Anyone want to experiment with calibrating the compass next to a magnetic source?

 

[msinger]

Just standing near the table and doing the calibration may create a bad calibration, with no error message from the drone.

Standing near the table and taking off could negatively affect the compass, but you should not be able to successfully calibrate the compass if you're close enough to the table for it to negatively affect the compass.

Anyone want to experiment with calibrating the compass next to a magnetic source?

I've tried this a few times before and found the calibration process always fails. After you finish the second rotation, the DJI GO app will display a message stating the calibration failed due to strong magnetic interference.

 

[Mike4884]

My Mavic, asks me to calibrate the compass at startup about 90% of the time. No big deal, to me it is just part of the preflight process.

Then your mavic is messed up dude! Mine never asks..... part of your PreFlight check is always to calibrate your compass? are you sure that it's even asking you? Does it say it in Red Letters? " calibrate compass" if not I think you're doing something wrong or misinterpreting something

 

[sar104]

Just standing near the table and doing the calibration may create a bad calibration, with no error message from the drone.

Now, it's possible, as you say, that the DJI software might detect that the magnetic field varied in a non uniform way, but I think it will have no way of knowing. Anyone want to experiment with calibrating the compass next to a magnetic source?

Based on the evidence available and the current understanding of the DJI implementation, I don't think that's true or is, at least, very unlikely. During calibration the FC is simply trying to distinguish (and characterize) the apparently unchanging magnetic field (due to the aircraft components) from the external moving field (the earth's, plus any other local field). Unless the external field is too strong for that to work (in which case the calibration will fail), it really doesn't matter what the external field comprises, since it is subtracted out in the measurement of the aircraft's field.

The only real concern that I can see is if the local field also changes during the calibration process, which could happen if the local field gradient is large relative to the size of the aircraft, for example. However, the chances of that still producing an accepted calibration seems very small since the fluctuations would still be dominated by rotation, and a robust calibration algorithm should detect that.

I cannot recall any recent Phantom or Mavic logs that have been analyzed that have indicated a bad calibration. It's always magnetic distortion at IMU initialization or later in the flight, or magnetic components added to the aircraft without recalibration, which supports @BudWalker's assessment.

That was not true back in the days of the P2, but the calibration process seems to have become much more robust. That said, I have not calibrated either my P4 or Mavic since purchase, because they fly just fine and have not requested recalibration.

 

[geigy]

Based on the evidence available and the current understanding of the DJI implementation, I don't think that's true or is, at least, very unlikely. During calibration the FC is simply trying to distinguish (and characterize) the apparently unchanging magnetic field (due to the aircraft components) from the external moving field (the earth's, plus any other local field). Unless the external field is too strong for that to work (in which case the calibration will fail), it really doesn't matter what the external field comprises, since it is subtracted out in the measurement of the aircraft's field.

The only real concern that I can see is if the local field also changes during the calibration process, which could happen if the local field gradient is large relative to the size of the aircraft, for example. However, the chances of that still producing an accepted calibration seems very small since the fluctuations would still be dominated by rotation, and a robust calibration algorithm should detect that.

I cannot recall any recent Phantom or Mavic logs that have been analyzed that have indicated a bad calibration. It's always magnetic distortion at IMU initialization or later in the flight, or magnetic components added to the aircraft without recalibration, which supports @BudWalker's assessment.

That was not true back in the days of the P2, but the calibration process seems to have become much more robust. That said, I have not calibrated either my P4 or Mavic since purchase, because they fly just fine and have not requested recalibration.

Would you recommend, then, that a beginner calibrate their compass, just to make sure things are OK?

 

[sar104]

Would you recommend, then, that a beginner calibrate their compass, just to make sure things are OK?

I would recommend calibrating if there is any indication of a compass problem that is not explained by a poor location and, after adding any new components, if it has been exposed to strong magnetic fields (strong enough to change the magnetization of components, including screws) and perhaps on a new aircraft. Provided that it is done carefully, and after reading the manual, it is unlikely to cause any problems. Mistakes are a poor choice of calibration location and calibrating while wearing magnetic bracelets, etc.

I would not recommend calibrating every flight or simply because you have changed location (despite what the recent DJI article said) because that has nothing to do with what calibration is trying to achieve. It's a waste of time, it represents messing unnecessarily with a critical part of the FC, and "why try to fix something that isn't broken" applies here.

 

[Helo]

It looks pretty cold there. I wouldn't rule out temp as a factor. Just think how cold the Mavic is flying at 30 mph.

Wind chill does not effect to a mavic. I does effect humans because it evaporates moisture on exposed skin witch causes cooling so unless your mavic is sweating your OK.

 

[sar104]

Wind chill does not effect to a mavic. I does effect humans because it evaporates moisture on exposed skin witch causes cooling so unless your mavic is sweating your OK.

Yes and no. Yes if the Mavic is in thermal equilibrium with the air. No if it is warmer which, in general, it will be, since higher wind speeds will increase its cooling rate towards thermal equilibrium.

 

[imsportsdad]

I would recommend calibrating if there is any indication of a compass problem that is not explained by a poor location and, after adding any new components, if it has been exposed to strong magnetic fields (strong enough to change the magnetization of components, including screws) and perhaps on a new aircraft. Provided that it is done carefully, and after reading the manual, it is unlikely to cause any problems. Mistakes are a poor choice of calibration location and calibrating while wearing magnetic bracelets, etc.

I would not recommend calibrating every flight or simply because you have changed location (despite what the recent DJI article said) because that has nothing to do with what calibration is trying to achieve. It's a waste of time, it represents messing unnecessarily with a critical part of the FC, and "why try to fix something that isn't broken" applies here.

Question: would going thru airport security scanning equipment (strong magnetic fields) perhaps be a reason to re-calibrate the compass?

 

[ac0j]

A compass looks for one thing. Magnetic north.
If you get a compass error, how does this tell you that you must be in a magnetic field? The compass points to the magnetism, so the craft has nothing to believe but the compass. It tells me the compass isnt communicating with the mavic. Or there are two conflicting compass's on board the Mavic.
Near magnetic interference, IE. metal structure, car, whatever, The compass may point to the object thinking that it is north. Once in the air, the normal compass heading would redirect the compass to proper north seamlessly, and I would expect a smooth adjustment by the UAV.
BUT, how in the world would the Mavic know that it is pointing at a false north or magnetic north? It gets that info from the compass. The only way it could display a compass error is if it isnt working I would think.

 

[Spy808]

I would return it n exchange... Or if u can't just replace the GPS\Compass... Do u reelly want to take a chance n lose ur birdy..? My electronic board designer friend said specifically n pointed it out to me what fails n whÿ... Up to u if u wanna pull the trigger n take a chance of losing it... It shouldn't be a battle to calibrate if it is then face it something is wrong...

 

[sar104]

Question: would going thru airport security scanning equipment (strong magnetic fields) perhaps be a reason to re-calibrate the compass?

I have not had any problems with airport scanners, which are mostly x-ray units, but who knows what kind of magnetic fields might be present. I can't rule it out.

 

[meens]

A question about calibration, in the manual it is indicated you have to rotate the aircraft 360° but does not precise whether this is clockwise or anticlockwise. However the drawings shows it anticlockwise. Is the direction of rotation important. I recently did a video on the compass calibration and did it clockwise, DJI GO4 told me the calibration was correct and indeed when I tested it afterward the compass worked flawlessly.

 

[msinger]

A question about calibration, in the manual it is indicated you have to rotate the aircraft 360° but does not precise whether this is clockwise or anticlockwise. However the drawings shows it anticlockwise. Is the direction of rotation important.

No. You can spin the Mavic in either direction.

 

[lweekspors]

My MP never asks me to calibrate either.

 

[BudWalker]

A compass looks for one thing. Magnetic north.
If you get a compass error, how does this tell you that you must be in a magnetic field? The compass points to the magnetism, so the craft has nothing to believe but the compass. It tells me the compass isnt communicating with the mavic. Or there are two conflicting compass's on board the Mavic.
Near magnetic interference, IE. metal structure, car, whatever, The compass may point to the object thinking that it is north. Once in the air, the normal compass heading would redirect the compass to proper north seamlessly, and I would expect a smooth adjustment by the UAV.
BUT, how in the world would the Mavic know that it is pointing at a false north or magnetic north? It gets that info from the compass. The only way it could display a compass error is if it isnt working I would think.

You're right. The compass just points to geomagnetic north. But, there is a little more to this. The Flight Controller's Yaw value doesn't come just from the compass. Yaw is a computed value derived mostly from the gyros and accelerometers with a small input from the compass. If the FC sees a discrepancy between the computed Yaw and the compass it decides that the computed Yaw is suspect and will indicate a "compass error". It can't know the cause of the discrepancy, just that there is one.

The Yaw value is being updated continuously. At every iteration it mostly uses the current Yaw value, gyro data and accelerometer data to compute the next Yaw value. The magnetometer data is also used but at a lower gain. This means that when the compass was incorrect but then becomes correct the resulting Yaw correction is delayed. Depending on the platform, size of the error, etc it can take anywhere from 5 secs to more than 30 secs for the Yaw to become correct.

 

[djaco]

I have flown many quad copters prior to the Mavic, all homebuilt, using DJI controllers. Many pilots suggested calibrating the compass at each new flying site, due to local magnetic differences. Not sure if it is overkill, but it will not hurt the Mavic to do the calibration.

 

[ScubaBob]

You have misinterpreted the error message on the screen. You compass has completely failed, not GPS.

The compass is a different piece of hardware. Without a functioning compass, the aircraft can not tell what direction it's pointing. This can be inferred while moving, but it is far too imprecise for flight control. The sensitivity (both angular and sample rate) of the digital compass is necessary for flight control.

As a consequence, the aircraft switched to ATTI mode because all the flight controller can do is maintain attitude via the ICU. Even though it knows location via GPS, it can not hold position because it doesn't know it's heading.

Since the aircraft was close to you (and the ground, and surrounding ground environment) it probably was affected by some external magnetic field. A good candidate would be an electrical transformer, telephone patch box, etc. However, at this point it's near impossible to find out what unless it can be repeated.

I agree. The Compass Error message is very obvious on the display and I think that is where your problem originated. At the least you have to remedy that problem. Were you standing near a lot of metal or some place with a metal roof?

 

[ac0j]

I sometimes wonder how my pockets arent stripped of their metal contents when I walk on concrete or next to metal things. These things dont usually affect my old school pocket compass, or the fancy digital one in the mirror of my car. I cant help but think there isnt a little crying wolf going on with this mavic compass stuff.