DJI Spark Compass Calibration

[JoeCec]

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies

 

[AeroMirage]

Well done!
Very informative.

 

[LuvMyTJ]

Sticky'ed

 

[JoeCec]

Thank you!

 

[Mako79]

I thought it was nose down for the second phase of compass calibration (as per manual and previous DJI quads).
I really think the app was too lazy to get a proper (side) drawing of the spark.
I guess not many have complained about this or maybe we are so lucky that sideways compass dance works just as well as nose down.

 

[JoeCec]

I agree, I overlooked that while creating the video... I created a poll to ask if it should be nose down or if it even matters! LOL

 

[Dronason]

I thought it was nose down for the second phase of compass calibration (as per manual and previous DJI quads).
I really think the app was too lazy to get a proper (side) drawing of the spark.
I guess not many have complained about this or maybe we are so lucky that sideways compass dance works just as well as nose down.

I have done it as in the app and rechecked the calibration from a .DAT file record using CsvView. That's just fine. I think they just calibrate the Z component in the second step and the position don't mater as it is rotation over the Z axis.

Same rotations as during the calibration:


Plot of compass output using the MagDataPlayer of CsvView:

 

[Lancelot]

Wait wait

I agree, I overlooked that while creating the video...

Are you the introvert speaks? haha, i loved your detailed video man!

Tbh, the second rotation doesnt really matter whether it is camera down or camera sideways, both works.
On the app its sideways, on guide its down.

 

[JoeCec]

Wait wait

Are you the introvert speaks? haha, i loved your detailed video man!

Tbh, the second rotation doesnt really matter whether it is camera down or camera sideways, both works.
On the app its sideways, on guide its down.

Yes, I am The Introvert Speaks Thanks for your support!

 

[JoeCec]

I have done it as in the app and rechecked the calibration from a .DAT file record using CsvView. That's just fine. I think they just calibrate the Z component in the second step and the position don't mater as it is rotation over the Z axis.

Same rotations as during the calibration:
View attachment 951

Plot of compass output using the MagDataPlayer of CsvView:
View attachment 952

That's great stuff man! Thanks for sharing. It's interesting why it would stress the nose down orientation, but in the end not really matter. Maybe they use a default App template for all the drones and that was overlooked? Or the technology passed the need for that step and they neglected to remove it. Either way, I hope DJI's Quality Control starts to pick back up soon.

 

[Daniel Liew]

Thanks for the informative video! However, I find my compass calibration almost always fails. Maybe 8 out of 10 trials will fail. Nowadays I just depend on the auto detection and do not calibrate when there's no prompt.

 

[DesertWindAero]

Thanks for the informative video! However, I find my compass calibration almost always fails. Maybe 8 out of 10 trials will fail. Nowadays I just depend on the auto detection and do not calibrate when there's no prompt.

You are better off calibrating only when prompted to do so.

 

[Mako79]

That's great stuff man! Thanks for sharing. It's interesting why it would stress the nose down orientation, but in the end not really matter. Maybe they use a default App template for all the drones and that was overlooked? Or the technology passed the need for that step and they neglected to remove it. Either way, I hope DJI's Quality Control starts to pick back up soon.

Another good tip is to hold the quad above your head. Spin 360, drop nose down and spin 360 again. Why? The higher off the ground, the better.

 

[DesertWindAero]

The higher off the ground, the better.

I've never heard that before. Where did you read this...?

 

[Daniel Liew]

I've never heard that before. Where did you read this...?

Heard this before too, and I usually hold it at least 3-4 ft above ground. Still I'd like to know what I've been doing wrong all this while to always fail calibration

 

[Dronason]

I've never heard that before. Where did you read this...?

It is all matter where would come disturbance. Often the one you cannot see are under ground.
So increasing the distance will be good, but the best would be to go a more suitable place for calibration.

 

[Dronason]

You are better off calibrating only when prompted to do so.

Yes but don't calibrate at the place you get the message without understanding why you get the message. Often you move a little or take-off from hand and you are good.
The message says "Calibrate" but it should be understood as "Compass report wrong data, don't take off, without checking".

 

[Lancelot]

Because on the ground you can have steel bars, concrete filled with i dont know what. This will affect the magnetic field and thus the compass.
And do not calibrate it with you turning with it, it needs to be on its axis. Why, cause it might have been you that cause the interference (or devices on your body).

 

[JoeCec]

An important reason to calibrate while at a new location is for the Spark to get the interference levels of the new region. This will help the Spark understand what is "normal."

You also have to remember it only has one IMU and compass, no backup like the Mavic or Phantom. This is an important bit of information to keep in mind while flying.

Calibrate the Spark on ITS axis, not YOURS. Spin the Spark not your body. Calibrate in a smooth steady fashion. Then go have fun flying!

 

[BudWalker]

An important reason to calibrate while at a new location is for the Spark to get the interference levels of the new region. This will help the Spark understand what is "normal."
......

Not sure what is meant by "interference levels". But, it's a myth that a calibration determines and compensates for magnetic anomalies external to the AC.