Compass Calibration, A Complete Primer

[GadgetGuy]

There's no fixed distance based deciding factor. Sooo many compass based misunderstandings still circulated here.

Care to elaborate? The point was not 150 miles but certainly at anything less than that, there is certainly no need to recalibrate, just because you are in a different location than your last good calibration. Frankly, I have found that straight out of the box P4P's require no compass calibration!

 

[N017RW]

Mentioning specifics creates misconceptions. I could ask you how you chose 150.

 

[GadgetGuy]

Mentioning specifics creates misconceptions. I could ask you how you chose 150.

Not mentioning specifics leads to too frequent compass calibrations onsite which are far more likely to replace a good compass reading with a bad one. Do you disagree that anything within 150 miles of your last known calibration does not require a new calibration, unless compass errors are reported and moving the aircraft does not remove the errors? Most drone flyers are flying within a 150 mile radius, unless traveling. No need for recalibrations before every flight, and doing so causes more harm than good!

 

[N017RW]

Yes. My experience is 600 miles so far.

 

[ianwood]

If you're referring to geoMagnetic declination

No. Not referring to declination. Please read some of the back posts in this thread as that is covered numerous times.

Calibration does not measure declination. Calibration measures deviation which are the distortions both from the craft as well as in the location. The key difference between the two being that one stays static with the craft's orientation, the other is static to the earth and is encountered differently as the craft changes orientation.

The causes of the latter are based on geomagnetic anomalies that occur throughout the earth's magnetic field. This can be caused by ore deposits on the earth's crust, tectonic plate composition, etc.

These may or may not vary from location to location enough to influence the magnetometer's function. But it is hard to know one way or the other. Which is why, as a rule of thumb, it is a good idea to re-calibrate if you cover a reasonable distance.

And as we can see, there is a fair bit of guesswork with regard to what is a reasonable distance.

 

[GadgetGuy]

Yes. My experience is 600 miles so far.

Cool! We are in agreement about at least 150 miles. 600 miles works for me, too!
The main takeaway is that whether the number is 600 miles or 150 miles, unless you are travelling a great distance, a new calibration is unnecessary, and will likely cause more problems than it will solve. One good calibration is all that most will ever need.

 

[BudWalker]

No. Not referring to declination. Please read some of the back posts in this thread as that is covered numerous times.

Calibration does not measure declination. Calibration measures deviation which are the distortions both from the craft as well as in the location. The key difference between the two being that one stays static with the craft's orientation, the other is static to the earth and is encountered differently as the craft changes orientation.

The causes of the latter are based on geomagnetic anomalies that occur throughout the earth's magnetic field. This can be caused by ore deposits on the earth's crust, tectonic plate composition, etc.

These may or may not vary from location to location enough to influence the magnetometer's function. But it is hard to know one way or the other. Which is why, as a rule of thumb, it is a good idea to re-calibrate if you cover a reasonable distance.

And as we can see, there is a fair bit of guesswork with regard to what is a reasonable distance.

"Ore deposits on the earth's crust, tectonic plate composition, etc." cause distortions that have the same characteristics as those causing declination. In fact these anomalies could be considered to be a component part of declination. The algorithm that computes declination in DatCon is based on the World Magnetic Model which comes from actual measurements of the geomagnetic field. I suspect the DJI drones all use the WMM to compute declination. Take a look at
World Magnetic Model | NCEI
for more info on this.

Like declination it's effectively impossible for a compass calibration to discover and/or compensate for these anomalies. In particular, changing the orientation of the compass will not yield variations in field strength. The article I referenced previously has a good explanation of this.
Compensating for Tilt, Hard-Iron, and Soft-Iron Effects | Sensors Magazine

If you really believe that these anomalies can be discovered by a compass calibration and then compensated for then I encourage you to publish your idea. You'll be famous.

 

[ianwood]

UGGH! I'll say it again and hopefully for the last time:

Declination is NOT part of calibration. Deviation IS. Declination and deviation are two different things. Deviation is caused by near field anomalies whereas declination is not. Deviation consists of anomalies that maintain orientation with the craft AND those that don't. Declination offset is consistent to any heading whereas deviation varies based on heading and is often noted at cardinal points on aircraft.

And not even considered in this is a third magnetic anomaly called inclination. Let's not even go there!

 

[BudWalker]

UGGH! I'll say it again and hopefully for the last time:

Declination is NOT part of calibration. Deviation IS. Declination and deviation are two different things. Deviation is caused by near field anomalies whereas declination is not. Deviation consists of anomalies that maintain orientation with the craft AND those that don't. Declination offset is consistent to any heading whereas deviation varies based on heading and is often noted at cardinal points on aircraft.

And not even considered in this is a third magnetic anomaly called inclination. Let's not even go there!

Maybe I misunderstood. But, it sure looks to me like you were saying that "ore deposits on the earth's crust, tectonic plate composition, etc." can be detected by a calibration and then compensated for. These are declination type features and cannot be detected by a calibration.

 

[KongWerks]

Flew in Vancouver BC Canada last week, Long Island New York this week. No request for compass recalibration. Didn't have any issues. I see some of the software versions for other Phantoms have removed the compass calibration from the main screen. It's only a compass. You aren't taking a bearing off of it so as long as it points the nose in the right direction no calibration needed.

 

[ianwood]

It's only a compass. You aren't taking a bearing off of it so as long as it points the nose in the right direction no calibration needed.

it works right up until it doesn't. By the time the drone can detect something is wrong, the compass is so far off, you would surely crash. The compass is the sensor most implicated in uncontrolled flight resulting in a crash. If the compass and GPS disagree by more than a few degrees, you'll have a lot of fun maintaining control.

Maybe I misunderstood. But, it sure looks to me like you were saying that "ore deposits on the earth's crust, tectonic plate composition, etc." can be detected by a calibration and then compensated for. These are declination type features and cannot be detected by a calibration.

Declination = molten ore flows in the center of the earth that influence global polarity.
Deviation = tectonic or other surface-based magnetic anomalies. Localized.

 

[BudWalker]

.......

Deviation = tectonic or other surface-based magnetic anomalies. Localized.

You can label it how ever you want, but that's not the issue. If the anomaly such as "tectonic or other surface-based magnetic anomalies" or "ore deposits on the earth's crust, tectonic plate composition, etc." is external to AC then it can't be discovered by the calibration. If the anomaly causes variations in field strength dependent on orientation then the anomaly is effectively ambiguous; change the calibration location by a small amount and the results of the calibration are different. OTOH, if the field strength doesn't vary then even the existence of the anomaly can't be known, let alone it's characteristics.

 

[KongWerks]

it works right up until it doesn't. By the time the drone can detect something is wrong, the compass is so far off, you would surely crash. The compass is the sensor most implicated in uncontrolled flight resulting in a crash. If the compass and GPS disagree by more than a few degrees, you'll have a lot of fun maintaining control.



Declination = molten ore flows in the center of the earth that influence global polarity.
Deviation = tectonic or other surface-based magnetic anomalies. Localized.

AFAIK... The compass and GPS have no real interaction with each other. I guess we will have to agree to disagree on this one.

 

[N017RW]

Interaction?
Maybe that term is obscuring their interdependency... a better term IMO.

The compass indicates bearing- the direction which indicates forward to a MR. This is important cause it can move horizontally in ANY direction [as opposed to a fixed wing aircraft].

So when the direction required to transition between two (GPS) coordinates/points are needed the compass is used to determine the heading needed to get there. This allows the MC to coordinate the motors to move appropriately. Even if your autonomous program instructs an other than forward motion the compass must still be used to maintain the appropriate heading.

So RTH, waypoint, or any other autonomous flight is not possible without both sensors and accurate compass calibration (compensation).

 

[ruggb]

The sensor in the AC is a 2 axis sensor similar to a HMC6352 device. We know this because the cal procedure is only specified in 2 axis. The following is extracted from that data sheet.

User Calibration
The HMC6352 provides a user calibration routine with the “C” command permitting entry into the calibration mode and the “E” command to exit the calibration mode. Once in calibration mode, the user is requested to rotate the compass on a flat surface at least one full circular rotation while the HMC6352 collects several readings per second at various headings with the emphasis on rotation smoothness to gather uniformly spaced readings. Optimally two rotations over 20 seconds duration would provide an accurate calibration. The calibration time window is recommended to be from 6 seconds up to 3 minutes depending on the end user’s platform. The calibration routine collects these readings to correct for hard-iron distortions of the earth’s magnetic field. These hardiron effects are due to magnetized materials nearby the HMC6352 part that in a fixed position with respect to the end user platform. An example would be the magnetized chassis or engine block of a vehicle in which the compass is mounted onto. Upon exiting the calibration mode, the resulting magnetometer offsets and scaling factors are updated.

If anything, to improve the result turn slowly. The AC firmware determines if it has sufficient data and is programmed for one turn. You can play with ANY device - phone, tablet, etc. - and see the effects of turning too fast or too slow. It will argue about it. It is the same device/process.

This whole thread is based on the fact that if you get a good cal - away from any "hard-iron distortions" - doing another cal in the vicinity of these distortions will make things worse, not better. The distortions are local and have nothing to do with "declination". Declination is determined from GPS co-ordinates referenced to a table. They change VERY slowly and would be compensated for with updated firmware - except your AC and you probably will not live long enough to see the change.

This is from thecompasstore.com explanation and chart
Below is an actual declination chart from 1990, showing the worldwide declination situation in 1990. The current declination values for today will be very close, if not identical,

When the AC is in the air, it is normally far enough away from any local distortions so that if you calibrated it near a distortion, it is now uncalibrated in flight. - Hence, you made it worse. Fly around power lines or metal structures and all bets are off.

So do it right once and leave it alone. The only reason for it to change is that components of the AC became more or less magnetized. Those are local distortions that you cannot get away from. You may have changed a motor or added a metallic accessory, like an antenna or metallic decals.

I hope that makes things more clear. OH and BTW, declination only matters if you are navigating via TRUE NORTH instead of MAGNETIC NORTH. Then you must account for DECLINATION.
I'm betting the AC does not even care about TRUE north.

 

[GadgetGuy]

So do it right once, and leave it alone!

This part says it all!

 

[N017RW]

It's a 3axis magnetometer. The newer models actually have two.

The compensation routine is to cancel/normalize the on-board distortion from both hard and soft iron components. This was most significant in earlier Phantoms when the user could add OSD and filter boards, gimbals & cameras.

Notice the routine does not require motors active which would present a much more intimate source of magnetic fields.

 

[BudWalker]

The sensor in the AC is a 2 axis sensor similar to a HMC6352 device. We know this because the cal procedure is only specified in 2 axis. The following is extracted from that data sheet.

User Calibration
The HMC6352 provides a user calibration routine with the “C” command permitting entry into the calibration mode and the “E” command to exit the calibration mode. Once in calibration mode, the user is requested to rotate the compass on a flat surface at least one full circular rotation while the HMC6352 collects several readings per second at various headings with the emphasis on rotation smoothness to gather uniformly spaced readings. Optimally two rotations over 20 seconds duration would provide an accurate calibration. The calibration time window is recommended to be from 6 seconds up to 3 minutes depending on the end user’s platform. The calibration routine collects these readings to correct for hard-iron distortions of the earth’s magnetic field. These hardiron effects are due to magnetized materials nearby the HMC6352 part that in a fixed position with respect to the end user platform. An example would be the magnetized chassis or engine block of a vehicle in which the compass is mounted onto. Upon exiting the calibration mode, the resulting magnetometer offsets and scaling factors are updated.

If anything, to improve the result turn slowly. The AC firmware determines if it has sufficient data and is programmed for one turn. You can play with ANY device - phone, tablet, etc. - and see the effects of turning too fast or too slow. It will argue about it. It is the same device/process.

This whole thread is based on the fact that if you get a good cal - away from any "hard-iron distortions" - doing another cal in the vicinity of these distortions will make things worse, not better. The distortions are local and have nothing to do with "declination". Declination is determined from GPS co-ordinates referenced to a table. They change VERY slowly and would be compensated for with updated firmware - except your AC and you probably will not live long enough to see the change.

This is from thecompasstore.com explanation and chart
Below is an actual declination chart from 1990, showing the worldwide declination situation in 1990. The current declination values for today will be very close, if not identical,

When the AC is in the air, it is normally far enough away from any local distortions so that if you calibrated it near a distortion, it is now uncalibrated in flight. - Hence, you made it worse. Fly around power lines or metal structures and all bets are off.

So do it right once and leave it alone. The only reason for it to change is that components of the AC became more or less magnetized. Those are local distortions that you cannot get away from. You may have changed a motor or added a metallic accessory, like an antenna or metallic decals.

I hope that makes things more clear. OH and BTW, declination only matters if you are navigating via TRUE NORTH instead of MAGNETIC NORTH. Then you must account for DECLINATION.
I'm betting the AC does not even care about TRUE north.

As noted by @N017RW the Phantom series have 3 axis magnetometers. To see this look use the Go App sensors page. You can also look at the .DAT file. Both show X, Y and Z magnetometers.

The Phantom series uses true north for navigation. This can be seen by using CsvView/DatCon (obtained here). Yaw is set to magYaw (heading derived from magnetometers) at batteryOn. When GPS info is available Yaw is then adjusted by the declination derived from the GPS data.

I'm not sure what the reference to hard-iron distortions is about. Hard-iron distortions are due to anomalies on the AC. I.e., hard-iron distortions are not external to the AC. The "User Calibration" text that you included affirms this.

It's mathematically impossible for the compass calibration to incorporate and compensate for any external magnetic anomalies. This is addressed in post #167 and following. In particular, from post #172.

If the anomaly such as "tectonic or other surface-based magnetic anomalies" or "ore deposits on the earth's crust, tectonic plate composition, etc." is external to AC then it can't be discovered by the calibration. If the anomaly causes variations in field strength dependent on orientation then the anomaly is effectively ambiguous; change the calibration location by a small amount and the results of the calibration are different. OTOH, if the field strength doesn't vary then even the existence of the anomaly can't be known, let alone it's characteristics.

 

[Desertbear54]

I'm sure somewhere in this imformative thread someone has asked that unless you get a compass notification in the Go App you should not need to perform a compass calibration. That is what my manual states for my P4P. Should I perform it to just be extra cautious because I have traveled more than 700 miles and I have updated the fw twice and the bird has flown flawlessly with no compass calibrations notifications? Sorry if this has been addressed.(DB)

 

[N017RW]

I haven't. Travelled 600 miles (from cal location) and back with no issues.
Isn't the proof in your flights?