Recording Magnetic Heading on a UAV

[air]

Hi,
I'd like to be able to record the magnetic bearing that a UAV has as it flies between waypoints.

For example say that the plane is tracking perfectly such that the heading is very close to the bearing to the next waypoint all the time but there's a crosswind so that the plane is crabbed into the wind by 20 degrees during the time it travels between waypoints.

I'm looking at perhaps carrying a garmin etrex legend hcx which includes a magnetic compass but I've yet to find out if it records the compass bearing seperately in the track file along with the GPS heading.

Also has anyone got experience in using a magnetic compass in an electrically noisy environment like a small electric UAV?
I know some people use them on mikrokopters.
Essentially I guess I'm just looking for a gps track logger with a magnetic compass log also.

Thanks.

[matttay]

I'm not sure what the magnets on the motor might do if things are tight. The earth's field is about half a Gauss, and the small magnets used in motors are probably 10's if not hundreds of Gauss. So you'd need to take that into account. If you have the stuff on hand, perhaps you could bring a compass near a running plane and see if it changes. I've also thought it'd be useful to have a compass on board, and there are some very small ones that are getting smarter every day.

PS. Is Kerry County the location of the Kerry mountains referred to in the song Whiskey in the Jar?

[Tom Harper]

I have the Honeywell compass but have never been able to get it to work.

Anybody around who knows the magic?

Tom

[matttay]

The are very sensitive to tilt. Is that perhaps part of the problem?

[air]

Hi Mattay,
Yes, thats what I'm worried about. I can seperate the unit from the motors by about 50cm, I'd be interested to hear from someone that uses a magnetic heading reference system, I know the Carvec GPS system uses it as well as mikrokopter units with GPS position hold as mentioned.
Tom, which Honeywell unit did you try (worked for them for a bit ;) ). I think some of the ones they have are quite robust as regards their orientation.

@Mattay - Yes on the Whiskey in the Jar issue! I was born in Cork to complete the reference. "Between the Cork and Kerry mountains". Great song.

[dmgoedde]

I can offer my thoughts, and soon my experience. This is an area of research for me at the moment.

Ideally, you'd want a 3-axis magnetometer, because the Earth's magnetic field is approximately parallel to the ground only at the equator. On top of that you need to know the plane's pitch and roll, and the latitude. Using all of this info, you would then be able to solve for the magnetic vector that is the local North. Not trivial!

[air]

Hi Dean,
Thanks for that, yes I know they are lots of fun!
Even better is that hte magnetic declination for a fixed location changes all the time! I recently read a post where a guy mentioned that he found a chart for his area drawn up in the 60's that was now 4 degrees out!

Tom, which compass did you try out? I was considering a HMR3300 some time ago & I see they're still on the market.
Ideally I'm looking for a ready to go solution that is already interfaced to a data logger, I havent figured out yet if there are any consumer type devices that would fit the bill.

[DrDon]

I agree with dmgoedde that the derivation and processing for compensation for tilt of a magnetometer is NOT trivial. We have developed a complete IMU system with integrated GPS/IMU/Compass.

I highly recommend it if you ever get a chance to play with 3D magnetic sensors as the readings will surprise you. I have included a plot of our inflight results of comparing GPS Course over Ground and Compass Heading.

DrDon
http://www.iat-llc.com
Automation Sensors, Systems, and Services
http://www.iat-llc.com/COGvsCompass.jpg

[kbosak]

I have the Honeywell compass but have never been able to get it to work.

Anybody around who knows the magic?

Tom
Their I2C uses clock stretching. Must use pullups on SCL also -that's the magic.
This calls for very good I2C library (rare) or using hardware TWI/I2C interface in the uC. They also tend to be less accessible after switching to continuous measurement mode. Getting Raw X and Raw Y data from them lowers the bandwidth significantly, making tilt correction very difficult in our applications.
You also need to know the tilt and local value of Inclination in order to compensate virtual Z-axis for tilt.
Knowing local declination might be useful also.

[kbosak]

On top of that you need to know the plane's pitch and roll, and the latitude.
Not exactly the latitude, but declination (sometimes very small, unless Canada), and inclination (that may or may not be close to latitude).

[kbosak]

I have included a plot of our inflight results of comparing GPS Course over Ground and Compass Heading.
...and they seem to agree very well, given the fact that COG and HDG differs because of wind. I would say it is 'piece of cake', until I see larger plot.