See also our Sensor Comparison page
A magnetometer measures magnetic fields. In this case, we're measuring the Earth's magnetic field and thus using it as a "three dimensional compass" to get an absolute attitude (direction) reference to aid our navigation algorithms. Note that what the magnetometer actualy gives you is a 3D vector (magnitude and direction) of the local magentic field; it's up to you to interprete what that actually means since the vector of the Earth's magnetic field is position dependent (and, surprisingly, time-dependent).
We'll most likely be using Giant Magneto Resistive (GMR) sensors because of their small size and weight. As far as I understand, they have a meagneto-resistive material which changes resistances based on the local magnetic field. Flux gate magnetometers are the other type: Basically coils which are driven by sine waves in saturation, and the nonloinear affects (due to the saturation) from the local magnetic fields can be measured. It looks much, much harder to deal with than GMR, as well as much larger and more power hungry, but it's probably more accurate and has greater range. Since we're just measuring the Earth's magnetic field, we don't have a huge range and we're not terribly interested in super-high accuracy.
Questions to ponder: What resolution do we need? what accuracy? What kind of bandwidth do the sensors have? What kind of update rates can we get? We obviously need to mount the thing in the nose cone... what other precautions do we need to take to make it accurate?
JAP:Which A/D to be used? 'The ADS1158 has been proposed.'
The Earth's magnetic field resembles that of a simple bar magnet. This magnetic dipole has its field lines originating at a point near the South pole and terminating at a point near the north pole. These points are refered to as the magnetic poles. These field lines vary in both strength and direction about the face of the earth. In North America the field lines points downward toward north at an angle roughly 70 degrees to the earth's surface. This angle is called the magnetic angle of inclination. The direction and strength of the earth's magnetic field can be represented by the three axis values Hx, Hy, and Hz. The Hx and Hy information can be used to determine compass headings in reference to the magnetic poles.
The Earth's magnetic field magnitude is about 0.5 to 0.6 gauss and has a component parallel to the Earth's surface that always point toward magnetic north (which is why compasses work).
Aircraft convention defines the attitude parameters in terms of three angles: heading, pitch and roll. These angles are referenced to the local horizontal plane. That is, the plane perpendicular to the earth's gravitational vector. Heading is defined as the angle in the local horizontal plane measured clockwise from a true North (earth's polar axis) direction. Pitch is defined as the angle between the aircraft's longitudinal axis and the local horizontal plane (positive for nose up). Roll is defined as the angle about the longitudinal axis between the local horizontal plane and the actual flight orientation (positive for right wing down). If a compass was sitting in the local horizontal plane, then the roll and pitch angles would be zero and the heading would be calculated as:
Heading = arcTan(Yh/Xh)
where Xh and Yh represent the earth's horizontal magnetic field components. As the aircraft is rotated, the heading would sweep 0 to 360 degrees referenced to magnetic north. If the compass were now tilted, the tilt angles (roll and pitch) and all three magnetic field components (X,Y,Z) must be used in order to calculate heading.
ADG: Where's the LV2 reference frame paper?
Heading accuracy is affected by:
- A/D converter resolution
- Magnetic sensor errors
- Temperature effects
- Nearby ferrous materials
- Compass tilt errors
- Variation of the earth's field
- Sample rate from sensor = 200 Hz.
- Use THS41XX Differential amp (single supply???),(slew rate??).
- Use feedback (DAC to current source)
- Multiplex feedback for X and Y axis.
- Possibly use internally mux amp stage of ADS1258.
- Over sample at A/D.
- Use constant current supply (magnetometer).
The magnetometer will give us the direction of the magnetic field relative with respect to the vehicle frame of LV2. But having the magnetic field direction doesn't give us attitude. To find that, we need a model of the Earth's magnetic field given a position and time. Here's a link to our model page:
They make almost exactly what we want: The HMR2300r strapdown 3D magnetometer. Here's the data sheet: http://www.magneticsensors.com/datasheets/hmr2300r.pdf. But, it's $800 and talks RS-485 instead of CAN.
The next step down is their HMC2003 3D analog module which has the sensors and instrumentation grade op-amps on the small module board. It's a nicely put together module, but it's $200.
Finally, we get down to their sensors. They have a bunch of types, and they're all about $20 each. So that's $60 for the three sensors. They have a $75 ball grid module which is 3-axis, and for $100 you can get it pre-mounted on a 16-pin DIP socket.
Frankly, I'm not sure where the best cutoff is. My guess is that it might be worth it to get the $200 module and just handle the ADC and CAN interface from there. Or, with a bit more pain, we can go for the cheaper sensors and do our own amplifiers.
10/15/2008: Our current choices in order are: HMC1043 or HCM1053 (what the heck is the difference). We ordered samples from our local Portland rep., Delta Technical Sales 503.646.7747
- http://www.semiconductors.philips.com/pip/KMZ51.html (appnote: http://www-us.semiconductors.philips.com/acrobat/various/SC17_GENERAL_MAG_98_1.pdf)
- Yamaha YAS529
- Alps HSCD series
Small GMR projects:
- http://server4.oersted.dtu.dk/research/CSC/publica/ - Mostly research grade flux gate sensors
- http://kadc.kao.re.kr/paper/jass/pdf/17/17_35.pdf - Korean university flux gate magnetometer flown in a sounding rocket
- http://www.ips.gov.au/IPSHosted/NCRS/wars/wars2000/commh/bish.pdf - A nice space-grade flux gate built for
|2007-03-28||Jason, Dan, Jonathan, Tim, Andrew on the next generation mag.|
|2004-06-30||Tyler and Andrew try to try out the digital sub. method|
|2004-05-02||Tim, Tyler and Andrew meet and do the Spencer Webb Design Cycle: start simple, go completely nuts, and then simplify.|