Magnetometers

[DGN]

Not surprisingly there's not a forum called 'Magnetometers' so I hope I'll be forgiven for posting here instead.

I'll be using the MAG3110 magnetometer in a surveying application.  This application will use the magnetometer to find compass heading from the earth's magnetic field.  To get the best accuracy the sensor should not have any ferromagnetic materials nearby.

My question is how diligent do I have to be in keeping magnetic components off  the circuit board?  If I'm forced to use them then how far away do they have to be from the sensor to minimise their influence?

I've source non-magnetic resistors and capacitors as well as a low magnetic signature battery.  But non-magnetic crystal oscillators seem less common.  I don't want to spend lots of dollars and time on sourcing non-magnetic components unless there is a real benefit, but I don't know how to quantify that benefit.

[bigmac]

Hello, and welcome to the forum.

When choosing components I would try to avoid those using ferromagnetic materials.  However, I would not have thought that standard SM resistors and capacitors would have been an issue.  Do you really need to use a crystal?  Can you make use of the internal reference within the MCU (assuming a MC9S08 MCU type)?  If you need an accurate bus frequency, perhaps the use of a low frequency watch crystal would be better than a higher frequency crystal.

I think that you will ultimately need to experiment with the placement of any suspect components, to test their effect on the 3-axis measurement.  I would certainly check the intended battery placement.

Another consideration might be the placement of PCB tracks in the vicinity of the sensor, especially for those carrying any significant current.  I guess there should not be any of these tracks immediately below the sensor.  For a PCB track that passes the sensor at a distance of 10mm, the flux density at the sensor position caused by the track current would be about 0.02uT per mA, using the formula B = u0*I /(2*pi*d).  Compare this with the sensor sensitivity of 0.1uT per LSB.

Regards,

Mac