Hi. I have a question about Mag calibration.
When i give board movement to Filter, it is changed as follows. and i become to see that it is calibrated and I can get a accurate value.
is there any standard for Mag calibration?
For example, for Mag calibration, I have to move a specific sensor about 180 deg
Thank you.
Solved! Go to Solution.
There is a defacto standard for the form of the solution: 3x3 inverse soft iron plus 3x1 hard iron offset. But there is no standard for the process of computing magnetic calibration coefficients, not even a defacto one. The Freescale magnetic calibration functions utilize a "total-least squares" fit to compute the inverse mapping required for hard and soft iron calibration. Basically, you need to collect enough data points on the measurement ellipsoid (see http://blogs.freescale.com/sensors/2011/03/hard-and-soft-iron-magnetic-compensation-explained/) so that the solver can reach a numerically significant result.
There is a defacto standard for the form of the solution: 3x3 inverse soft iron plus 3x1 hard iron offset. But there is no standard for the process of computing magnetic calibration coefficients, not even a defacto one. The Freescale magnetic calibration functions utilize a "total-least squares" fit to compute the inverse mapping required for hard and soft iron calibration. Basically, you need to collect enough data points on the measurement ellipsoid (see http://blogs.freescale.com/sensors/2011/03/hard-and-soft-iron-magnetic-compensation-explained/) so that the solver can reach a numerically significant result.
Thank you very much
Hi Mike,
Could you let me know when you are going to fix the link for " http://blogs.freescale.com/sensors/2011/03/hard-and-soft-iron-magnetic-compensation-explained"?
I checked the other link like matrix computation as well. They are all not accessible now?
Thanks,
Christie
Christie,
It looks like it got dropped when they merged the Freescale and NXP sites. I don't have direct access to the site, and will have to request it be reposted when I return to the office in January. In the meantime, I'll post a PDF of the old site in another post. It's not pretty (the PDF formatting leaves a lot to be desired), but should help fill the gap.
Mike