Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

What magnetic position sensor to choose?

  1. Jul 12, 2017 #1
    I want to record the displacement of a small magnetic ball moving sideways across the face of the sensor. I've had a close look at AMS's catalog of magnetic position sensors and I'm not sure how to choose one.

    If I'm working with a movement range of <5mm and want a resolution of 5-10μm, and also want one that can handle wet conditions, which would I choose? I assume a linear position sensor would be preferable to an angular one, but from there I'm unsure.

    This is the list http://ams.com/eng/Products/Magnetic-Position-Sensors
    Perhaps another manufacturer would be better?
  2. jcsd
  3. Jul 12, 2017 #2


    Staff: Mentor


    I don't know much about those sensors, but I wonder if you could consider optical measurements and ignore that that the ball is magnetic.
  4. Jul 12, 2017 #3
    I could use optical measurements, but they'd have to run continuously for many measurements. I've also found Hall effect sensors to be cheaper. What sort of optical sensor would you use?
  5. Jul 12, 2017 #4


    Staff: Mentor

    I can think of two.
    1. Video camera with image analysis. You could have a graded scale in the image background.
    2. A laser range-finding device. (Analogous to a surveyor's range finder, but at a much smaller scale.)
  6. Jul 12, 2017 #5


    User Avatar

    Staff: Mentor

    As the ball rolls, the magnetic field will spin as well, right? It seems that the changing magnetic field would make position measurement difficult...
  7. Jul 12, 2017 #6
    The only problems with these are that these both record distance from the sensor, where as I need something recording movement that is parallel to the sensor. I also would like something small. I thought an IC would be ideal.

    The ball won't be rolling, it will be translating only. It's similar to the head of a pin.
  8. Jul 12, 2017 #7
    Hi, just so I understand, you are looking to measure sideways movement of a small object, "Head of a pin" this object can move up to 5mm from left to right (in relation to the sensors view point) I'm assuming the object cannot move up or down or closer or further away from the sensor?

    I cannot think of anything that would reliably suit your application, optics and cameras are an option if you have the know how or patience and the budget, however I assume you were looking at magnetic sensors due to the "wet environment" which would interfere with an optical device.

    There may be another way around to solve your application, but it would require further explanation of what you are ultimately trying to detect.
  9. Jul 13, 2017 #8

    Thank you for your response, I think you've clarified my question quite a lot. I only want to record the displacement on one axis, so moving up/down or toward/away from the sensor does not really matter. But it should only move left-to-right.

    Other than using magnetic Hall effect sensors, I've considered using linear potentiometers. My concern with potentiometers is how reliable they are on the micrometer scale.
  10. Jul 14, 2017 #9


    User Avatar

    Maybe... but probably not the one I used. [COLOR=#black].[/COLOR] :oldgrumpy: [COLOR=#black].[/COLOR] lol

  11. Jul 16, 2017 #10


    User Avatar
    Gold Member

    The idea of using a camera will work, but there are some things to be considered:

    - The images must be of good quality ( e.g. white ball, black background, good illumination ).

    - The image must be at least 1000 x 1000 pixels

    - The lens will distort the image ( lens distortion ) so that the image will not be dimensionally stable. You will have to calibrate the lens and develop some algorithm that converts the recorded image to a dimensionally stable image. The distortion will be radially.

    What you will have to do is:

    - Record an image.

    - Locate the edge of the ball ( the edge pixels ) by means of a spatial filter.

    - Convert the locations of the edge pixels into dimensionally stable locations.

    - Locate the the position of the ball by means of a Hough transform.

    Due to the statistical nature of the Hough transform, you will be able to determine the position of the ball within 0.1 pixel. The method is very robust, but a lot of computer power is needed ( the calculations are simple, but there are 1 Megapixels ). An "IC" will not do it, unless it's a huge FPGA-device. I think that a PC will do the job within about 2 sec., depending on the number of edge pixels.
  12. Jul 16, 2017 #11


    User Avatar
    Science Advisor

  13. Jul 17, 2017 #12


    User Avatar
    Gold Member

    Well, yes, but a line imager only sees a line.

    Say that there is a disturbance in the image, e.g. an unwanted reflection near the edge of the ball and on that line, the line imager will be completely confused: The position of the ball will not be measured correct.

    In a spatial image most of the circumference of the ball can be seen, and this greater amount of data causes that a disturbance can be ignored no matter where it occurs: The position of the ball will be measured accurate anyway.
  14. Jul 24, 2017 #13
    For those curious, the AS5510 is a very good magnetic linear position sensor for small resolutions. It is generally paired with an AS5000 magnet.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted