Shop OBEX P1 Docs P2 Docs Learn Events
magnetic field sensor — Parallax Forums

magnetic field sensor

sport270sport270 Posts: 82
edited 2016-05-12 14:35 in BASIC Stamp
Need to make a non contact sensor that would put out 0-5v over a 2" range of motion. Motion Is 2 steel plates overlapping with about 1.5" gap .
Sensor will be mounted around 8" from pivot point ...so small arch pattern of rotation.
I made a simple voltage division circuit and them attached 10 glass reed switches in it , each with a different resister value in line. Theory was to accumulate voltage in .5v increments with the closing of each switch. And would accumulate the total of the 0previous closed switches. Worked great until I couldn't arrange magnets in way that would keep them all closed with full overlap. It would drop some off. Thought about spreading the reed switches apart and having each individual one be a higher value stepped in .5v steps .
And again I can't get just one switch to always be closed. Sometimes its 2 or 0.
Found a magnetic field transducer AD22151 but can't figure out what the output would be...
Honeywell SPS-A100HAWS
would be perfect except it's made for rotation inside the other arch. Where I need the sensor to face the magnet and overlap in rotation over 2" in movement ..... Anyone have any ideas.

Comments

  • Chris SavageChris Savage Parallax Engineering Posts: 14,406
    Does it have to be magnetic?
  • ercoerco Posts: 20,257
    edited 2016-05-12 15:43
    How about a sketch of what you need? Electrically it sounds like a variable capacitor with 2 steel plates moving wrt each other.
  • kwinnkwinn Posts: 8,697
    erco wrote: »
    How about a sketch of what you need? Electrically it sounds like a variable capacitor with 2 steel plates moving wrt each other.

    Like erco posted, how about a sketch. I can think of several methods to do this but they depend on the exact geometry and motion of the steel plates.
  • sport270sport270 Posts: 82
    Doesn't necessarily have to be magnetic.
  • sport270sport270 Posts: 82
    In the 2nd pic you can see the sketch of where the reed switches layout.
    3rd pic is of the stacked bracket that protects the reed switches on pc board.
    4th pic is with the outer cover that will ramp up to the 3/4" height.
    image.jpg
    2052 x 1532 - 471K
    image.jpg
    2052 x 1532 - 272K
    image.jpg
    1532 x 2052 - 329K
    image.jpg
    1532 x 2052 - 450K
    image.jpg
    2052 x 1532 - 343K
  • sport270sport270 Posts: 82
    Wanting ideas of anything with a more guaranteed signal.
  • ercoerco Posts: 20,257
    edited 2016-05-17 02:24
    Looks like a whole lotta steel goin' on there. Maybe your reed switch idea is the best.

    Reed switches have a lot of idiosyncracies, hysteresis and favored orientations for best and most reliable triggering. Did you experiment with rotating the switches in all axes to check for asymmetric make & break?

    Please show a wiring diagram of your multi-resistor voltage divider & reed switches. Seems like a strong enough neodymium magnet should close all switches as it sweeps by. Wired properly, multiple switches closing simultaneously isn't a problem. At that point, you just need to fine-tune the reed switches' location and orientation for proper function.

    Is your magnet fairly isolated (not buried in steel) so the field is very concentrated? That's important. Using some green magnaview paper is helpful to visualize the field and flux lines.
  • kwinnkwinn Posts: 8,697
    That's the kind of thing that is normally done with an LVDT sensor like this, but it requires physical contact to move the plunger in the tube. You could probably build a non contact version since the basic sensor consists of 3 coils, but it would take a bit of experimenting and some circuitry to get it to work for your application.
  • Beau SchwabeBeau Schwabe Posts: 6,568
    Personally I would go with something a little more industrial and reliable for that application ...

    Something like this ...
    http://www.midoriamerica.com/products_detail.php?eid=00036&category=Tilt+Angle+Sensor&type=Contactless
  • stamptrolstamptrol Posts: 1,731
    Definitely an industrial type installation and I'm assuming more than $1.95 is available to implement this project.

    Two things come to mind.

    The first is a Temposonics type measuring device that is inserted into the hydraulic cylinder and gives a direct analog readout of the position of the cylinder which can be easily related to the main boom and stick boom relative positions.

    The second is a resolver unit such as those called Duropot made by Astrosystems. The unit acts like a pot at the hinge joint except has no sliding contacts or wearing parts and is totally waterproof.

    Along the same lines, you might be able to adapt a one-revolution absolute encoder unit at the hinge point and generate a repeatable position signal.

    Cheers,
Sign In or Register to comment.