potentiometer for angle measurement

Discussion in 'The Projects Forum' started by Gadersd, Apr 9, 2014.

  1. Gadersd

    Thread Starter Member

    Dec 8, 2012
    98
    1
    I am trying to design a circuit that will get the angle of someone's arm. My first thought is to place a potentiometer on a hinge and attach the hinge to my arm. There will be a voltage to binary converter that will read the potentiometer's voltage output and output a binary value representing the angle of the arm. I am worried that the potentiometer will have low accuracy. Is there a better way that will give me better accuracy, or is my idea fine? I don't need it to be very accurate.
     
  2. #12

    Expert

    Nov 30, 2010
    16,248
    6,745
    Potentiometer accuracy has always been bad. You can "calibrate in software" or use an op-amp with offset and gain adjustments to convert anything that is pretty linear to a useful result.
     
  3. wayneh

    Expert

    Sep 9, 2010
    12,087
    3,026
    It'll work. Just be sure to use a linear pot (not "audio" or "log"), and then calibrate your device for angle versus voltage. You should get a fairly straight line. Don't forget to protect the pot from excess current. You can count on someone turning the dial all the way to either side.
     
  4. ScottWang

    Moderator

    Aug 23, 2012
    4,850
    767
    Using an encoder to counting the number, forward as count up and backward as count down.
     
  5. wayneh

    Expert

    Sep 9, 2010
    12,087
    3,026
    For what it's worth, there are dedicated angle sensors. The pot approach will not likely be better than ±5%, or maybe ±5°. Maybe better if you have a good calibration curve. I haven't used the sensors but I would imagine ±0.5° is not too difficult.
     
  6. MrChips

    Moderator

    Oct 2, 2009
    12,415
    3,354
    My opionion is that the potentiometer will work well for your application.
    If you wire the potentiometer as a potentiometer (duh!) you wouldn't have a problem with excess current.
     
  7. #12

    Expert

    Nov 30, 2010
    16,248
    6,745
    Or drive it with a constant current source.

    Many ways to skin the cat.
     
  8. MaxHeadRoom

    Expert

    Jul 18, 2013
    10,500
    2,365
    Accuracy but what resolution?
    What are you using to gather and display? Pic etc?
    Look up wind vane (360°) sensors, also as well as the encoder solution there is the resolver, you can use the sin/cos signals for any resolution.
    Max.
     
  9. #12

    Expert

    Nov 30, 2010
    16,248
    6,745
    Probably in keeping with the idea that most people can't tell if their elbow moves one degree differently from yesterday.
     
  10. ifixit

    Distinguished Member

    Nov 20, 2008
    638
    108
  11. #12

    Expert

    Nov 30, 2010
    16,248
    6,745
    Where do you plug it in? :D
     
  12. Brownout

    Well-Known Member

    Jan 10, 2012
    2,375
    998
    I vote for a pot. But make sure you get a precision pot, and drive it with a precision voltage (or current)
     
  13. atferrari

    AAC Fanatic!

    Jan 6, 2004
    2,644
    759
    Hard to use, this is the best. It would make for additional arm training as a bonus.
     
  14. djsfantasi

    AAC Fanatic!

    Apr 11, 2010
    2,790
    827
    @atferrari - I like this one.

    @Gadersd - It looks like the turn screw is solderable, allowing one part of the arm to be attached to a small circuit board holding the pot; soldering a brass arm in the slot can be used for the other arm segment to attach.

    With a 12 turn 10kΩ trimmer, and an arm motion of 180°, the range of values is 0 - 416Ωs. Each degree corresponds to a ~2.3Ω change. I don't k now if this is sufficient resolution for your application; I expect it would be.

    I can draw up a sketch if it would be useful.
     
  15. djsfantasi

    AAC Fanatic!

    Apr 11, 2010
    2,790
    827
    Or this single turn trim pot gives ~13.8Ω per degree of arm movement around the elbow (this is a 5kΩ pot; there are larger single turn pots with an extended handle available). The handle axis becomes the hinge between the arms.

    Other trim pots, single turn, higher values can be found on Jameco here.

    I thought further, and velcro straps can be used to hold the measurement arms to the person's arm.
     
    Last edited: Apr 10, 2014
  16. burger2227

    Member

    Feb 3, 2014
    190
    24
    OK, how do you go about attaching the pot to one's arm?
     
  17. #12

    Expert

    Nov 30, 2010
    16,248
    6,745
    You don't. You attach the pot to 2 long pieces, probably metal, and strap the metal to the human. I like DJ's idea of using a circuit board to hold the pot. I don't like the idea of using trim pots. Too fragile and a common 1 turn pot will cover the required range of motion.
     
  18. Brownout

    Well-Known Member

    Jan 10, 2012
    2,375
    998
    The arm has an 'elbow' where one member pivots with another member. Affix the pot body at that exact pivot point, the body attaches to one member and the shaft attaches to the other member. Now, you only need to figure out how to make the physical attachments. Building robots is all about solving issues like this.

    Note: only do it this way if it's a 1-turn pot.
     
Loading...