Opto Joint Angle Sensor!

Discussion in 'The Projects Forum' started by jegonz, May 13, 2008.

  1. jegonz

    Thread Starter Member

    Jan 25, 2008
    19
    0
    Hello, I was hoping some one can guide or advise me with this project I want to make.
    The thing is that I have a robot hand and I want to measure the angle of each of the joints movement. Of course I tried the bending sensors, but I have 2 problems: Precision and Friction. So I was thinking that maybe infrared optics will give me a good result, because there will not be friction and the precision I might be able to make it good.

    1. using a led and a photodiode or photo transistor in the other side. The linear distance that will move is about 1mm to 6mm. attached file sensor idea

    2. using a reflective distance (1mm to 6mm) sensor and a bend reflective surface to maintain the refective distance as straight as possible. attached file sensor idea 2. for this one I will use a distance sensor from OSRAM SFH 9201.

    I'm not sure how linear I can make the output of the sensor, but I guess it will be enough. I will build up the circuits and test it to see how it goes, but any advice will be appreciated! When I finish the test I will post the circuit design and the results!

    Thanks for the comments! :)
     
  2. beenthere

    Retired Moderator

    Apr 20, 2004
    15,815
    282
    That may not be a good means of reading ange as a function of illumination. incident light can have a major influence on the absolute illumination of the photosensor. A change in incident lighting will affect the sensor and make it impossible to determine if the angle has changed.

    If your joint is constrained to less than 300 deg rotation, then a potentiometer can read the angle as a function of voltage between 0 and 5 volts. Converting the wiper voltage lets the computer see the angle,

    There are other methods, but this is pretty easy to rig and is inexpensive.
     
  3. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    What Beenthere said ;)

    Along with that though, have a look inside a PC mouse. There are slotted wheels that pass through IR sensors. Any time a slot in the wheel passes by the sensors, a signal is output. It's not just a single blip, either - the pulse train is somewhat complex. On the front end of the signal, there are a couple of synchronization bits, and then a directional bit. Somewhere in the middle of the pulse train is a window that has a varying width, depending upon how many pulses have gone by. If you have an O-scope, you might take a look at the output signal and consider scavenging the guts of a cheap mouse to get the arm angle input for your microcontroller.
     
  4. jegonz

    Thread Starter Member

    Jan 25, 2008
    19
    0
    by incident lighting change, you mean a change in the intensity of the source LED or ambient (or external) light? I don't understand that very well.
    I guess the problem will be the same for both ideas I have?
    Because if you mean external light, well just cover up in a box the sensor.
    The POT idea is the easiest I guess for the electronics, but then we will have the friction and to build up the mechanical gears to move it. So at the end will not do what i need. I though of the mouse wheels too SgtWookie, I even already tear appart one mouse to take the parts, but I need to know direction, so there are several opto-chips that act as encoders

    ( http://docs-asia.electrocomponents.com/webdocs/0028/0900766b800286ff.pdf
    )

    that will do the job, but also was thinkning that it might be a little difficult to build the mechanical part... to attach the wheel and all this, that is why I thought of doing the sensor the way I posted. So if you tell me that the idea is not very good... then to attach a wheel to the joints! :)

    Thanks!
     
  5. beenthere

    Retired Moderator

    Apr 20, 2004
    15,815
    282
    By all means, conduct an experiment with the optical sensing. That should be very easy to do. The formula for a disaster is to go directly from an idea to the finished product without breadbording and experimenting.
     
  6. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    As I mentioned in my first reply, there IS a bit that indicates direction! In the mouse I looked at, it's like the 3rd bit in the train.

    A potentiometer would certainly be an easier option, and it would give you about 270° of travel. However, pots get "noisy" as they age/wear, and that would likely cause your readings to jump all over the place. A small cap to ground would help smooth the bumps out.
     
Loading...