Robotic finger

AImanAIman Posts: 499
edited February 2007 in Robotics Vote Up0Vote Down
I have been working on a humanoid robot for some time and was able to get the protoype finger to move fast enough to strip the gearing from its servo. The finger·is controlled via fish line (50lb test) to contract and rubber bands to pull it straight.·Motion control has been tested to much better then .10 of a second. The original prototype is made out of 1 inch wood doll rod with metal eyelets for cable running and rubber band attachment.

Plans are already made to build a finger from PVC and run cables via the interior. PVC was chosen as the second version because of its light weight, cheap costs and availability. Version 3 is in the works which will be built out of milled aluminum, but is not now being built·because of re-occuring issues of the fish line getting stuck if the finger is in the wrong position. This probelm was solved by restringing the fishline through washers. Crude but effective. Version two is meant to work out design flaws prior to makeing version 3 which will cost more.

There is nothing special about·the programing other then bumping up the PWM to around 5. The BS2 so far has been able to complete all opperations. The servo holds the finger nicely and the tension from the rubber bands allows the hand to curl like a humans when not in use.

Cables with servos were selected·over motors, pnumatics· and hydraluic power for several reasons. Motors are harder to locate in the space of a humanoids hand or arm and get them to function correctly. Pnumatics (Air Power) was rejected because of complexities involved with valves, locations to place the compressor, power·and bounce. Pnumatics have power and speed but wear out faster and when contracting will stop and bounce back and forth slightly. Hudraluics were rejected for lack of space for the pump, power usage and valves. Because of these things cables with servos were selected even though the stringing of cables can get complexe at times. The other benefit is that cables and servos are easy to come by where getting small valves and cylinders can be challenging at best. Research into using pnumatics or hydraluics included checking out pricing for unusual parts and it was found to be rather high.

Post Edited (AIman) : 10/9/2006 7:37:28 PM GMT


  • 10 Comments sorted by Date Added Votes
  • Tim-MTim-M Posts: 522
    edited October 2006 Vote Up0Vote Down
    Sounds like a great project... we'd love to see some pictures!

  • AImanAIman Posts: 499
    edited October 2006 Vote Up0Vote Down
    I will try to get them out soon but no promises, right now I am working on a handful of projects at the same time.
  • AImanAIman Posts: 499
    edited October 2006 Vote Up0Vote Down
    By the way, I found a program to balance a robot on two wheels and that will fit into this project perfectly.
  • AImanAIman Posts: 499
    edited November 2006 Vote Up0Vote Down
    Ran into a few issues with the PVC finger.

    First problem is that the cables on the inside either get frayed or gouge the PVC. So that will need to be resolved.

    Second problem is fastening the stuff on the inside. Glue is good but comes off easy. To screw some thing in place means cutting the pipe in half and then figuring out how to put it back together. On bigger pieces that isn't a problem, but trying to work in the space alotted creates unforseen obsticals.
  • Tommy BotTommy Bot Posts: 60
    edited November 2006 Vote Up0Vote Down
    Any R/C shop for airplanes will have what you need. I forget what they are called but it is a piece of platic rod inside of a plastic tube. They are used to connect the servo arm to the actual control surface i.e. ailerons, flaps, rudder, elevator. The outer sleave is secured in place with glue and sometimes kevlar thread, and the inner rod moves freely. You can even curve it to some extent, and they come in many sizes.

    Of course if you wanted to use these with the fishing line, you would need a push pull system, maybe two per "finger".
    Might help.

    (Frequently heard from other's)

    Tommy, I know it wasn't designed to·x, but can you make it·do x·anyway?

    Post Edited (Tommy Bot) : 11/7/2006 7:24:10 PM GMT
  • AImanAIman Posts: 499
    edited November 2006 Vote Up0Vote Down
    I thought of using a·push rod, however the problem wasn't getting them to push but getting them to go around a finger. For example, a knuckle can curly upto 90 degrees or more. to do the tip of the index finger means no less then 180 degree turn from start to finish. I went with fishline/cables to avoid binding issues and extra force requirements from solid rods.

    So far I haven't been able to come up with any system that can move a finger without using bidirectional forces. The one uni-directional system used, required rubberbans or springs to pull the fingers straight and that was very klunky.
  • Chris SavageChris Savage Parallax Engineering Posts: 14,406
    edited November 2006 Vote Up0Vote Down
    ·· This isn’t an already completed project?

    Chris Savage
    Parallax Tech Support
    Chris Savage | Engineering Tech | Main Office: (916) 624-8333 | Direct to Tech Support: (888) 997-8267 | Website | Twitter | Google+
  • AImanAIman Posts: 499
    edited November 2006 Vote Up0Vote Down
    The finger is completed and working.

    Figured since there was interest in the finger I would let people know whats going on as the newer versions are developed·and built.
  • willy1067willy1067 Posts: 107
    edited February 2007 Vote Up0Vote Down
    I am also working on a anatomical robotic hand, my project uses only one motor to control al 20 finger joints. the wrist requires more force, so I will create another method to drive it.· here are some pictures.

    I just added two new pictures, machining the rest of the hand this weekend.

    Fernando Gomez

    Never compare yourself with anyone else, there will always be someone bigger·or·smaller·than you.

    Post Edited (willy1067) : 3/27/2007 9:59:15 PM GMT
    1744 x 1328 - 799K
    2048 x 1536 - 623K
    1363 x 2022 - 231K
Sign In or Register to comment.