Monday, August 2, 2010

The telepresence finger: Take 2

I redesigned the prototype finger over the weekend. The controlling tendons are now inside the finger structure out of the way instead of hanging loose. I have a single tendon across the top side of the finger as you can see here.

Dual tendons run across the bottom of the finger and are exposed now only at the joints.

For scale, the prototype finger measures 100 mm from knuckle to fingertip, the length of my own index finger.  It is slightly wider than my finger at this point.  There is nothing to stop me from narrowing it down, now that I have the design issues pretty much addressed.

I reduced the joint shaft diameter slightly to reduce the need to clean the part after printing.  As well, I've adjusted the joint resistance to increase with each joint from the first to the third (knuckle).  This enables the finger to contract beginning at the fingertip and curling down in a natural manner as the inside tendons are pulled.

Full contraction or extension requires roughly 38 mm extension of the affected tendons.  Given that I am initially using a 120 degree servo motor I will be requiring a servo rotor of about 18 mm diameter to effect the full range of motion for the finger.

My decision to use standard plastic model airplane cement (butyl acetate) seems to be a great success.  The joints are permanent with ABS and I avoid the whole drama and expense of sourcing #4-40 (equivalent to M3) nuts and bolts.  Unlike epoxy, butyl acetate joints set within a minute and are ready for use within an hour.


  1. This looks really great, Forrest! I really like the improvement from your first model to your second. It's drastic. Are you gonna do a third to narrow it down to a human-like width before continueing on?


  2. I'm going to try to rig what I've got to a servo first and see if I've got enough torque to make it do what I want. After that, I plan on taking the finger through another design iteration.