One of my task was to redesign the fingers. Nathaniel and I liked the original finger shape more or less, but not the actual dimensions. We wanted the the thumb to have a different thickness than the other fingers so that it could pick up smaller objects.
We had many struggles with this design because we wanted the fingers to do more presicion work, but also wanted to give someone the ability to still toss a ball around with their friends.
Our first problem was that I did not design the finger around the other part that connects to the knuckle, so when I tried to put the finger and the other part together in a rhino model the hole that were supposed to line up to put the pin through did not line up. I basically had to redesign the finger several times to get it just right, because either the holes were too small to put the string through or many other basic careless interference mistakes.
Because we wanted two different sized fingers, we also had to create a whole other finger from scratch. We wanted the thumb and the index finger to touch when you clamped we had to measure how long the index finger should be to be able to touch the tip of the thumb.
I learned to always model around pieces as to not mess up any measurments and to think methodically about which pieces to create first so that you did not have to work twice as hard to get the next pice done.
We had also found another problem with the original design of the RoboHand. We did not like how on all of struts there were two, and you had to put like ten pins on each side just to have that piece be pretty flimsy. So Nathaniel and I thought, why not just make it all one piece. The problem in our design for the knuckles though was that the knuckles were slanted, so when we printed the first prototype of the struts, it did not align right. We had to go back into Rhino and go and edit the struts to make one of the pieces longer.
Another problem we had was the connector piece to the knuckles. This piece was not hard, but I just was not thinking when made it. I made this piece after the finger, which was the wrong choice, because the piece did not give the finger maximum angle to rotate. Also there was interference with the knuckles, and it gave it almost no angle to rotate. Eventually I made the right decision and modeled the connector piece around the knuckles and the fingers.