For the final iteration of my inflatable tentacle I decided to attempt one additional feature than the original four air chambers I had conceived. I decided to try and add an additional air chamber in the central cylinder in order to lengthen the tentacle when inflated. I cast the four quarters first. I then used a pvc tube to cast the central cylinder. I used a small bamboo skewer thickened with electrical tape to create the chamber at the end of the cylinder. I also ran fuel tubing up the cylinder to the chamber in order to inflate the cylinder. I then created a cap with a thin piece of flexible plastic and sealed the end of the cylinder. Finally, I started to attach the four air chamber quarters to the central cylinder. As I was attaching them, I quickly realized that four quarters, each with a flat face would yield an awfully boxy tentacle. Instead of continuing with four, I decided to only continue with three of the air chambers. Once they were all attached to the center, I trimmed off the excess ecoflex from the air chambers. Because of their trapezoidal shape there was a lot of excess and removing it made it much easier to inflate the chambers.
Unfortunately, once everything was assembled, I discovered some flaws in my castings while testing. The central cylinder failed to inflate, I suspect due to the ecoflex not adhering well to the fuel tubing and leaking . Also, of the three exterior air chambers, only inflated properly. One other was too thick and requires a lot of air pressure to inflate. The third was very thin at the end closest to the tip of the tentacle. This caused the end of the tentacle to curve dramatically in the top third of the tentacle, but not at all in the rest. It also appeared very fragile since the ecoflex was so thin in that area. In order to try and better distribute the air in each chamber, I wrapped the entire tentacle in ribbon. Unfortunately, the first test with this in place caused the thin section to balloon out between ribbon lengths and pop. So, of the four degrees of freedom I designed into the robot, only one worked as intended.
Future work on this is limited by the difficulties in consistently casting complex pieces and ensuring their are no leaks in any of the seals between pieces. Unfortunately, because of the need to have the thin buffer layer with paper or plastic on the inside of the tentacle, casting the whole thing as one piece seems infeasible, however there may be some additional optimization in reducing how many pieces are cast and minimizing the complexity of the seals.