... makes (almost) anything
Soft robotics

In the past couple of years I've gained an interest in soft robotics, but I have never made soft robots myself. So this week was a perfect opportunity for me to try some techniques.


The Soft robotics toolkit is developed at the Harvard Biodesign Lab, and is a collection of shared resources to support the design, fabrication, modeling, characterization, and control of soft robotic devices. First, I gave their soft gripper tutorial a try.

The soft gripper inflated at first, but overinflation caused a leak before I was able to take pictures or videos of the inflated state. So for my next test, I went with a bit more precise and sophisticated approach.


The Science Buddies website has a tutorial for a similar soft gripper, along with the necessary stl files for molds.

I printed the stl file they provided, and modeled one myself to serve as a "bath tub" to mold the other side of the gripper.

The step of spreading out the Ecoflex and putting the cast on top of it kept failing on me. It seemed like the Ecoflex flowed away when putting the cast on top, and as a result the air channels were never sealed. So, following the technique that Merav used, I used my two 3D prints to create a bottom and a top part of the gripper, and then glued those together with a thin layer of Ecoflex.

This worked better (at least the fingers inflated), but the bottom layer was uneven (next time I should make sure it cures leveled). Upon inflation, the fingers were uneven. At some point the finger that performed best started to stretch instead of bend. I'm wondering if the bottom layer detached from the top layer to cause that. And.. eventually it popped, so it is not working anymore...

Last try before I give up! I made sure the oven was leveled before I did any curing. I tried both a smaller, tight "bath tub" and the previous larger one to mold the bottom layer and glue the bottom and top layer together. The smaller bathtub resulted in large air bubbles. The larger bathtub resulted in a well sealed soft robot. Still.. inflation is uneven, which can cause overinflation in one finger and result in leaks, but this one seemed more robust than the previous ones I made.

The stl for the top part of the soft gripper (from Science Buddies) and the "bath tub" I modeled for the bottom part.

My final design and final try performed the best. I milled these molds in the Shopbot: mold with air channels and two bases (1mm and 2mm).

Sticking the piece with the air channels together with a 2 mm base worked well: it gave the most even performance among all four fingers. The 1 mm base proved too thin.