I should soon be receiving a batch of test prints for the artificial muscle designs. A big thanks to Susan from, who was able and willing to print out so many parts. These parts have been printed in PLA and PVA. The PVA parts are designed to dissolve inside the muscles and leave them with a more complex internal structure, than otherwise possible. I am also testing a number of flat muscles which will have parallel or converging fibres, as well as single or multichamber designs.

I will need to build a form of Universal testing machine to test these designs out properly.


I’ve already got ideas for what next to try, and I’ve decided to focus on the face and head muscles for now. These are smaller, but just as complex. I’ve gotten a skull model now that I can start taking measurements on and so it shouldn’t take long to start making designs, once testing is done.

Plus it might open up options for animatronic design.

I just posted a new video on youtube. It’s a prototype of some of the elements I was talking about in the previous post. Here it is below.

So this test was actually from 6 months ago. As you can see the robot was hardly exactly finished, which is one reason I hadn’t posted it yet. But I’ve realised that I’m likely not going to have a series of ‘finished’ robots before I design the first real prototype anyway, so I might as well post the half (or quarter) finished stuff.

The principle behind this design is that the human body builds joints from hard and flexible materials. In a typical robot, joints are held in place due to rigid material constraints. E.g. in a ball and socket joint the socket needs to encapsulate over 50% of the ball surface in order to hold the ball in place. The sockets in human ball and socket joints barely encapsulate a quarter of the ball. So how do they stay in place? Well the answer is ligaments and the muscle/tendons that anchor it. The shoulder joint in this case supplies a force into the socket cup. It’s enough force to hold it under small forces like gravity, but low enough that there’s not much resistance to it moving.

The spine is slightly different. It uses the principle of ‘tensegrity’, where elements under tension (typically flexible cables) hold rigid components in place. In this case the under tension elements are dragon skin 10 silicone.

The idea moving forward will be to replace the passive silicone pieces for function Pneumatic muscles. This is the next set of testing and once I’ve worked out the characteristics of the muscles I can start working on the practical side of constructing the full size robot.