It is 49 stand 1mm dia stainless steel fishing wire. Dyneema is hard to get.

It has been suggested to add a stiff spring in series with the wire. The problem is the creep/compression in the 3D printed parts.

Do you have any links to you work?

That's clever to use steel fishing wire! I did not know that they use wires with such a high number of strands. Ok, here in Switzerland, Dyneema ropes are really easy to get in stores for sailing ships. You may want to have a look at these two articles (use scihub.ee if you have not access yourself): 

- "Guiding effects and friction modeling for tendon driven systems", doi: 10.1109/ICRA.2014.6907852 

- "Cable-Driven Actuation for Highly Dynamic Robotic Systems", doi: 10.1109/IROS.2018.8593569

I am sorry, I cannot link my work yet because I just started to write my first manuscript for publishing. 

Regarding the creep of 3D printed parts: What I did so far was to clamp the cables in a mini-slider that can be tensioned by a simple screw. This makes it very easy to adjust the tension. But indeed, stiff springs might be a solution. In our lab, we are working with some devices of Force Dimension and they use springs as well. It is, however, a pain to change the cables.

I am already looking forward to seeing the next update on your project. 

Great design! What kind of cables are you using and where did you get them? At my university, we are working with capstan drives as well (in our case it is for haptic devices) and I have made the experience that it is hard to get good cables in low quantities. Dyneema ropes might be a viable alternative that we are probably going to test this spring. Depending on the type, they do creep very little compared to other ropes. Did you already consider or try using ropes?