Robot wrist joint, timing belt differential  

The ratio between joints is 1.6

It has about 6 solid contact points....taking in the fact that these are plastic gears...I would bet it's pretty strong for its purpose. I'm more worried about the wear of those belts in those extreme angles...

@@b03tz At the end you can see wear on the edge of the belt. And if all 6 points of belt contact mesh properly, it's either a miracle, or very careful design.

@@criznachhow would that be a miracle? If you design a GT2 gear, it would be obvious those teeth all align. Only problem is that the belt goes straight for a tiny bit that might pull on the teeth. The belt wear is a bigger issue I think!

@@b03tz Might? Definitely. Number of teeth on an arc is different than number of teeth on a straight line.

@@criznachfor sure for sure! And in that case I guess if he calculated the rollers so that exactly 1 or 2 teeth are “skipped” it would be quite amazing!

Primitive is exactly what "prototype" means. It is meant to display the concept, not be perfectly reliable

@@DEFCON5307 the concept itself is what's flawed though.

The lifetime of belts will be reduced, but no one knows if that would be significant enough to invalidate the concept

You're primitive

@@gryn1s In commercial joints like in Kuka arms, the belts twist like this, but the length of the twist section is much longer. This creates a smaller ratio between the required outer length on a helix to the required center length of a straight line. Thereby reducing wear on the belts.