Here I put together 14 rods, 4 of them having stronger springs, because I ran out of the weaker springs, so I think this is probably equivalent to 16 rods. This is the maximum that the Logical Engine should use.
A lot of force was required, and when I tried to drive the bar using the rotating cylinder drive, the bar wouldn't move. It would just bend. I also had to replace some of the casing with aluminum parts, because the force was too great for the acrylic to handle.
Looking at the forces, it is clear that there is a lot of extra energy going into friction due to parts twisting:
So on the drive home, I came up with another idea, which looks like this:
Instead of having a complex bar which pushes the rod stops at an angle, I can push the rods straight out, with the only friction being the bar sliding against the rod stops. All other wasted forces go directly into the pivots where they don't do much harm.
However, I'm still left with the cylinder drive, which has a tendency to push down on the bar. I think I can solve that by cutting a profile into the slope like this:
The bar rests against a flat, not a slope, and the flat is helically wound around the cylinder. This is extremely similar to the point on a screw, or to a fusee, which is a cylinder with varying radius around which is wound a groove for a chain. Fusees were used in clocks to keep the torque going into the clock the same even as the spring winds down, delivering less force.
I might be able to use the lathe at NextFab to cut one:
Due to the geometry, the diameter of the large end needs to be a little over 2.3 inches, and I only have a 1.625 inch steel round. So it's not likely that I will get the bigger round by this weekend. I will only be practicing fusee cutting this weekend, but hopefully next weekend I'll get to try the improved drive system.