Anti-backlash gears

[Ryan Carlyle]

I really like the concept of anti-backlash gears, because they make traditional gear trains much more practical for 3d printing. Backlash is our mortal enemy, so much so that -- aside from extruders -- we almost never use gears at all. We stick with backlash-free belt/pulley profiles like GT2. That's fine, but it's nice to have more tools in the toolbox, particularly since printed herringbone gears are so neat. There are some mechanical designs where rack-and-pinion drives would be useful, for example.

You may be familiar with the traditional anti-backlash or "scissor" gear:

http://www.brighthubengineering.com/machine-design/65169-how-does-an-anti-backlash-gear-works/

What I don't like about these is that any impulse load (such as a corner "jerk") that exceeds the spring force will cause over-travel. They're set up so that the springs have less mechanical advantage than the load on the teeth. Meaning it could add to ringing, unless you use very strong springs. That in turn may contribute to tooth wear. (Which admittedly is less of an issue than normal since the springs also take up the slack caused by tooth wear.) They're also not super common/cheap parts compared to a standard spur gear. 

Another option is variable center distance. If the gears are closer together, there's less backlash. But you don't want to just jam the gears together, because you don't get the proper low-friction rolling contact that involute-profile teeth are supposed to have, and you end up with a lot of wear at the tips and roots of the gear teeth. 

I really like this patented idea (2006):

https://www.google.com/patents/US20070295136

The drive pinion is mounted on a rocker arm that pushes it radially towards the main gear. This spring arm preloads the pinion against two idlers, so that they push on the main gear in opposite directions. You get the same "opposing load paths" concept as the scissor gear, but much more flexibility in using standard parts and arranging the spring to optimize force/travel characteristics. (It surprises me that there doesn't appear to be prior art -- the idea seems too simple to only be a decade old.)

I think it should not be very difficult to get around the patent claims. For example, you could drive one of the "outer" gears and use an idler for the center sprung gear. 

[Ryan Carlyle]

Two places I'm tempted to build something like this:

1) A zero-backlash Wade's style extruder

2) A zero-backlash printed rack and pinion drive (super RepRap style)

I haven't really come up with any other good uses.