# How to let a "screw" leave and reenter a thread

The setup

Consider a cylinder inside another cylinder (placed with their symmetry axis horizontal).

The inner cylinder is about Ø100 mm in diameter and a centimeter or two in lenght/thickness. On the perimeter there is a thread. This cylinder is fixed in rotational motion, and cannot rotate.

The outer cylinder is much longer. It has a thread on the inside through half of its' lenght or so and this thread fits with the inner cylinder thread.

The question

Now, when rotating the outer cylinder, the inner cylinder will move back and forth (the inner cannot rotate but can move translationally).

My issue is that the outer cylinder might rotate many revolutions (it follows a bicycle wheel), and therefor at some point the inner cylinder will move out of the thread of the outer cylinder.

Is there a mechanical and smart method to let such a cylinder leave and reenter the thread of the outer cylinder? E.g. the outer cylinder would rotate in one direction untill the inner cylinder leaves the thread. Then when the outer cylinder is reversed in rotational direction, I want the inner cylinder to reenter the thread and start moving the opposite way.

• Why not block both ends to prevent the internal cylinder from exiting the hole.
– Fred
Commented Feb 25, 2015 at 10:52
• I don't understand where everything is. Make a sketch.
– mart
Commented Feb 25, 2015 at 11:51
• @Fred: To prevent blocking the rotary movement of the outer cylinder in the "outer" direction. I believe this mechanism is meant as a uni-directional clutch with a plenty of backslash: disengage entirely in one direction, engage only after several turns in the reverse direction. This could be done with blocking both ends combined with a standard unidirectional clutch.
– SF.
Commented Feb 25, 2015 at 14:44
• ...and on related note, could the question author elaborate why would so much of backslash be needed? Wouldn't a simple unidirectional clutch be sufficient?
– SF.
Commented Feb 25, 2015 at 15:01

To have the thread automatically reengage some thought will have to go to the exact type of thread used. A coarse thread and maybe a slight taper will help the threads to properly engage while minimizing the chance of cross-threading or binding.

Mechanism

You said that eventually the inner piece will disengage from the outer piece. This doesn't sound like the problem. Your problem is that you then want the pieces to reengage when the rotation is reversed. To do this, the inner part needs to have some force on it to keep it in contact with the outer part. This sounds like a perfect case for using a spring.

The spring would start to compress as the inner piece neared the end of the outer piece. Once the inner piece is free, the spring would provide enough force to press the inner piece back against the outer piece. This way the threads could engage once the rotation is reversed.

A downside to this is that there will be some wear between the threads of each piece. The spring is constantly pushing the two together even as the rotation is keeping the threads from engaging.

• A bit of unthreaded shaft extending past the thread of the cylinder, matching the outer piece's inner diameter would help against lateral misalignment of the two.
– SF.
Commented Feb 26, 2015 at 8:06

The simplest solution would be to use a spring such that it exerts a small force to push the moving cylinder back in. As it 'unscrews' the spring will be tensioned (or compressed) and as the wheel starts spinning the other way, the small force will help the thread to re-enter.