Quote Originally Posted by Tenson View Post
Perhaps slightly off-topic but I was wondering if half the problem with unsupported rails comes not from the rail itself, but from the commonly used clamping method to hold it?

I can imagine those rather immaterial aluminium screw clamps allow a fair bit of movement and flex under use. On the other hand, if the rails precisely slot into a much more solid piece of metal and are directly threaded and screwed up tight, I bet the movement is reduced a lot. This is how the CNC3040 and 6040 do it.

I'm guessing most who have used unsupported rail in DIY machines have gone the clamping route, because the cost of end-machining the rail to have a thread and perfect flat end makes it easier to buy supported rails.

Supported rail will be better, of course, but I'm just wondering where the main area of weakness might be.
The main weakness comes from the fact the beam is not supported along its entire length.

Think about a springy length of piano wire sitting on a support at each end:
Imagine pushing down with your finger in the middle of the wire.
Now clamp each end of the wire firmly to each end support and assume the supports cannot move. Imagine once again pushing down with your finger in the middle of the wire.
Finally put a third support in middle of the wire and imagine pushing down directly over the middle support...

You may like http://www.clag.org.uk/beam.html , see 'Centre load on beam with two simple supports' and 'Centre load on beam with 2 fixed supports'