Acme Screws for X/Y/Z

[ecnerwal]

Dumb question of the day; I am, on 3rd thought, as I find my newly re-set-up cable drive PR 96 bot offsetting its zero after only very minor exercise, considering a quick and dirty R&P self-upgrade - the cable drive goes 0.75 inches for one full turn of the motor, so a no-gearbox direct swap of a 20 pitch 15 tooth gear or a 16 pitch 12 tooth gear should be exactly equivalent for smallest step size and power. I still think a whole new machine is the eventual upgrade path when serious money comes into it, and I'm not looking for more power or speed, beyond whatever speed I can get by slapping in a higher voltage supply I already have in hand, but have not hooked up.

However, I note while shopping around that, if I was willing to go through the agony of joining two Acme threaded rods to make the X drives (max I can find is 6 ft) the materials for an Acme threaded rod drive are far less expensive than for a rack and pinion drive - in particular the rod is far cheaper than the rack.

Has anyone done this, or does anyone have thoughts on problems with it other than the problem of accurately joining two sections of threaded rod end to end?

[gerald_d]

Joining the rods accurately is one thing, but strongly is another.

The pitch on that acme screw is relatively fine, so you will need to turn it faster than your motor. Speed-increasing gearboxes are a pain (friction, inertia, backlash, price). High speed shafts tend to "whip" - longer is worse - that's one of the reasons they don't make long acme screws.

Then you need to keep the screw clean.....

[ecnerwal]

A bit more shopping around turned up 12 foot long 2 turns per inch (they are 10-pitch 5-start) 1/2" diameter precision Acme screws at a reasonable price from McMaster-Carr. Shipping will probably be less reasonable, but the whole setup would come in quite a bit less than gear racks. They also have a 1 turn per inch option, but not in 12 foot lengths.

Whip (which I'd certainly expect on rod that long and thin) should be controllable with some sliding keeper bearings.

Speed would be cut by 1/3 from present speed, which I could probably stand, possibly a bit more if the inertia of the screw requires more ramping. I might also get a short section and speed up my Z axis (presently 16 pitch non-precision threaded rod).

An offline discussion also lead to the idea of using toothed rubber timing belt as the "rack" - that probably does requre a reduction step, but should be quite robust for dirt, and drops the overall cost even further - there is the question of how much force a particular sized belt can take (resolvable through research), and stretch.