Hey all. I'm designing a machine for a very specific kind of coil with a special wind. I've settled on a mechanism which depends on a shaft which is free to rotate (only about 10-20 degrees... it needn't ever complete a full turn) but whose linear, coaxial displacement is controlled. Crucially, the shaft only needs to move coaxially a maximum of 5/8" of an inch. The shaft is otherwise fixed in space.

The motion of this rod has some interesting constraints. It will never be under high load. The machine will never make fast movements, so dynamic loading is of very low concern. The motion also needs to be PRECISE but NOT ACCURATE. It merely needs to repeat to the same spot about 200 or 300 times over 20 minutes within a half thousandth of an inch, but it can drift ten times that over hours or ten times the number of cycles. Low backlash is also extremely imporant. I think a strong spring preload would be appropriate due to the low load and the need for repeatability.

Cost is also a concern. I'd like to see how cheap I could pull this off with mostly 3D printed parts. I would love a mechanism which is similar to those used in cheap consumer electronics like VCRs, CD-ROM drives, Printers, and the like, which uses a lot of plastic and spring preloading.

So given all these constraints, I've settled on a few different ideas... one is a cable drive. I place the shaft in plain bearings so it's free to rotate and move coaxially, pull on the shaft coaxially with a cable wrapped around a drum driven by a stepper, and then spring load the shaft coaxially to pull counter to the force exerted by the cable. Since the shaft doesn't need to rotate more than 20 degrees, putting a small twist in the cable is not a big deal.

My second idea again uses two plain bearings but spring preloads in the opposite direction. We then use a non-captive screw stepper motor to push coaxially along/into the shaft. Non-captive screw stepper motors are quite expensive, though... so I really favor the cable drive. I just wonder how repeatable/precise such actuators can be made. Again, I don't care about long term drift.

Thanks for having a look!