Comment by ssl-3
7 hours ago
If I understand the question correctly then you're asking about a sprag clutch.
When used in-line on a shaft like this, a sprag clutch permits torque to be transmitted out of the drill motor but not into the drill motor.
Any operation as a support bearing for the spinny-bits is kind of a secondary function at very best (and may be ill-advised; a sprag clutch assembly may be shaped kind of like a roller bearing assembly, but that doesn't mean that it will live long in that kind of role). :)
Thank you.
Would it need a standard bearing on only one side of the sprag clutch, or should I double them up on both sides?
That's a very good question. If I were a mechanical engineer then maybe I'd be able to provide a very good answer, but I'm not. So the best I can do is spitball:
I'd like to think that the sprag clutch would be happiest if it didn't see significant deviation or loading in any dimension other than axial rotation, which suggests that at least one bearing on each side is good. Maybe even two. The constraint needs to be sufficient.
But I also understand that shafts can be overconstrained, and that this can cause non-obvious problems because shafts are seldom ever perfectly straight or perfectly aligned, and the support structure for the shaft is seldom perfectly rigid.
Empirically, this is how we wind up with things like slip yokes, flex discs, and also U joints on the drive shaft for a front-engined RWD car -- even when the differential is bolted in place to subframe members. Those are all things that seek to soften constraint of the shaft. (Some of that stuff tends to remain necessary even on front-engine Corvettes, where the transmission and the differential are also connected with an ideally-rigid torque tube that surrounds the rotating drive shaft.)
It's also how people in the Cartesian 3D printing hackspace often find a reduction in print quality when they see the Z axis lead screws flop around a bit on travel and then endeavor to "fix" that by adding more support bearings, replace flexible motor couplings with rigid, and otherwise stiffen up the spinny-bits.
The combination of the imperfect leadscrews becoming overconstrained imparts stronger forces on the gantry in the X and Y axes than the previous state (doing ~nothing at all) did.
If they can get past the "WTF! More rigid parts must be better!" stage, take these extra parts out, and learn to embrace the slop, then print quality tends to go get better again. The Z screws go back to (mostly!) being just Z screws again instead of also being unintended movers-and-shakers for X and Y.
It's a really interesting topic, but it is also one that I've never taken the time to get my head all the way around so I can't explain it simply. Hopefully I've given you enough words and ideas to let you conduct some useful homework.