I was wondering if anyone might be able to offer an account of how a ball screw mechanism works, with a particular quality in mind. Is it actually possible to push the load with the ball mechanism along the screw? I think what I'm describing is happening in this video:
http://www.youtube.com/watch?v=fWyCxByYcFw

If I have that right, then despite the much greater efficiency of the ball screw, doesn't this present a problem compared to a lead screw? What I mean is, won't the inertia of the load cause it to coast a little way after stopping, since the ball screw's providing little resistance? I don't know why, but I think I had always presumed that a ball screw worked the same way as a lead screw, my thinking being that the lead screw wouldn't allow the load to be pushed along the screw's length. But I guess I had that wrong?

Any anecdotes greatly appreciated!

 

Hi Hank,

From what Wikipedia says (take it for what it's worth), yes: you can make them so that you can backdrive them. http://en.wikipedia.org/wiki/Ball_screw

And that's all I know about that. A little armchair mechanical engineering. No personal experience myself, and all that.

What are you working on?


Torben

 

i think you are probably correct about the inertia, but i believe they are probably useing the stepper motor to stop it.

 

If they backdrive, is there a braking system besides that of the holding torque of the stepper? Given that ball screws are used (I'm guessing, but pretty sure) in high-end CNC machines, wouldn't they need some way to ensure precise braking points? If the workpiece is of considerable weight, isn't it possible that the inertia at some point could considerably over power the stepper's holding torque? I know they probably verify distances with encoders, but even then the error would have to be rare, if just to make sure the machine runs fairly smoothly. So ball screws = needs brakes, too?

I'm just running things in my head, thinking about making a walking robot using lead screws. The way I see it, you could make a fairly realistic, albeit slow-moving, walking robot using lead screws to control the motion of the various joints. I'm presuming a kind of robot/walking that would be balanced at any given time during the cycle, so kinetic balance is not an issue (hence speed of the robot is not an issue).

The way the ball screws came up was I was kind of anticipating a more evolved, efficient version further down the line. But if ballscrews don't have the holding resistance of lead screws, then that's a new issue I'd have to factor in farther along, too.

You know what's weird? I have a model I made out of Lego to demonstrate my thoughts on slow-moving, but balanced, walking robots. I guess I'll try to get a video of that on Youtube this weekend. But the thing is, the model I made looks very similar (to my mind, anyway) to the first version of Asimo from back in 1986. That, as far as I'm aware, is pure coincidence.
http://world.honda.com/ASIMO/history/e0.html

In my opinion, my Lego model has a significant improvement over the initial Asimo design. As counter-intuitive as it might seem, I think it's a big mistake to overlook the importance of keeping a considerable portion of the mass of the robot above the legs. This higher mass oscilates in a pendulus motion to counter what would otherwise be a lateral shift in the centre of gravity of the robot as it moves from leg to leg.

Attached Images

hank's walking model.jpg (46.0 KB, 15 views)

 

Usually you have a ramp up and down profile which the software uses to calculate the maximum speed you can transverse on one of these screws before overshoot becomes a problem.

I worked on a big 3 axis machine many years ago which used some very high torque servo motors with 2 meter ballscrews and getting it set up initially was good fun . Once we got the accel/decel profiles correct it ran like a dream.

 

Okay, so my follow-up question is: can you conceivably backdrive a lead screw? I guess my idea of the mechanics of it were that that was impossible, but presuming (and I admit, unrealistically) no friction on the screw, would it backdrive? Or is there some basic element of the mechanics that's different between the ball screw and the lead screw?

Just a hypothetical to try and illustrate my thoughts: at the gym I go to, there's some thread stock suspended from the ceiling that's used to support the ventilation system. Suppose on one of those instead of attaching the ventilation system I hang a large weight on a single nut. Presuming an unrealistically greased-up screw, resulting in nominal friction, should I expect to see the weight slowly revolve around the screw downwards? Now that I'm typing, I expect I would, but I guess in reality what happens is the friction caused by and mass/force on a conventional screw counteracts what would otherwise be backdrive.

Do I have that right?

 

i was taught in machincs school that a nut or a bolt holds its torque on the part of the nut that the washer would touch. thats why they make lock washers. i'm sure you could use our freind pathagerous's therum to culculate exactly how much torque would be exerted on the nut per downward force on it just by useing the angle of the treads. then factor in friction. after all treads are just an inclined plane in a spiral.

 

I think you're right. Thanks!