How to model a ballscrew and nut and simulate motion

[CNCnUtZ]

I was wondering if someone could tell me the proper steps to modelling a screw held in bearings and a nut attached to a table and then simulate motion of the table via rotary simulation in SW2003?

If someone could send me a sample file that would be great too.

Thanks

[jimmyb]

Have you had any responses yet? I started working on it, but man its exhausting. I've modeled a primitive ballscrew and a base it will sit in. Thats as far as I've gotten at this point, lol.

[CNCnUtZ]

Originally posted by jimmyb
Have you had any responses yet? I started working on it, but man its exhausting. I've modeled a primitive ballscrew and a base it will sit in. Thats as far as I've gotten at this point, lol.

Im sorry.havent been here in a few.
But I have gotten no responses.

But I did make some progress...
I modelled the ballscrew just fine but if I use the standard size threads on the ballnut to match the ballscrew, it mates fine but has collision like crazy when moving and simulating..... and no matter what I do , I cant seem to get rid of the collisions.

So what I did was make the sweep of the profile for the threads on the ballnut, much smaller, and just made the thread sweep on the screw, tangent with the face on the end of the sweep(helix)for the nut and it simulates fine...just takes a longggggg time to process.
Its strange that the profile swept at a larger diam. will not mate correctly with the screw cut sweep using the same helix ,plane ect. Mind boggling.

Ive nearly completed modelling the mill...just have the z column to dimension just right and mate everything.

This is probably the largest assembly that Ive ever done...approx 142 mb.... the ballscrew sub assemblies on all three axiis, taking up nearly 70% of that!

I would have liked to make the ballnut as authentically modelled as possible. But I can settle for the threads of the ballnut not being to spec considering they will be hidden anyway.
I have been to a few forums and some advice Ive gotten from a few "so called" proffesionals with SW, that it couldnt be done.
I figured ,there had to be a way!

I can send you an assembly of the ballscrew if this is at all unclear.
Let me know and thanks for the reply..

[cliffm69]

Send me the file for the ballscrew assembly and I will take a crack at it.

Cliff

[hardmill]

I'd also like to see the file.
e-mail it to me.

[j45acp]

I too would like to see the file. Could you e-mail it to me also?

j45acp@hotmail.com

[HuFlungDung]

CncNutz,

That sounds like an interesting project. I would point out that in a typical ballscrew, the threads do not engage one another. Therfore they cannot collide. Without the balls, the screw would simply spin inside the nut. I'm sorry if I am stating the obvious

The balls do not run on their "tops and bottoms" either. Typically a four point contact is assumed, I believe, with the balls contacting maybe at 20 to 30 degrees each side of a normal to the screw axis. This would mean that the form of the grooves is not a perfect hemisphere, because then it would be contacting the ball on every possible chordal circumference, and this is a real life collision: the ball cannot simultaneously roll at one circumferential velocity when it has many different chordal circumferences in contact.

This same dilemma occurs whenever someone builds a V shaped wheel rolling on the corner of an angle iron. The whole V cannot contact the whole surface of the angle without causing inherent skidding and wear to either the track or the roller. The track passes beneath the wheel at a uniform linear speed, but each different diameter of the V in the wheel has a unique angular velocity. Something has to give

The points of contact of the ball must be at exactly the same chordal circumference of the ball, both on the screw and in the nut, otherwise, the ball would have to "skid" on one surface or the other, because the velocities would be different. This would be due to the chordal circumference velocity at one contact point at a given angle on the ball, would be different for every other contact point that did not lie on the same chordal circle.

I hope that this might shed a bit of light on the problem. If I were going to attempt to model something like this, I would not use a hemispherically shaped groove, because in real life, that is not going to make a good screw because of skidding of the balls.

[vacpress]

never tried anything like this, but i think:
To accurately "simulate" a ballscrew in solidworks would be quite a project. after getting "EXACTLY" the right 3D model, you then have to use some sort of FEA package to do a basic analysis of the assembly. then animate the thing, and let the FEA do it swork also. So you get dynamic FEA data about the movement.

[cadman]

Sounds like an interesting homework assignment. I'll squeeze this in between projects and post the results when I finish.

[Zipdrive]

I think you are missing a piece of software. MSC Dynamic Modeler.

Trying to manage a million configurations and use the animation toolbar in solidworks is a horrible.

[ESjaavik]

Seeing how long it takes just to redraw a V-thread in SW, I'm pretty sure your "virtual mill" will not be good for HSM.

I'm following this thread closely. It will be interesting when you
start hooking it up to your CNC program.

[ynneb]

This same dilemma occurs whenever someone builds a V shaped wheel rolling on the corner of an angle iron. The whole V cannot contact the whole surface of the angle without causing inherent skidding and wear to either the track or the roller.

I know what you mean. Its just not smart engineering. However it works. Its a bit like a brake shoe that stops a car. Its not the best option but still it works. Or an automatic gearbox and its perpetual slipping. This kind of engineering annoys me in some ways. But it works. I have learnt that nothing will ever be done, if the 100% best way is waited for. Sorry but another example comes to mind. A propeller, Imagine before airplanes were invented and some one proposed to push against air itself to move an object. I would be first to say, its too inefficient and never give it a second thought. The crazy thing is, push enough air in one direction, and you get propulsion.

I think pioneering engineering thinks outside the box. I must say that I am generaly an in the box person though. This is what limits me.

Sorry to not be on the subject of this thread at all.