Large ballscrew

[juzwuz]

Hi! I just bought a large precision ground ballscrew from ebay with a 36mm (1.4inch) diameter, 8mm per revolution, 43 inches of travel, class 2 accuracy. However I didn't really realize how big this thing is until it showed up on my doorstep. My question is what size stepper motor will I need to spin this thing? I have a NEMA 34 around 280 oz-in stepper...will this work? My moving gantry for the router is maybe around 35 lbs. I don't care too much how fast it cuts (as long as it is reasonable). Thanks!

[widgitmaster]

So what your saying is you screwed up!

[DerekZahn]

"screwed up", har har!

I can't say for sure what size stepper you need, but I can make a couple of observations.

1) The diameter, length, or weight of the ballscrew doesn't really matter much in terms of the amount of pushing force your stepper will produce.

2) The pitch of the ballscrew DOES matter. In your case, if you direct-couple a stepping motor to an 8mm pitch ballscrew, the maximum pushing power (in pounds) is roughly Torque-of-stepper (in oz-in) * 1.25. So for your 280 oz-in stepper, that's a maximum pushing force of 350 lbs. It will actually be less because the ballscrew is not 100% efficient (and of course a bunch of that goes into moving the table around, depending on how your machine is constructed). As steppers speed up, they lose torque so the pushing force will decrease as speed increases.

3) A long large diameter ballscrew will have a high inertia. If you spin that baby in your hands and try to stop it by grabbing it you'll feel it. This will limit the acceleration and deceleration you will be able to do. Figuring out exactly how limited is a calculation involving a bunch of letters and symbols that I don't really remember, but a big screw like that will probably impose SOME limit.

That's all theory. Maybe somebody with experience with a similar piece of equipment will be able to give practical advice.

[juzwuz]

Thanks DerekZahn! I plan on driving the ballscrew directly so it seems like if I stay under say a 100 lb gantry, I should be OK in terms of pushing force. How is accel/decel accounted for? Is there something in the cnc software that accounts for it? Does anybody use a large diameter ballscrew (36mm and 43 in of travel) who could give me some insight into their setup?

[NEATman]

If you want to "go through the numbers" you can use the formulas located on this site:
http://www.idcmotion.com/engineers/index.html
From this page, go to the "Rotary System Selection" link.
This page will give you all of the variables and formulas you should need to accurately calculate the torque from the moving mass.

Neatman

[hardyn]

the last large ball screw set-up i did was with a driven nut system, and worked very well...

the screw was clamped rigidly and the nut was belt driven, we got away with 200in/oz servos, but the machine was slow. the motor actully doesn't see much in the way of acceleration as the pitch of the screw that was used was actully quite small, so a self-locking effect occurs. thus the inertia from the decell was transmitted into the screw clamps. and just make sure you run your numbers for acceleration.

i don't get the feeling that you are diving the nut... same thing applies... the motor will not see a whole lots of the braking force, although it is proportional to the pitch of screw being used, the nut mounting device and screw fixing bearings will take the braking force.

whip is the biggest concern with long screw appications.

[juzwuz]

Not exactly sure what whip is. Is it something to do with spinning the screw too fast and it has some sort of unstable motion? Do you have a picture of a driven nut system? This large diameter ballscrew is probably more appropriate for a big cnc mill conversion than a light duty router that I'm planning. I actually just bought another ballscrew with a 0.875 inch diamter that I think will be much better for my application though it didn't come with end bearings. A driven nut system sounds interesting if I can just fix the ends of the ballscrew. Do you need some sort of bearing on the ballnut?

[hardyn]

A driven nut is not an easier system to impliment and it is more expensive but the performace is better.

http://members.shaw.ca/hardyn/photo/profes_perf2/7.jpg

here is a pic of my driven nut system... it used a VERY large screw. the nut was mounted to a spool which was mounted with bearings to the inside of an aluminum housing. again, lots of machining, and lots of expense, but worked very well.

the easiest way to think about whip is to think of a skipping rope. you can put it tight, but it will still sag in the middle. now rotate that skipping rope... as the screw spins faster the inertia of the screw will cause the sagging or whip to increase. with enough speed it honestly looks like a skipping rope.
this kind of effect will vibrate a machine to pieces and cause all the bearings to fail just about instantly.

hope this helps.

[r_fl_z]

Hardyn:

How about some more details on your design?

Thanks,

r

[futura]

Hello Hardyn,

I'm looking at various options for my own build and am intrigued with your design and have some questions.

Is the nut you use a ball nut or a regular nut?

Is the tolerance improved, if yes what is it?

Is a stepper or servo motor used with the design in the picture, is this pulley ratio you use the sweet spot for your motor or is it a rational compromise of sorts?

What type of belt/pulley system produces the most precise positioning?

Is the maximum feed rate of this design improved over a spinning shaft design using the same motor and pulley ratios?

I hope I am not being overly enthusiastic with my list of questions and greatly appreciate any additional inspiration you can share.

Thanks!

Larry

[hardyn]

Accuracy is not improved in this design, what is improved is you are not spinning that big screw therefore the possibilty of whipping is eliminated. it also reduces the inertial requirements of the motor. that was huge screw and would require some serous torque to get it up and running, and slowed back down again. the nut is only a small fraction of the weight of that screw.

yes, it is a ball nut and rolled screw. if you look many manufacturers web-sites you will see that the maximum speed of a sliding nut is quite low. the screws and nuts used were high quality german made components, and were of minimal backlash without using ground screws or "double-nuts"

we were using servo motors, therefore pretty constant torque. The drive ratio was really more arbitrary than anything. The large pully was sized on spacial contraints alone. we used a trapizoidal profile timing belt. cheap and available anywhere. they are kevlar belted and have minimal stretch. there are many belt profiles available, you should do your own homework on that.

by using large pulles, the teeth on the belts will not affect positioning much, so the positioning is affected by pitch of the screw, belt ration, and resolution of the stepper or servo.

maximum speed does increase, as i mentioned above. no whipping, your feed rate is limited by the rating of the ball nut, and inertial constraints; but there is a penalty. Price.

Nook Industries has a good 3d table for calculation screw size based on screw speed, lenth, dia. and is very easy to determine if a driven nut system is right for you... a conventional system is much easier to impliment.

[futura]

Hello Hardyn,

Thanks for the great reply.

It seems like the most critical and time consuming part of the design is the spinning adapter used to connect the ball screw to the bracket.

I can't tell for certain from the photo but it seems that the ball bearing side of the ball screw flange is attached to the adapter and rotates within the large diameter hole in the projecting bracket. Was this design chosen to reduce binding on the shaft?

Would a similar spinning bracket adapter mounted on the opposite side of the ball screw flange with a smaller diameter hole and bearing work as well or would the resulting additional leverage make it less effective?

Thanks again for sharing your knowledge.