[meincer]

I am interested in using this method
http://www.5bears.com/cnc16.htm
to turn my Roton ball screws.

My problem is finding the Delrin. Could I find it locally (Atlanta area) or would I need to wait for an online order?

[rowbare]

http://www.lairdplastics.com/content/section/16/55/

[MichaelHenry]

McMaster-Carr sells Delrin and I believe that they have a warehouse in Atlanta.

www.mcmaster.com

Mike

[scudzuki]

I think the author's obsession with concentricity is a bit excessive, and his methodology flawed. For one thing, what guarantee do you have that the OD of the screw is in fact concentric with the ball grooves (which is in fact the feature that needs to be concentric with the bearing journal to prevent excessive side loading). What guarantee do you have that the ball screw is in fact perfectly straight? Perfect concentricity with the major diameter of the screw at the far end may in fact result in GREATER side loads on the screw when the slide is in the position where it spends most of its life. Seems like way too much trouble to go to when there are so many unknowns.

Joe

[port_huon]

So scudzuki, what method do you recommend?

I have some ballscrews that need turning and I'm interested in the best method to do this.

[scudzuki]

The only part of the methodology in the linked article that I question is the need for a delrin bushing and a 4 jaw chuck indicated within a few ten thousandths to turn down the end. If one has a collet that is in reasonable shape or a good 3 jaw chuck (mine runs within .0015 t.i.r.) that should be sufficient.

Joe

[LeeWay]

I do agree. This is especially true for rolled ball screws. I just turned down one end on some Thompson screws from McMaster Carr. I turned the journal down to 1/2", then 7/16"-20 threads and then 3/8" for the coupling or timing pulley.
I did notice that I could make out the ball grooves and they were a tad wider on one side of the screw.
This may have been the result of using a 3 jaw chuck, but I have it trued up pretty good. It all turned smoothly enough as well with the live center.

With rolled screws, there is likely that much play in the nut itself to account for such minimal error. Also given the fact that you never need to run the nut dead up against the bearing, you should see no issue.

Granted, it does need to be a nice job, but perfect isn't necessary on a rolled screw.

He was turning the lazy end of the screw too. In some applications, a screw won't even require this.

He mentions not to let it get too hot. I annealed the end of mine using a propane torch. I kept a wet rag around what I didn't want annealed, then heated the end to be turned for about 20 minutes. Then let it slowly cool overnight. It turned very nicely the next day. The key is not to quench it when you have it hot.

Here is a thread I started about annealing.
http://www.cnczone.com/forums/showthread.php?t=63112

[dansutula]

I have 3 Techno Isel 25mm rolled ball screws that I sent out to a local shop for machining. I was contemplating the ideal method of gripping the screw root, by using a coil spring of approximately the same wire diameter as the balls, and the same lead as the screw. The spring would be threaded onto the lead screw and and the OD of the spring would be gripped by a chuck or collet. But prior to going to that extreme, we did a few test cuts on one of the screws (mainly to check the hardness and machinablity with carbide cutter - and without any annealing). We gripped the OD directly with a collet and as we turned off the threads approaching the root of the grooves, the grooves appeared (to the naked eye) to be deminishing uniformly. I believe that this a decent indication that the ID/OD were fairly concentric - and close enough for me. The hardness was not a major issue and we were able to get a decent surface finish. Thus, I abandoned the idea of the spring adapter and instructed the shop to simply grip the screws on the ODs. The shop has not competed the machining yet, so I can not yet report on how accurate they "turned out".
Regards,
Dan

[Jason3]

Hi,

For a ground preloaded ballscrew concentricity is very important - it will wear fast and bind up, the more it's 'out' the worse it will be. The lands are ground too, it'd be pretty crappy for a ground screw if they weren't concentric with the threads. The 5bears guy's ballscrew was a ground one.

Rolled ballscrews are different because of the process used to form them. To center a rolled screw based on the threads rather than the flat lands where your chuck will grip it, you can use a dial indicator (assuming the ballnut is already mounted on the shaft). Fix the base and the nut and run the screw through the nut with the DTI point on the part of the screw you will be clamping in the chuck. Measure and note the runout, mark the high spot, then use the DTI to center it the same in the 4 jaw chuck on your lathe.

Best regards,

Jason

[scudzuki]

I believe that a ground ballscrew is likely to have greater concentricity between the thread minor diameter and the screw major diameter than a rolled screw. However (for example), if I am trying to hold a few tenths on a bore, I don't measure the OD and subtract twice the wall thickness... I measure the bore directly. Using balls between the ball thread and a machined tube to indicate the screw in for machining would at least be using the appropriate feature of the screw for the reference point(s). I just wouldn't go to the trouble for the "lazy" end.


Joe

[cadmonkey]

I gripped my rolled Nook screws directly with the 3 jaw on my HF 7x10 lathe. As opposed to annealing I used a pneumatic grinder mounted to the toolpost to grind through the threads with a constant coolant flood to keep the heat down, the sparks down and the grit controlled to a limited area, and got down to the meat which I was able to turn with standard brazed carbide tip tooling. Concentricity was pretty good - I had my brother take one screw into his shop and check it and he was impressed with what I did with what I had - by no means PERFECT but worth the 280 some odd dollars I spent getting the lathe and everything else I've used it for other than the screws.

All 3 screws run very nicely with reasonably heavy preload on the 40 degree AC bearings and varying preload methods on the ballnuts (one with no preload [Z], one with positive preload [X] and one nut with oversize balls [Y]) and the accuracy of the machine (an X3 from Grizzly) is great and there is no skidding or galling. I'd agree on a ground screw I'd probably go to greater lengths to assure that I got a perfect turn (and I don't think I'd consider using my little 7x10, I'd want a beefy lathe - a purchase slated for the future) but with the rolled screws there are a lot of variables as were mentioned above. You'd have to get fairly skewed to induce enough error to impair the functionality of the system.

If you want to see some details of my screw turning there are picture in the gallery on my website - www.distinctperspectives.com

HTH - my .02 - YMMV