Ball Source

[SteveT44]

I've been acquiring parts to build a CNC router. Three used ballscrew/nut assemblies that I've purchesed are a little loose. I've been reading through the posts regarding re-balling ballnuts with larger balls. Is there a good online source for getting the replacement balls? Also, what grade of replacement ball should I be looking for?

TIA

Steve

[HuFlungDung]

Check here. I have bought from these guys myself. 1/8" balls cost about 20 cents each at that time.

http://www.precisionballs.com/

What grade? I buy the round ones I think they are all precision, but what you pay for is the uniformity of the sort size. Without having enough to actually do a trial load of your ballnut, it is tough to guess how much larger you need to go. I bought an assortment in steps of .0001". Cost me quite a bit, though.

[Adobe Machine]

also you may want to try Salem Speciality Ball Co.,@ www.salemball.com/index.htm.We were advised to use chrome steel ,alloy 52100 for our ball nuts.They have every size and increment..chrome steel .125 plus cost us .23 ea, ordered various incremental sizes ( even some metric) above .125, just had to be patient and experiment..our smoothest(spelling ?) and .0005 backlash was .1279.just do not get in a hurry loading !

[SteveT44]

Is it possible to guesstimate the size of replacement balls by measuring the backlash of the ballnut before disassembly? I.e., add the backlash amount to the average diameter of the old balls.

[HuFlungDung]

The ball contact angle is somewhat unknown, by us amateur ball screw reconditioners, anyway

If the ball made contact at its midline (pitch circle of the screw) then I guess the backlash would exactly reflect the amount that the old balls are undersize. However, the ball should be making contact somewhere near a 45 degree angle (I assume), so there will be some trigonometric formula to apply to your initial estimate.

If the race ways are worn a bit 'long', then there is potential for a larger ball to make contact at 90 degrees to the screw axis, before it takes up all the axial lash.

So it is a bit of a guessing game, especially if you are aiming for zero lash.

[Adobe Machine]

I agree w/HuFlungDung...an estimate...but the wear in the ball screw is always uneven:The wear at the ends is usually less than in the "work area" ( middle?)...We start by indicating ACTUAL backlash( load table -x(or-y),turn .100 while indicating,then turn back100. Record error.Do at least 3-4 times, and check other areas, recording each error. If you have a lot of difference in error between ,say the end of travel, middle of travel, 3/4 of travel etc,you may want to re-think trying to repair that ball screw/nut yourself,but if all the errors are the same or nearly so, try a set of balls 1/2 oversize the error ( as as close as the balls on hand) and take the ball nut through the full travel by hand with out the seals ( and boy are those fun to load ). If there is no binding, and you have reduced the error to your specs, load and use with new seals/wipers. The worst thing you can do is get the ball nut so tight it binds in places, as this will cause lots of weird errors and scrap( or mad customers).When your through, you will respect why ball screw repair is so expensive, and frustrating.Caution, inform wife or significant other to keep pets, children,herself and mostly all others away from you untill finished..it is a very frustrating ,time consuming job if done corrrectly.( a little luck does not hurt either.