[109jb]

Why u don't buy a ground c5 , c3 ballscrew ?
Is more easy to use it , and u have w little backlash .

- - - Updated - - -

Why u don't buy a ground c5 , c3 ballscrew ?
Is more easy to use it , and u have w little backlash .

Yes a C5 or C3 screw would be great, and they also cost a lot more money. I also already have the ballscrews that we are talking about.

The subject of the thread was about how to determine the size for oversize balls, not which ballscrews should be used.

[kawazuki]

I have heard that some people have had success with repacked nuts containing a combination of sized balls. I'm guessing the larger balls will take up the gap whilst a smaller ball placed alternately will just act as a spacer to reduce binding. When all said and done its all an experiment to achieve whatever is an acceptable lash to your own required tolerances.

[handlewanker]

Hi....just my 2 pennoth worth.......a ball screw can be likened to a ball race that only accepts radial loading....... if you subject a single radial ballrace to any form of axial loading eventually it will display end play .....this is precisely the same as a ball screw in it's construction.

To get no axial movement in any ball bearing you need to have 2 opposed faces spaced to resist the load......even radial ball races can resist axial thrust loads but angular contact types do the job better.

The point is......if you reloaded a ball race with bigger balls to make it work it will by nature of the design cause extra loading on the ball tracks and eventually pitting and failure......this is the same as a ball screw.

The short answer to anti backlash properties is a double ball nut, either dedicated as a design or just two single nuts placed back to back with a resilient spacer.

If the thread is pedantic enough to the point of dismissing this option then bigger balls don't mean backlash solving even if it can be done..........a brand new cheap rolled ball screw has no backlash while the components are still new, unused and pristine.....this is the same as a new radial ball race under the same conditions.

BTW.....if you are getting backlash in a single ball screw......it's mainly in the middle where it does the most work.....bigger balls will be too big on either side of the wear path.
Ian.

[arizonavideo]

Double nuts main advantage is you have twice as many balls so you can preload to a higher level and have less ware. They are called a two point contact system because each nut has pressure in one direction only so only contact on one ball face. Most single nuts are called 4 point systems. Then a few maker have offset ground nuts with the ball tracts offset to make the contact points tow point in a single nut. So a single nut acts like a double nut.

I cant really say single nuts are better in any way but I have re packed over a hundred and most will give really really low backlash often around .001" and they can always be re packed in 5 years or so for cheep. For a hobby mill making a part on the weekends for the local RC club they will last for many years. If you want to run all day long for production go double but then you might not want to have a hobby mill at all at that level.

Repacking, its fun!


https://www.youtube.com/watch?v=8a14YbmUEHE

[109jb]

handlewanker,

Although I see where you are coming from, I have to disagree with much of what was stated.

First, This statement:

.a brand new cheap rolled ball screw has no backlash while the components are still new, unused and pristine.....this is the same as a new radial ball race under the same conditions.

While this may be true for a ball bearing in which the races and balls are closely matched, a ball screw and ball nut can have significant backlash even when new. Here is a direct cut and paste from the Thompson ball screw documentation regarding backlash.

Backlash is the linear independent motion between the ball screw and the ball nut and
can be controlled by preloading the nut. With skip-lead preload the lead is offset within
the ball nut to provide a precise preload. This type of preload is typically used where
both repeatability and high stiffness are needed. Double-nut adjustable preload involves
the use of a compression spring to axially load two ball nuts against each other. It is
typically used for positioning applications where repeatability is critical. Where no
preload is used, axial play is present between the screw and nut, typically 0.002 inch to
0.008 inch depending on the size. No preload is typically used for transport or vertical
applications.

Note the sentence I highlighted in red. So for a non-preloaded screw/nut combination the backlash is typically 0.002" to 0.008" and in my case 0.004" is right in the middle. However, this is not acceptable for the machine I am building because I know it can be better, and the controller I will be using with it does not have backlash compensation.

Next:

The point is......if you reloaded a ball race with bigger balls to make it work it will by nature of the design cause extra loading on the ball tracks and eventually pitting and failure......this is the same as a ball screw.

While I will agree that loading a ball nut with balls so large that they cause binding will result in extra loading, I disagree that this is always the case. A ball nut can be loaded with balls larger than it came with and reduce the backlash while retaining free rotation. As long as the balls are not so large as to create more than 2 contact points, there will be no increased load or wear. For example, I want to reduce the backlash in my ball screws from 0.004" to about 0.001". If there is 0.001" of backlash in the actual screw then there is still enough clearance to allow that backlash and therefore there will only be 2 points of contact. I agree that loading a ball nut with oversize balls will not "eliminate" backlash, but it can definitely reduce it.

Next:

The short answer to anti backlash properties is a double ball nut, either dedicated as a design or just two single nuts placed back to back with a resilient spacer.

I do consider this the best solution for anti-backlash, but this thread was not about anti-backlash, but simply reducing the backlash in the nut/screw combination. There is a distinct difference

Finally:

BTW.....if you are getting backlash in a single ball screw......it's mainly in the middle where it does the most work.....bigger balls will be too big on either side of the wear path.

The ball screws that I was inquiring about when starting this thread are brand new. As such, there is absolutely no wear in them at all. Yet, as stated earlier, they have 0.004" backlash, which maybe not common for new screws is within the manufacturer's specifications.

[109jb]

Here is some additional information about pre-loading and different ways to do it, one of which is by selectively sizing the balls, which is what is trying to be accomplished.

[handlewanker]

Hi 109.......I agree with what you arte saying as it is what you said in the beginning.....to "reduce" the amount of backlash by repacking with slightly bigger balls to better give the .001" you wanted.....also you have a new screw, so no problems with tight spots at the ends etc.......other people will have tight spots at the ends as they will no doubt be repacking to extend a worn screw/nut combo for further use......and that is what I was specifically referring to.

My point that I still hold to (but not applicable to your case) is it would be better.....if it can be achieved in the space you have.........to fit a second ball nut to back against the first one and so give you an infinitely variable and adjustable double ball nut with preload......BUT.......you would need a precisely ground spacer washer to give the preloading that is endemic to a double ball nut, so if a hard rubber spacer is fitted instead of a hard ground one it can give a variable preload to cater for a screw that is not in the best condition but still usable for incremental purposes.

Fitting a resilient spacer will not be a problem as when the load is applied in either direction against one or the other of the nuts, the force that the screw exerts against either nut probably won't compress the rubber if it is hard enough but can still be resilient to hard spot variances.

Anyone fitting a hard spacer between two nuts on a used screw is missing the point completely and will have problems.

BTW....this is a DIY double ball nut solution proposal that can give good results if thought through carefully.
Ian.

[handlewanker]

Here is some additional information about pre-loading and different ways to do it, one of which is by selectively sizing the balls, which is what is trying to be accomplished.

Hi, I would think for most people, having a varied selection of balls would be difficult to achieve, but can be done if you are in the market to repack nuts regularly.

Have you considered that if the ultimate aim is achieved and you attain a .001" backlash figure initially.....it will not remain in that condition and can only get worse........would this be the low end of a tolerance to not accept .002" when the nut/screw ball combination beds down after a period of running and you have to work with the final .002" figure?

My considered opinion would be, from the outset, if the .001" is a critical must have figure, is to design to achieve that with a variable double ball nut combo as it can be adjusted to maintain that figure.

I think a "loose" ball nut with a second "loose" ball nut will work in tandem like an angular contact pair that are adjusted to give a close relationship for high speed running etc., and provided a resilient spacer was not overcome by the pressure of the screw thrust could maintain zero backlash .....forever.
Ian.

[109jb]

The must have for this machine is simply the lowest achievable backlash possible with a single nut. 0.004 isn't the lowest achievable and it also isn't acceptable. All I am trying to do is bring that down. Realistically, 0.002" would probably be fine but if I can get it lower to begin with, why not? For a hobby machine such as this I doubt it would ever wear enough to even notice an increase in backlash.

It was previously mentioned that a double nut was not feasible due to space limitations and the length of a double nut would reduce the amount of travel available. For the Y axis of this machine I will not be using a double nut. If backlash was the only consideration I would simply use a double nut, but backlash is only one of many considerations. Let me state it again. For the y axis of this machine I will not be using a double nut.

I have been extremely busy with work, family, and some other things, so this project hasn't progressed very far, but I will be getting back to it.

In any case, this thread wasn't created to discuss the merits of this method. It was created to find out if ways other than trial and error had been found for selecting the ball size. Alternately, any rules of thumb, like reducing backlash by 0.001" generally requires increasing ball size by 0.0002", would be helpful. That is all I am interested in discovering.

[arizonavideo]

Well I have hand packed over a 100 of 16mm ball nuts and many pack up just fine with under.002" under load and no backlash or almost no backlash with no load. The range of ball sizes is kind of large but the better screws give really nice lash. You can pack them to get what they call 4 point contact which gives a fair amount of friction but they turn smoother with load as you move to 2 point contact under load. A nut packed like this will last a long long time.

As to how to tell the correct size. I have sorted balls by hand to remove balls that were .0001" too large to let the nut run smooth. Balls in a given size are not all the same size. It is a range. When things get tight the larger ones might need to go.

[handlewanker]

Hi Arizona........I liked your video on the repacking.....very clear..........one question, is there any truth that alternate balls are or must be undersized to act as free running spacers to prevent friction from crowding for smoother running?......it seems logical to have it that way.

Two balls running in a track will have their adjacent faces running in opposite directions and therefore will rub against one another and even more so when under a load.........I have to wonder if this is also a case with a needle bearing
Ian.

[109jb]

Hi Arizona........I liked your video on the repacking.....very clear..........one question, is there any truth that alternate balls are or must be undersized to act as free running spacers to prevent friction from crowding for smoother running?......it seems logical to have it that way.

Two balls running in a track will have their adjacent faces running in opposite directions and therefore will rub against one another and even more so when under a load.........I have to wonder if this is also a case with a needle bearing
Ian.

handlewanker,

That is one aspect that I plan on implementing when I eventually get around to repacking the ball screws I have. Since I will have to get larger balls to reduce the backlash, I will have the original smaller balls. I intend to alternate large-small-large-small, etc.I think the concern is the load carrying capacity of the combination when doing this since the load is taken by only 1/2 of the installed balls. One could surmise that the load carrying capacity would also be reduced by 1/2. On my project that isn't a concern because it is a small machine and even 1/2 the load capacity is more than enough. I do think it is a good idea if reducing the load carrying capacity isn't a concern.

[handlewanker]

Hi, thanks for your expertise........it's something that's been in my mind for a while.

One more question......would I be right in thinking that there are two separate ball tracks in each ball nut, hence the two plastic plugs on the outside?

If that is the case.....with 58 balls in the nut each seperate track will have 29 balls so it will not be possible to have alternate big small balls, unless one ball is left out of each row.

I've been working on a design that has a ball nut cut in half so that it can be "remade" as a double ball nut but in the same envelope as a single.........true that it will have only one row of balls to carry the load but the second row also carries some load as it will act as a preload function.

I know that there is a design that has the two tracks in a single ball nut slightly pitched apart to make the balls act as preload items, but the design I'm thinking of is adjustable axially to make a backlash free nut using a standard nut.
Ian.

[Eldon_Joh]

I've been working on a design that has a ball nut cut in half so that it can be "remade" as a double ball nut but in the same envelope as a single.........true that it will have only one row of balls to carry the load but the second row also carries some load as it will act as a preload function.

i wouldn't waste time doing that. it might be very interesting to axially slice the ball nut with a .025" thick cut off blade, and then compress the nut somehow, but that leaves a gap the ball has to jump over. also, its going to distort the ball nut non uniformally.. you may want to try this on a round ball nut rather than a square ball nut. in fact it may be better to grind a taper on the outside of the ball nut, and grind an internal taper in your housing to match and then axially press the ball nut into the tapered housing.

Another option: the single ball nuts that I got from someone, which are these 5/8 X .200 Right Hand Recirculating Lead Screws & Nuts for Power Transmission - Roton Products, Inc.

If you can drill two holes in the right place with a carbide drill bit and pull that off.. and if you make another ball return track, you can get three more rows of balls in the stock single nut.
the stock nut has 4 turns of balls.. about 75 balls are needed.

The screw and nuts I got from someone were not used that much, all of them measured .12425. I dropped .125" balls from mcmaster carr (2$ for 500 of them) in them and they work just fine.


why the double ball nut is 4 times the cost of the single i have no idea. i'm guessing that they have to grind the inside of those ball nuts and the longer the ball nut the exponential amount of effort is needed to make the same inside diameter on one end as the other.

[handlewanker]

Hi...thanks for the info.....I was under the assumption that there might be two separate ball tracks... one at each end...... cutting the nut radially at the centre is not an option if the tracks run all the way down in a continuous spiral and are cut into.

To achieve a simple DIY double ball nut all it takes is two regular nuts flange to flange with a spacer etc....much simpler even if the double length reduces the table travel by 40mm or so.

I think the reason the double ball nut proper is more expensive is because there is a keyway and a spacer to fit so that the pre load can be ascertained precisely........possibly the sales volume for double nuts is not cost effective in the cheaper rolled screw range.

Going to a double ball nut solution to eliminate backlash is half of the battle, the rest is getting the extra precision for the cost.....anyone opting for a double ball nut would automatically be thinking in the long term precision of a ground screw as well...........if the job(s) warrants it.

Purely for hobby use a DIY double nut on a rolled ball screw will give you the infinitely adjustable anti backlash feature for peanuts without the un-needed extra cost for more precision.
Ian.

[arizonavideo]

Double nuts are better and the cost is only about $30 more per screw. The problem is the China double nuts are shimmed with a steel shim so most have lash and if they are tight then will have tight spots because of lead error in the screw and no way for the nut to adjust under load. You have the same problem with single nuts but you don't really have anything you can do with single nuts if you have tight and loose spots.

With double nuts you can use a spring system to allow for lead error and maintain zero lash. My 20mm X axes screw on my IH mill is a double nut Hiwin screw that came with spring preload and still has no lash after 6 years.

I re do the 16mm double nuts with a spring system and around 150 pounds of preload and it works perfect but I have to make 6 parts to do it and each nut still needs to be fitted for spring pressure so it all take time and adds to the cost.

If someone plans on running parts all day for years then this is the way to go. I am going to be doing CNC kits for the IH/Charter Oak and the really nice PM-940M big mills soon and they all will get the spring preloaded double nuts. I have them mostly working now.

I have done the big ball small ball thing and it does help but you run into problems real fast when you get to positive preload with the fit. It really depends on the screw. For a long time many of the RM-16 screws were almost perfect but lately they have not been quite as good. No mater what, if the screw has error in diameter you cant pack them too tight or they nut will just stop turning. BTW they have three tracks.

I did do one trick with the Roton screws and still have the nut and screw to play with. I used a diamond saw to cut a thin slice across half the nut. I then re packed the nut with much larger balls. It was tricky because you had to have the nut flex away from the cut so the balls would not hit a bump where you cut the slice. This worked well and even with the really not good Rotons I got good preload and zero lash with no tight spots. I love the system for a router table where you have light loading and want no lash. I have yet to try to cut the RM16 nuts because with the three tracks you would have to cut into a return or really close to one.

In short after a point it just as well to go to double nuts. The cost is not too bad and after you re do the preload system they are great.

I just finished my PM-25 mill and its in the shop now and I did single nuts with it. Mainly because I will only be doing a few jobs a week with the mill and the lash turned out quite low for it. I still have .002" on the Y. The nut was tight but the Y axes needed the gib set fairly tight so It has extra loading. I will see where the Y is after the axes break in. I would prefer .001"

[handlewanker]

Hi....that is what I was thinking too when I thought the spring loaded double nut design would be better than having a solid spacer.........determining the spring pressure is the hard part.......probably just keep tensioning the springs until all your circles turn out round........an indication that backlash is zero.

As all the accurate stuff is in the final pass, the spring pressure only needs to be sufficient to keep the nuts together without being pushed apart from cutter forces.....a few hundredths of a MM at most......the roughing passes aren't all that fussy.

My opinion is that it would be better to have two nuts that were slightly loose....from new.... because that would not matter when the nuts are pushed together.....or apart....whichever, the effect is the same as pre loading 2 standard radial bearings to make them act like angular contact types.

Using 2 radials with preload for light duty work can be advantageous as they are sealed and contain lube for life which makes the housing a simple design, as opposed to the need to have add on seals for an angular contact design.
Ian.

[arizonavideo]

I use sealed AC bearings, they are hard to find but I found them.

For the 16mm screws they rate them for about 600 pounds each. I don't know the exact force but I increase the pressure until rolling resistance starts to go up to lets say 30 or 40 Oz inches.
One thing you may find is the stock double nuts may or may not have all the same size balls. Stock "loose" nuts can have a wide range of ball sizes and you don't notice but when you add preload if you have a big ball it will stick each time you go around so then you have to either replace them or measure them. When I have a tight spot when rotating the nuts if the spot happens about every 1.125 turns for the full length of the screw its a big ball. If its only in one spot its the screw.

I would not use standard radials or even the sealed double AC bearings that have about .002" of lash. I did use standard radials on my router table but they are preloaded with a rubber washer so have a light load.

[handlewanker]

Hi....that would be a worry....buying a new nut you'd expect it to be pristine and functional......odd balls would be a no no for even basic needs.

If I was a purist, I'd strip a new nut down before it ran and check the balls or repack any new nut with balls selected to give smooth running and a close fit......your video makes it seem so easy....once you know how........the big problem would be to stock sufficient balls in the right sizes to do it......but it would be worth while for the reliability factor.

At any rate I don't know of a supplier in OZ that stocks balls in micron variable sizes.........I did buy 100 balls from VXB in the USA as they were regular 1/8" balls for a project I have in mind for a roller ball screw.....been discussed on the forum at great length a while back.......I already have a metre length of 1605 ball screw to experiment with.
Ian.

[lcvette]

I am wanting to repack my double nuts on a 1605 and 2505 screw, I measured the 1605 nuts and they have .125500" balls, that screw feels a little sloppy to begin with, have not yet checked backlash, as the machine isn't quite finished but would like to order a few sizes of balls and was curious what the consensus was on the larger sizes that might be correct. Doesn't look like an answer was discovered here and wanted to see if anyone had an update?

Thanks!

Chris

Sent from my QTAIR7 using Tapatalk