[overgrownchild]

Hi all,

I have two 400mm lengths of ballscrews - is it common to join these together somehow to make a single 800mm length?

I was thinking about tapping both ends and joining with a small threaded rod - does anyone have any better suggestions, or should I not bother to begin with? ;-)

Thanks

[ViperTX]

No, it is not common! Can you do it? Well, yes. Will it be without problems......well maybe and maybe not.

Without a perfect mate.....everytime the balls cross the threshold they will be scarred.

If you decide to go forth...let us know how it works out in a couple of years.

Paul

[allesg]

I love the previous answer.

However, I'm wondering how you are going to align the ball path if the joining rod is threaded? I think that forces you to align the thread path and THAT DOESN'T SEEM LIKELY TO ME.

If you used some sort of UBER epoxy and a straight joining rod and you had perfect center alignment and a great magnifying glass to use to visually align the ball path and it all aligns right - MAYBE.

But even then you might see an error as balls skip, dance or jiggle over the join in their path. It really all depends how much of an error your product can tolerate.

If it were me, I'd just buy the 800mm ballscrew and get on with life.

[overgrownchild]

Thanks for the replies! I looked at it again and lined the screws up and they appear perfect without a break in the thread... but yes agreed, joining them perfectly would be a bit hit and miss.

So the massive cnc machines use one continuous leadscrew?

I also had an idea kicking about in my head, which is when you hit the limit of one leadscrew, it starts a separate connecting adjacent leadscrew. Is that done? ... The problem I see with that is the accuracy and calibrating it - but seems possible. Please shoot me down if it is not !! (So that would cost much more than just buying a longer leadscrew, but may have use in some applications?).

[keebler303]

In the real world the most economical thing will be to buy the correct ballscrew for the job. Fiddling with alignment, reliability, accuracy, for what?

Even at a hobby level it is not worth your time to try it.

Matt

[mcarvey]

Not saying it couldn't be done, but you you need to be quite experienced in machining/welding/heat treating, have the right equipment, and be REALLY motivated to do so. Even then, it would probably end up taking less time to make one from scratch.
There are other more straightforward ways of doubling the throw of the ballscrew (pulley, rack and pinion, etc) if you were really in a pinch and absolutely needed to use the 400mm screw. All of these methods are more complicated than getting a 800mm ballscrew.
And there are tons of other uses for that extra ballscrew! Either for your next build (Does anyone on CNC Zone stop after making only one machine?) or another axis, 2 z-axes, etc.)

-Matt

[handlewanker]

Short answer....don't even think about it.

This is like attempting to join two ballrace tracks that have been cut in half across the tracks......the balls go bumpity bumpity bump as they hit the join lines.

What you are attempting to do is make a long screw thread that transfers the pitch of one thread EXACTLY to mate with the pitch of another thread....how you intend to join them in line Gawd alone knows.

Even a soft Vee thread would be challenging.
Ian.

[Andre' B]

Short answer....don't even think about it.

This is like attempting to join two ballrace tracks that have been cut in half across the tracks......the balls go bumpity bumpity bump as they hit the join lines.

What you are attempting to do is make a long screw thread that transfers the pitch of one thread EXACTLY to mate with the pitch of another thread....how you intend to join them in line Gawd alone knows.

Even a soft Vee thread would be challenging.
Ian.

I understand where you are coming from but your example has the disadvantage in that you can buy split ball and roller bearings.
Could not find an example of a ball bearing but here is a split roller bearing.
http://www.bsahome.org/tools/pdfs/Split_Roller_web.pdf

[keebler303]

I understand where you are coming from but your example has the disadvantage in that you can buy split ball and roller bearings.
Could not find an example of a ball bearing but here is a split roller bearing.
http://www.bsahome.org/tools/pdfs/Split_Roller_web.pdf

The two halves are ground flat, put together and then ground as one. They don't just take two, cut them and half and then put them together hoping they work. They are held in the rest of the bearing very accurately so there is no "seam" so to speak, only a tiny gap between the two pieces. No doubt if you put the two ballscrews together somehow and then grind them that way, they will be pretty good. How taking two screws, making it into one and then grinding is better than taking one screw and grinding it I have no idea.

No doubt roller bearings are easier to make in pieces than ballscrews.

It would never be done commercially because its cost prohibitive, its perfectly possible. It is not done on a hobby level because it is outside the realm of an average hobbyist to get it to work properly.

Matt

[handlewanker]

Andre'B, I've worked with the split roller type of bearing in the 50's.....not much call for them except in exceptional circumstances.

The split was at an angle as far as I can recollect, so making the join progressive.

I don't think the ball type would work as the ball would hit the join and destroy the bearing....the same would apply to a joined ball screw.

A typical use for the roller split bearing would be a big end bearing in an engine where multi cylinders and a one piece crankshaft are in use.

For the cost involved, sell them on Ebay and buy a longer one....quicker, less pain, more practical.

The fact that there are a number of balls in the thread form at one time would probably allow the balls to run over the join without bumping on it......maybe....but the allignment would have to almost perfect.

It would work if the screw was rotating fairly slowly, but at higher speeds the impact of the balls on the join would destroy it.

To get a near perfect join would require a spigoted interface, which means grinding and spark eroding the spigot bore.....then alligning the ball tracks perfectly, and if'n the screw was slow moving it wouldn't get wacked by the hardened balls at the join.

So it COULD be done, but what a path to tread without the equipment to make it.
Ian.

[rokag3]

So it COULD be done, but what a path to tread without the equipment to make it.

and even with the equipment !

definitely a bad idea ! unless you are lost in space do not even try it!

[handlewanker]

Hi, provided the drive was fairly slow rotating the impact of the balls at the join would not cause trouble, IE, no fast traversing, AND provided the ball tracks were PERFECTLY alligned axially and radially, but it is in the realm of hypothetical possibilities as opposed to practicalities.

A machine I worked on many years ago had a large ball race (100mm diam bore) on a drive shaft that rotated slowly.

When the bearing and shaft assembly were removed for a service problem it was found to have a cracked inner race, right across the race way.

The crack had not deteriorated to any degree, and the machine just got reassembled and ran for many years more without any trouble.

The fact that the shaft rotated slowly contributed to the ball track not getting damaged by the balls impacting on the crack, but if'n it had been a fast moving drive shaft....a different matter.
Ian.

[Andre' B]

For the cost involved, sell them on Ebay and buy a longer one....quicker, less pain, more practical.

.
.

So it COULD be done, but what a path to tread without the equipment to make it.
Ian.

Yes completely agreed.

The split ball bearings I ran across were large slow turning, a turn table type of setup.
It was a long time ago and someone elses project, I just got a short look at it. The more I think about it the balls may not have crossed over the split, it is possible they recirculated within each segment.

[handlewanker]

Hypothetically speaking, the only way that a suitable solution could be achieved would be to bore/spark erode/grind the end out on one end of a screw and turn/grind down a spigot length on the other screw, a length of approx 1-1/2 X diam, so that the two ends could be joined concentrically and Loctited together as a first consideration.

The next step would be to allign the ball tracks PERFECTLY so that there was no deviation in the pathways when the balls met the join.

The fact that in a linear bearing the load is carried across a number of balls might just make the join crossing smoother, seeing as how the ball at the join cannot deflect when it is in the join area, the same would apply to the spiral groove of a ball screw path.

The overlying factor would be the cost to achieve it.

Unused but second hand ballscrews like these at 400mm long probably could sell on Ebay for $150 each, total $300, whatever.

A new 800mm ball screw (Ebay?) would probably go for $400, so at the start you are only $100 out of pocket.

You don't have to be an economist to know where to draw the line.....LOL.

BTW, the longer the screw, the larger the diam required, or you'll get a lot of wippy springing when you try to push the load as opposed to pullling it.

So an 800mm brand new screw would probably have to be a larger diam, so cost a bit more.
Ian.