A different (better) kind of screw

[handlewanker]

Hi Dan, there's been a number of design proposals using ball races and screws in combination.....most have failed to materialise as a viable project due to the application of hard metals on soft metals that destroys the screw , namely an Acme screw thread, which is the most commonly used cheap screw drive, (as opposed to the ball screws) in routers and all manual mills......even expensive mills like the Bridgeport.

As soon as I get some tangible results I'll post an idea that might be another different approach to the ball screw alternative.
Ian.

[daniellyall]

if you Google ball races and look at picks there's a few different ways of doing it most are ideas a couple of company's sell different type of the same idea they cost a lot

[handlewanker]

Hi Dan....I didn't mean actual ball races......I meant a leadscrew alternative to a ball screw with ball races as the drive component or nut.

There's been lots of designs for possible alternatives to a ball screw using ball races......one has the ball races on the screw at an angle with the corners of the inner race in the thread form.....short life and not good at transmitting loads.....I'll post a pic of the one I mean.

It's a clever design....could work well under certain circumstances.....certainly not heavy duty.

Come to think of it.......if this design was to be used with a ball screw it might work very well as the ball screw itself is hardened and the thread profile might make a better contact with the ball race inner race than an Acme thread.

The only problem with that design is .......the ball races are designed to cater for radial loads, whereas screw transmission is all axial.
Ian.

[daniellyall]

yer I new what you meant.

you got to use the correct wording when searching I saw one some one made with 3 bearings he had a vid of it working on a rod he said wonder how well it work with a ball screw

[handlewanker]

Hi Dan.......yes, very clever design, looks like a commercial job.......problem is it has slip as it relies on the friction between the shaft OD and the hard outer race periphery of the ball race to do the moving......but...

If the shaft was a ball screw........ like you say...... possibly if the ball race was angle over more, the ball race OD edge could get a positive grip on the groove of the thread profile and drive even if it did have some backlash.....what cheap rolled ball screw doesn't.

Some spring loading might "persuade" it to stay tight in the thread form and be self adjusting.....for most average hobby type uses......like as a replacement on those aluminium CNC routers on EBAY that have Acme thread screws and Delrin or brass nuts, just crying out to have ball screw retrofits.

Even then, one edge of the ball race casing has a rolling contact and the other edge is rubbing against the thread form.....that's lots of friction going backwards.

In the pic you posted the ball race rides on the OD of the shaft, probably a hardened shaft, and so is a pure rolling contact........for Louie....almost!

With that design you would not have to worry too much about swarf getting into the works as most of it will be pushed off as the ball races rotated.

Even on a plain steel shaft...... with a corner contact, eventually it will dig into the surface and wear a track and so you have a self generating screw profile.....pitch unknown........encoders will help here......perhaps giving it a nudge in the right direction by cutting a shallow pitched groove in the lathe.

My main and only interest is to have a screw alternative that had pitch/lead accuracy and no backlash,
Ian.

[daniellyall]

its diy that thing, the guy said it ate in to the shaft that way would not work to good for ball screws if they where in line with the pitch with the correct off set they may be ok its another idea I have been look at since I said about it pages back it would be easy to do a test I have some bearing I might do a test pieces next week or week after.

I have to make a laptop table for a wheel chair next week so it will have to be after that, it will only be a days farting around to test it I might even put a radius on the bearings I got a ball turning jig I can mount the grinder to

- - - Updated - - -

another question hows the router going

[handlewanker]

its diy that thing, the guy said it ate in to the shaft that way would not work to good for ball screws if they where in line with the pitch with the correct off set they may be ok its another idea I have been look at since I said about it pages back it would be easy to do a test I have some bearing I might do a test pieces next week or week after.

I have to make a laptop table for a wheel chair next week so it will have to be after that, it will only be a days farting around to test it I might even put a radius on the bearings I got a ball turning jig I can mount the grinder to

- - - Updated - - -

another question hows the router going

Hi, the router is on the back burner slightly while I assemble another project also for welding.......gonna set up and do the two together......welding is a business you want to do carefully or you can burn the place down......LOL.

That device you posted might prove to be difficult to make especially with the off sets to the ball races and the angle they need to be to the screw helix.....difficult to adjust .......the way they are in the pic is on a plain shaft and they drive by the amount they grip the shaft by.

Being on a plain shaft means they aren't too fussy about how they are positioned to the shaft........different if you want to run them in a groove which is very pitch orientated.
Ian.

[louieatienza]

yer I new what you meant.

you got to use the correct wording when searching I saw one some one made with 3 bearings he had a vid of it working on a rod he said wonder how well it work with a ball screw

Good pic Daniel. The concept was developed by Roh'Lix. I brought up earlier the concept of making sleeves foe the bearings to fit in the threads. Uhing also has a similar concept, instead using the inner race. There is someone that implemented that concept to a screw a while back in this thread.

[louieatienza]

Hi Dan.......yes, very clever design, looks like a commercial job.......problem is it has slip as it relies on the friction between the shaft OD and the hard outer race periphery of the ball race to do the moving......but...

If the shaft was a ball screw........ like you say...... possibly if the ball race was angle over more, the ball race OD edge could get a positive grip on the groove of the thread profile and drive even if it did have some backlash.....what cheap rolled ball screw doesn't.

Some spring loading might "persuade" it to stay tight in the thread form and be self adjusting.....for most average hobby type uses......like as a replacement on those aluminium CNC routers on EBAY that have Acme thread screws and Delrin or brass nuts, just crying out to have ball screw retrofits.

Even then, one edge of the ball race casing has a rolling contact and the other edge is rubbing against the thread form.....that's lots of friction going backwards.

In the pic you posted the ball race rides on the OD of the shaft, probably a hardened shaft, and so is a pure rolling contact........for Louie....almost!

With that design you would not have to worry too much about swarf getting into the works as most of it will be pushed off as the ball races rotated.

Even on a plain steel shaft...... with a corner contact, eventually it will dig into the surface and wear a track and so you have a self generating screw profile.....pitch unknown........encoders will help here......perhaps giving it a nudge in the right direction by cutting a shallow pitched groove in the lathe.

My main and only interest is to have a screw alternative that had pitch/lead accuracy and no backlash,
Ian.

This design is designed to slip when a force exceeds the rating of the unit. This eliminates the need for overload protection or even switches for some mechanized operations; stops at either end instead of limit switches.

Again, as this relies on rolling friction the lead can vary depending on the load thus you need linear feedback for positioning.

[louieatienza]

Hi, that was just a comparison....no one really makes a ball race when you can buy them so cheaply.
Ian.

Exactly. Same with a ballscrew. HELLO???

[louieatienza]

Hi all......had a deep think about your design Louie..........nope, it is watch making at the best.....not in the making, but in the setting up.

First, I quite like the clever profile design of the rollers and how one interacts with the other.......the essence of good design is KISS........a point most designers forget when the bit is firmly between the teeth.

As far as the rollers being simple to make.....they are just profiles and can easily be made with a CNC or manual lathe, even a form tool to make the grooves is not a big deal........BUT.......you would not want to go and plunge a form tool into a spindle of that small diam and expect to have it accurate.

If you managed to produce the rollers to the form you designed, heat treated them and finished etc.....the last point in the saga is.........how do you adjust them for contact, alignment, and the biggest bogey thrust on 6 tiny rollers?

Plain simple radial bearings will soon not be radial any more and exhibit axial play as anyone who has used radial bearings in a thrust configuration finds out pretty quickly........Chinese designers are fond of this "cost saving".

If you use small thrust bearings as well as the radials on each end, the complication now goes deeper.

Having 3 sets of rollers and balls means they all have to be precisely aligned to one another.....a balancing act of gigantic proportions in a tiny envelope.

If just one roller in any one set is slack or too tight, the whole balancing act trembles.

If the rollers are carried on radial bearings.....what provision do you intend..... to adjust for positioning each set to the screw profile .....in two directions....and for thrust too?

The devil is indeed in the detail in this case.

Lastly, how do you intend to assemble the nut to the screw with all those small parts precariously balanced against each other.......a dab of grease would hold the balls in while you screw the screw into the nut, but subsequent servicing would be a nightmare.

BTW, contrary to what you have said in post #1014, the design I proposed is totally about a 3 point contact configuration......there being two sets of rollers and a third plain roller on top for adjustment, all acting together to maintain equilibrium.

There ARE axial forces to contend with on any router. mill or machine that pushes or pulls to achieve position by spiral grooved screws, and pushing or pulling has to be contained or you're back to square one with the bogey of uncontrolled backlash.

It's a nice try Louie, but you have to be able to make it or get it made, even just a mundane prototype, if you want to prove the pudding.....so if that is achievable then maybe NSK will make you an offer for the design.......but as it's public domain now, that won't happen......the Chinese are already streets ahead now they know how it works....LOL.......and they're experts at adopting ideas.

Meanwhile, as I'm satisfied the design I proposed is as far as I can develop it and doable, I'm working on another design, but this time it's a ball race screw.....totally different to the roller ball nut design, and possibly extremely DIY achievable.
Ian..

To address your concerns:

Many model engine builders make parts even smaller, this should be feasible even on a CNC Sherline lathe. You can even make a mount for a pencil grinder if you wish to grind the profile.

As to adjustment, I envisage something similar to ballscrew mounts. One end fixed, and one end free. The design is self aligning afterwards.

For thrust, radial bearings can carry thrust for light duty applications. There was no prerequisite to use this ro lift the head on an RF-45. Plus, for heavier load applications you just use a bigger screw and parts, and AC bearings become feasible.

There should be no need for alignment, that's what the retainers do. And with a CNC that's not an issue.

Assembly is also easy. Insert balls in location with grease, put a rod inside with same root diameter as the screw, thread ballscrew in, pushing rod out.

As to making it? I bet you a beer I havr a prototype done before you finish your router hahahaha....

To your point on 3-point contact... That's all fine and dandy in the 2D world, but as your design shows the races and grooves envelop the ball, thus line contact not point contact.

[handlewanker]

Practical 3 point location....that is what I call it and that is what it is.

The grooved rollers and the balls are held in 3 point suspension.

Being able to balance the 3 points against one another you get a triangle of forces all pushing in to a common centre point.....the three legged stool effect, or as used in a lathe fixed steady when you machine long shafts......it's the most stable configuration there is.

With this last design I only have one thrust bearing on each grooved roller....this enables me to have one adjustment for all the grooved rollers initially and one adjustment overall for the total pressure of the balls to the screw.

If you do make one to your design before I make one to my design I'll eat my hat.....the router is a work in progress project.......indulged in due to Defeng being late to the party, and that'll be as and when it happens.

At the moment I have a set of ball races 5mm X 4mm X 3mm, and a set of ball thrust bearings on the way, F2.5-6M type, which is 2.5mm X 6mm X 3mm......10 for $16 on EBAY, where else......... so I'm already half way there......LOL, and I already have a 1605 ball screw with a packet of 50 - 3.17mm ball bearings.

The burning question is.....why would anyone make a substitute for a ball screw when they are so cheap on EBAY?

Well, some of us like to think they can have something that will last forever.....forever interpreted as when you stop using it and it gets shelved as a display piece.

Many years ago in the early 60's I built a small lathe to do model engineering work, and almost 50 years later I still use it for small jobs.....at one time I used it for a business venture lasting three years.....it's a personal thing when you make something to use......I also made a 4 jaw chuck and a quick change tool post for it....that's what people do when they get personal about things.
Ian.

[louieatienza]

Practical 3 point location....that is what I call it and that is what it is.

The grooved rollers and the balls are held in 3 point suspension.

Being able to balance the 3 points against one another you get a triangle of forces all pushing in to a common centre point.....the three legged stool effect, or as used in a lathe fixed steady when you machine long shafts......it's the most stable configuration there is.

With this last design I only have one thrust bearing on each grooved roller....this enables me to have one adjustment for all the grooved rollers initially and one adjustment overall for the total pressure of the balls to the screw.

If you do make one to your design before I make one to my design I'll eat my hat.....the router is a work in progress project.......indulged in due to Defeng being late to the party, and that'll be as and when it happens.

At the moment I have a set of ball races 5mm X 4mm X 3mm, and a set of ball thrust bearings on the way, F2.5-6M type, which is 2.5mm X 6mm X 3mm......10 for $16 on EBAY, where else......... so I'm already half way there......LOL, and I already have a 1605 ball screw with a packet of 50 - 3.17mm ball bearings.

The burning question is.....why would anyone make a substitute for a ball screw when they are so cheap on EBAY?



Well, some of us like to think they can have something that will last forever.....forever interpreted as when you stop using it and it gets shelved as a display piece.

Many years ago in the early 60's I built a small lathe to do model engineering work, and almost 50 years later I still use it for small jobs.....at one time I used it for a business venture lasting three years.....it's a personal thing when you make something to use......I also made a 4 jaw chuck and a quick change tool post for it....that's what people do when they get personal about things.
Ian.

Well you'll have some bearings and a ballscrew spindle to add to that cut square tube on the shelf! You might even get your SVM-0 before you even begin to finish your router!

I built a CNC router almost 6 years ago, and while the frame had evolved over that time, the basic concept, and even the linear rails and ACME screw and Delrin AB nuts, are the same I've had on the machine the past 5. No big deal - everyone and their mother seems to be building them now. But I didn't build it to impress or humor myself. I built it to help me with my guitarbuilding, and it has also branched out to sign making and car parts. Hence the need for a mill. It would be nice to have the opportunity to build something just for the sake of it. Or even to prove a point. But time is finite, and time is money...

[handlewanker]

I wish....the SVM-0 is my wish list top priority.

I quite agree that you can work wonders with anything you have as you're more familiar with it's quirks and can accommodate it with cunning expertise........practically all if not absolutely all manual mills have Acme threads or square threads.

I have a Beradi jig borer, Italian make, that I bought in the late 70's and that has all square threads with compensating back nuts to take up the slack in the nuts.

I think that if you intend to walk the CNC walk with a self build, you really do need to consider having ball screws (and linear rails) right from the very beginning........retrofitting a manual is not a wish list item I'd like to contemplate.

I've looked at the latest design of the roller ball screw I proposed and I think when the time comes for the router drives, I will very strongly lean on the side of making the nuts to this design......because I want to.....just got to acquire some used ball screws in the lengths I need.....or brand newies if they sell the screws by themselves......you can buy the nuts by themselves.....perhaps buy the complete package in case my leap of faith ends up down a deep hole and I have to go to conventional ball screw drives....LOL.
Ian.

[louieatienza]

I wish....the SVM-0 is my wish list top priority.

I quite agree that you can work wonders with anything you have as you're more familiar with it's quirks and can accommodate it with cunning expertise........practically all if not absolutely all manual mills have Acme threads or square threads.

I have a Beradi jig borer, Italian make, that I bought in the late 70's and that has all square threads with compensating back nuts to take up the slack in the nuts.

I think that if you intend to walk the CNC walk with a self build, you really do need to consider having ball screws (and linear rails) right from the very beginning........retrofitting a manual is not a wish list item I'd like to contemplate.

I've looked at the latest design of the roller ball screw I proposed and I think when the time comes for the router drives, I will very strongly lean on the side of making the nuts to this design......because I want to.....just got to acquire some used ball screws in the lengths I need.....or brand newies if they sell the screws by themselves......you can buy the nuts by themselves.....perhaps buy the complete package in case my leap of faith ends up down a deep hole and I have to go to conventional ball screw drives....LOL.
Ian.

Before stumbling upon Defeng's thread my plan was to buy a mill, mill off the dovetails as much I can, and retrofit linear rails. Second plan was to purchase cast iron plate and machine the parts. I don't think you'll regret for a second just purchasing ball screws - even ground ball screws can be had for not much if you're patient enough. Get your machine running, then you'll have a better grasp how the ballscrews work "in practice" and how they can be improved. Though I think the current commercial design is as pure as can be, in it's simplest form.

[daniellyall]

to make something like what I posted is easier than you think there is only the angle of the pitch to get wright and the angle the bearings will need to be on they are the same all the way around that's what 3D cad is good for even if I just made a roller to sit in the grove what I would support with bearings I would also do two set back to back it would not be small

[louieatienza]

to make something like what I posted is easier than you think there is only the angle of the pitch to get wright and the angle the bearings will need to be on they are the same all the way around that's what 3D cad is good for even if I just made a roller to sit in the grove what I would support with bearings I would also do two set back to back it would not be small

I think this is very feasible and obviously simple design. You can use deep groove ball bearing...

[handlewanker]

Hi Louie......BTDT.......I have a small jig borer that I intended to retrofit, almost from the ground up.....but other priorities got in the way.

Making things is a main interest with me, so the ball nut alternative is one thing I intend to make....eventually.

I thought the early designs with the rollers were going to fly, but as the project developed it became obvious that the layout of wheels in a groove for an existing thread form wasn't going to work as all progress on all fronts slowly fizzled out......it's that dammed helix angle that blocks progress.

It also became obvious that commercial roller nuts were extremely exotic, and no way were they going to be reproduced by DIY means.......hence the deep thinking that produced the roller ball nut design.

In my opinion, there is nothing exotic about a roller ball nut that it cannot be made by DIY means.

You mentioned that there were an insane amount of rollers in the design, well I went back and had another look and came up with yet another design change to offset this.....sketch attached.

Basically, it consists of a split ring that loosely wraps round the grooved roller, in the groove, and ends at the ball position where the ends of the split are positioned either side of the ball.

The ring segments (6 of them, one in each groove) are attached to a cross bar behind the grooved roller that serves to secure it in position in the roller grooves.

We could then do away with the 2 disc rollers that formerly ensured the balls stayed in position, that you got excited about.
Ian.

[louieatienza]

The reason the DIY roller screws failed is that they really did not mimic the real deal. First is the clocking of the rollers by means of ring gear, and spurs cut in the ends of the rollers. Second, the grooves of the housing the negative of the rollers, transferring load thus no need for bearings. Third, with planetary roller screws, there are multiple starts, one per roller. Otherwise it would be impossible to thread the screw since all rollers are alike. In a differential roller screw, the rollers are threaded, as well as grooved to match the profile of the inner housing; and different nuts can be used with the same screw even left hand thread. The recirculating roller screw also has a ring gear, but a mechanism that allows the roller to hop back to the start of the inner thread; a single start screw is possible. The last point is that the threads are very fine, which allows lower helix angles while still capable of higher leads.

Compare to the DIY stuff...
- less rollers, and different to follow helix, and related to...
- single start screw with coarse threads give high helix angle
- thrust transferred to retainer plates, relying on bearings or bushings instead of transferring to housing
- rollers of softer material
- reliance on eccentric cam or other mechanical preloading instead of use of tight tolerances

I think today's hobbyists have much more sophisticated machinery and CAD/CAM than even 8 years ago when this thread started. Some folks even have VMCs in their home garages. So I think it's possible for someone to make a roller screw today. The problem is, if you have the means to make one, you probably don't need one anyway.

[daniellyall]

[/QUOTE]
I think today's hobbyists have much more sophisticated machinery and CAD/CAM than even 8 years ago when this thread started. Some folks even have VMCs in their home garages. So I think it's possible for someone to make a roller screw today. problem is, if you have the means to make one, you probably don't need one anyway.[/QUOTE]

that's very true.

but for the mad men out there who like to make there own toys they can end up making there own VMC or HMC seen a few diy VMC and one HMC not as big as a hass but can do the same work on a smaller scale,

I am one of those mad men as I make my own tools my thing is if something cost $0 to $100 i will by it, if something cost more and I can make it for less I do like the ad dons I have done for the big router a tool setter to buy them any where from $100 to a Lot my one with time added $100 its as good as a $300 one.

so for me to whack out the bearing idea just for fun posable wast of time and money it happens what ever