[handlewanker]

LOL......yes, some people would take the view that anything they come up with is the answer, no matter how ridiculous, and if you point out the shortcomings they get very annoyed and list you as a troll etc.

The ball screw solution is not a simple one. but it does demonstrate that many solutions can fit other problems too.

The biggest problem to overcome is the fact that the screw is a relatively soft material and the principle of the ball screw is a rolling one.......that means hard metal on hard metal.......have one of the pair soft and it gets "transformed/reformed" by the other.

The problem is centred around the need to provide a backlash, frictionless drive that will economically replace the ballscrew for DIY purposes.

The drive you showed will work within it's limitations provided the load is proportional to the strength of the materials, namely the softness of the screw, which is usually available off the shelf or already on a machine.

When it's on a machine it drives the table, whatever, by sliding the flanks against a sacrificial bronze nut.....the nut being the easiest part to replace......and that is the problem......the design of the screw will not be practical if you roll it against a hardened surface....getting reformed in the process.

If by some magical process the screw could become hard without changing the dimensions of the very precise pitch and linear straightness, then the internal race of the bearing you showed would drive against the hard screw flanks, actually on the corners only, but without proper lubrication would both eventually grind themselves to dust.....this is like a ballrace that has no lubrication in it.....it will roll freely for a while until the tracks and balls pit and destroy each other......add some lubrication and the set-up would probably last for a longer time, relative to the load applied to it.

The process of hardening a soft screw still leaves it with a soft core that would not last long under the conditions I stated, but within a certain load factor would probably be very friction free and last forever...IE, a hand driven device.

The device you thought of is as simple as you can devise it, but the moment you change the status of the screw by hardening it, other factors arise and so you get into a "chase my tail" scenario trying to eliminate all the problems as they occur.......the ballscrew solves all the problems except the cost factor.

The making of a ballscrew alternative is also relative to the cost of the labour hours to produce it, and taken purely as a labour/hour cost factor, the ballscrew is the winner all the way, and totally readily available off the shelf for retrofit purposes.
Ian.

[BrianTee]

DIY often has little to do with ROI's and time saving! But sometimes it just doesn't make sense either.

So, a concave bearing on a concave thread would wear excessively. Agreed. A convex on convex would not however. Take a small bearing angled equal to the thread pitch with the outer race pushing against the thread. Two bearings required, one to push each direction. Opposite that bearing 2 bearings (pressure rollers) to locate the ACME in place. Each 120 degrees apart around the threaded rod of course.

Maybe i need to do a picture.

[handlewanker]

Ahhhh, a picture, my kinghdom for a picture......LOL....that would save a thousand words.

I work with a Wacom graphics pad and create sketches in Photoshop, saving them as Jpeg images, which saves a lot of scrap paperwork, but you can do the same with a mouse and MS Paint.

The big problem still remains, hard steel rolling on soft steel, even if the outer race were to be ground at an angle to fit the thread which for Acme is 29 deg included angle.

You have to consider that the screw thread has to transmit axial thrust against the flanks of the bronze nut in the normal screw nut set-up, but if you introduce a bearing with a thread profile in place of the nut and fit it into the thread, the bearing is designed to carry radial loads whereas the screw transmits axial loads and will destroy the bearing when any axial force is applied to it.

So far, in the 69 pages of this thread, all these have been suggested and some tried, but the short answer is, the screw is naturally soft and being so is a non starter.

I agree that ROI is not a consideration for DIY, but expense without result is also a poor answer to an alternative for alternative's sake.

The most important part of a CNC machine is the production of an extremely predictable resolution, without which you do not have a CNC solution.

So far, on a $5,000 CNC machine, the ballscrew represents a very small part of the total investment to get that extremely predictable movement each and every time for a very long time.....that's a hard act to follow with an Acme screw thread basis......a ballscrew will make a silk purse out of a Sow's ear every time.
Ian.

[BrianTee]

Ok a picture. Regarding the axial thrust comment, the bearing against the thread creates the correct forces designed for a ball bearing.

simple pic



A ring of brass over the bearing would suffice as a softer metal.

[handlewanker]

Hi, the pic is self explanatory.....the problem of brass on steel with point contact in this configuration is that the brass would quickly wear into a groove and eventually become lumpy and bumpy, hardly the requirement for a smooth frictionless drive.

If you started with a brass ring and cut a groove to match the corner profile of the thread, it would probably last a bit longer, but the problem of the soft metal on a harder metal with point contact still remains.

If you recut the screw thread to have a 90 deg thread form instead of the 29 deg Acme form, and rolled the bearing outer race against one of the flanks at an angle, like you show in your pic, you might get a drive as the bearing would be running in the radial mode it is designed for, but the soft reformed screw thread would eventually have indentations from the hard bearing pressing outer race into it.......unless the screw was case hardened and possible ground it might work, but the labour input would make it only feasible as a DIY project.

Taken a step further, you would have to have a balanced configuration to maintain close contact, as you previously stated, a single roller and two plain rollers at 120 deg, the two plain rollers pressing against the top of the thread and the ballrace pressing against the flank of the thread form.

You would probably get a better result if you had three ballraces space at 120 deg round the screw and space at 1/3rd of the pitch of the thread, each pressing against the 90 deg thread flank.......one of the ballraces would need to have a spindle with an eccentric mount to allow depthing in the screw thread to maintain contact.

The next problem in this "design" configuration would be that it was OK for driving forward, but you would need a similar set-up for driving in reverse and each ballrace set tensioned against one another to eliminate backlash.

The last problem would be the accuracy of your lathe leadscrew when you recut the 90 deg thread form, for that determines the feasibility of resolution as opposed to just moving the carriage, and if this is not accurate then the whole project is null and void.

BTW, in the pic you show, the ballrace is pressing against one corner of the thread while the other corner is rubbing against the ballrace side, totally impractical.

In any alternative screw drive design we are running side by side with ballrace technology and plain bearing technology...one rolls the other slides.......so far the ballscrew has succeeded in ticking all the boxes for cost, life expectancy, load carrying and resolution by successfully going down the rolling technology path in all aspects of the design.
Ian.

[BrianTee]

Definitely the ball is the winner today, but new ideas are constantly thought up, just not this time, thanks for the feedback. You've been on this one since the start and Mike is still building really cool things.

I went back and read thru the whole thread and found the references to topics i missed and found the thread about the video i posted.

I'll keep thinking about the DIY solution, cheap and simple. I do think that many of the DIY CNC hobbyists are less concerned about long term durability and just want to build the thing and only use it once in a while. Especially the wooden machines that expand and contract during the day. They just want a workable inexpensive idea that is 'neato'. That might be important to keep in mind too.

[BrianTee]

BTW, Grinding the edge of bearing conical, or pressing on a tube over the outer bearing race of a diameter approx 3 times that of the thread would allow that bearing/tube to sit flat against the ACME thread and not contact the other threads.

I'll post again when i make a model, or not if this idea is crap. Probably the latter

Brian

[handlewanker]

Hi, you have patience reading all the thread.

There was a lot of work done on the roller screw, as it showed promise, but came up against the attrition of the bronze rollers against the steel screw.

There is a possibility that Delrin rollers would mate with a steel screw without being worn away as this material can flex in a small way and also not be subject to wear from lack of lubrication......food for thought.

I have a pallet truck that has Nylon wheels and that is designed to run on a hard floor with 2 tons load capacity.

The top corners of the Acme thread form would have to be smoothed off to prevent cutting into the plastic if this was tried, but for wooden or aluminium type gantry routers this might well be a very practical solution as it can exploit the resolution of the Acme thread pitch and also with care have very little backlash.
Ian.

[BrianTee]

Lots of UHMW plastic is used on seeding equipment that scrubs the ground all day and wears better then abrasion resistant steel. Stuff is tough but i have not found it to machine smoothly. And its soft.

[handlewanker]

Hi, you're right, the plastic type materials are resilient and resist abrasion because they give way and don't present a solid wall to be broken down.

I cut plastics with tools designed to handle the material, and that means you can't machine steel with a wood chisel or cut stone with a HSS tool designed to cut steel etc etc.

The wheels on my pallet truck are a typical example of the application of resilient material on a hard surface.

If I ran a pallet truck on my concrete floor all day and the wheels were made from steel, eventually the steel wheels would show some wear, but the concrete floor would be totally kaput........plastic wheels do not have that problem, neither does the concrete floor.

It is a possibility that a plastic roller could work with a steel thread if designed accordingly, but the key word is designed.

So far I have not seen anyone designing in plastic as a viable material mix for screw applications, thinking .... yes, but no tangible results to date.

I would go so far as to suggest that the roller screw, which died a slow death with the bronze rollers, could be a solution to a ballscrew replacement/substitute if the right materials were used, and plastic, namely Delrin etc comes to mind as a candidate.

I use plastic gears on my 5" Colchester Bantam lathe in the screw cutting train, and have had no trouble either in making them or using them.

Provided the contact point is a rolling application and not a sliding one, plastics can work very well.

It's a possibility that needs to be explored.
Ian.

[lwill]

I ran across this and it reminded me of this thread.
I think some one suggested some thing like this a while back.
DIY Ballnut using roller bearings - STEP / IGES, STL, Other - 3D CAD model - GrabCAD

[handlewanker]

Hi Iwill, now that's a very nice rendition of the device.

You're quite right, it has been noted before but not in quite as good a detail.

One of the problems facing the constructer using this type of design is that the soft screw thread will rapidly get "reformed" from the 29 deg Acme form, (or any other thread form), to something more like a Vee thread but with an almost 90 degree form, caused by the hard bearing rolling the soft screw like putty.

If a screw thread was cut with the form that agreed with the contact area of the bearing and then case hardened you would have a good substitute for the regular ballscrew and it would be well within the ability of the average DIY'er with a lathe and some extra money for the send out case hardening.......but any distortion would make it scrap from the start.

To DIY cut a special thread form away from the normal 55 or 60 deg standard threads is not hard to do, but to get pitch accuracy is quite a different matter.........your best effort will be whatever the resolution of the lathe leadscrew has, and any subsequent distortion from the casehardening makes it a good exercise for friction free drive but a poor one for thread accuracy, and CNC is totally reliant on the ability to have accurate resolution from the screw thread pitch.
Ian.

[lwill]

I just found it, did not draw it Just been lurking in this thread. I too feel there is a better solution out there, just haven't found it yet.
But if you look closely, I think there is a lip formed on the inside of the bearing to match the tread more closely. Could this be a ring pressed into the bearing? (not machine the hardened race itself)
I agree that it would still not be a perfect match since the bearings would be making contact to the threads in an arc, where the threads are more at an angle.
Like //(///)// on top and //)///(// on the bottom. if the threads were shaped to match with a curved profile sort of like _()_()_ it would fit better. Or maybe two starts like _(\_/)_{\_/)_ with each bearing in the correct start. But as things wear, the "pitch" or the bearings would get closer together, and on each bearing get wider causing more wear / resistance.
Even with a 90 deg thread the spacing would change as the angle of the bearing increased over time as it wore in unless everything was hardened to the extreme.( but then again doesn't this even happen with ball screws over dime?)
Maybe machine bearings / rings, harden them, run them repeatedly on a relatively soft threaded shaft to form the threads to match, then harden the shaft?
I have no idea how to do the math of trying to predict the resulting shape But you would still have the problem of the "pitch" of the bearing changing with the angle.

My $.02 as devils advocate.

[handlewanker]

Hi, there are 70 pages containing 833 posts on this topic, and you'll find many ideas attempting to replace the ballscrew, but when it comes to replacing.....the problem attacked so far is not one of replacing but substituting which means you use something completely different to do the same job...........but cheaper.

There is a very good reason why ball races are not made from soft steel.

If you like sleepless nights read up on the early posts on the roller screw solutions and then add your thoughts to the solutions posted so far.....you might get lucky and go forward but much further on the backs of giants....LOL.
Ian.

[lwill]

Yes, as I said, I have been lurking, from the beginning, and HAVE read it ALL.
This was just my input, so make it 835.

[handlewanker]

Hi, the Saga of the Roller Screw in 835 instalments, coming to your place this week.....LOL.

It isn't an easy task, but there have been a number of near hits so far, but nothing really to be a definite replacement for the ballscrew as it now stands both in price and simplicity.

Even though a solution in design might be found, the design is to replace the cost of the ballscrew not the design, and with that aspect the materials used by the average DIY'er would not stand up to the useage needed to get the resolution or lifespan that a ballscrew gives.

I think the positioning aspect should be the main drive for a substitute for a ballscrew, something that enables a table to be moved from A to B and Back to A by co-ordinate knowledge in the control mechanism........that would enable even Allthread to move a table without having to rely on the pitch of the thread to be the deciding factor for position.

If a linear glass scale can be attached to a table and the position of the table travel appears instantly on a digital display which the computer is aware of, what then do you need a ballscrew for?

This is so much like a GPS system but on a nano scale, and glass scales are available quite reasonably priced for any length the average CNC hobby machine would use.

Using a system like that means the drive mechanism per se, is totally taken out of the equation when it comes to backlash, wear and accuracy of resolution.....and glass scales do not wear out or miss steps from continued use or high travel rates, provided of course the read head does not miss reading a few lines as the scale moves past it.

I'm quite sure if I can read a 7 segment display and interpret the reading then so can a computer to determine how far the table has moved from zero to a position and back again.

One of the big advantages of a ballscrew is that it can hold a position without moving forward of back unless the drive motor is activated, and this enables climb milling to be achieved which is impossible with an Acme type screw thread unless some form of braking mechanism or buffer is employed.
Ian.

[lwill]

One of the big advantages of a ballscrew is that it can hold a position without moving forward of back unless the drive motor is activated, and this enables climb milling to be achieved which is impossible with an Acme type screw thread unless some form of braking mechanism or buffer is employed.
Ian.

I believe you have that backwards.

And the OP and title of this thread was about different (better) options, not cost.

[handlewanker]

Hi.....backwards???....you mean you think you can climb mill with an Acme threaded table....I've never seen it done in my 50 years of engineering.

BTW, you are quite right.....the thread has explored the different at great lengths, but not for the better so far.

I think I read in the first few posts (I'll have to go back and have a read to clarify that) that the object was to overcome the cost of the ballscrew, otherwise why attempt to make an inferior device that so far has no design credibility to write home about.

It would seem everyone has given up on the attempt as no new designs have been offered or made.

As far as cost is concerned, my time would be more valuable by the hour for a buy in ballscrew than spending a lot more hours at cents per hour to attempt to make something that the experts have full engineering facilities to achieve.

I think you will find that the prime object being expressed in most of the threads is to make a cheap alternative to the ballscrew despite the title saying otherwise.
Ian.

[handlewanker]

Hi all........has it ever been noticed that the current design of the ballscrew is based upon the design of the deep groove radial bearing in all it's parameters except the ball screw is a linear traveling device as opposed to the ballrace which is radial?

This puts the current ballscrew design in the same category as the deep row balrace being used to carry primarily a thrust load and no radial loads, and as such is working under very poor conditions.
Ian.

[jhovel]

That seems like very good analogy!
So maybe a design that mimics back-to-back angular contact bearings would be better. That would take two ball circuits with independent ball return channels. The ballscrews would need to be two-start with opposite geometries. Ideally, the nuts would be interleaved and adjustable in relation to each other to set the preload.
I don't think that is technically all that complicated or difficult.
Perhaps someone could come up with a demonstration design and sell it to a ballscrews manufacturer....

Cheers,
Joe