Out of curiosity, has anyone out here ever built their own ballscrew, i have seen threads and posts on thread screws, but if you had the technology of hard turning at work would building ballscrews be possible?

it isn't the screw, it's the nut that would be problematic. but I guess you could make that too if you had all the machinery available to you.

People do it with threaded rod all the time.

The true trick to a ball screw is that the "groove" is not a true radius. It is more like a gothic arch.

Thus, the groove is actually shaped like this:

\O/

this shape enables you to eliminate ball slop by using larger balls. Otherwise, you'd have a "bowling ball in a gutter" effect which is counter productive to accurate positioning.

"Threading" the ball screw would be done how by the DIY'er?

By using a lathe?

You'd then be using a rolled or cut acme to try to cut a ball screw which at best would be cheap copy of the thread pitch on the baseline acme. Probably be better off using good quality threaded rod (is there such a thing???).

Hardinge made a machine tool lathe that could be used to make a ball screw. They had super precise ball screws on it that had ABEC 9 bearings that were further sorted so as to hold axial lead error to within millionths...

Hmmm. If you owned or had access one of these, why would you even need/consider making your own ball screw??? At some point, the falacy of making EVERYTHING becomes apparent. Ultimately, some stuff is just better off being purchased....

Ultimately, MAKING a machine tool (part or machine) takes a MUCH more accurate machine than the machine component you're trying to make....

I agree with the sentiment of your posting, but isn't this -



Ultimately, MAKING a machine tool (part or machine) takes a MUCH more accurate machine than the machine component you're trying to make....


- a paradox ?

not a paradox but a recursion. who/what makes the machine that makes the accurate machine for making precise machines that make ballscrews? jehovahs witnesses might now chime in by saying "God, of course" but we all know that he is more into spiritual well-being rather than materials science.

but anyway, to get back on track and as nc cams says the point is not to make everything yourself. some things you're just better off buying.

Well he made the world in 6 day's, didn't he? That's material science pushing to the limit. So making an accurate machine for an acurate machine for an accurate machine must be peanuts. So for once I agree with the jehova's. Automatically comes the next question: can one pray for ballscrews?

Carel

pray for ballscrews? now there's an idea...

<gone prayin'>

Well he made the world in 6 day's, didn't he? That's material science pushing to the limit. So making an accurate machine for an acurate machine for an accurate machine must be peanuts. So for once I agree with the jehova's. Automatically comes the next question: can one pray for ballscrews?

Carel


DEAR BIG FELLA (WITH THE BIG WHITE BEARD) UPSTAIRS.....

PLEASE HELP ME FIND THE INSPIRATION, THE KNOWLEDGE AND THE SKILL TO HELP ME CREATE THE MOST PRECISE/ACCURATE BALLSCREW EVER MADE...

YOURS TRULY,

RHINO

Ultimately, MAKING a machine tool (part or machine) takes a MUCH more accurate machine than the machine component you're trying to make....

This is an interesting statement and not strictly true. If it were true, then we would have no machines at all. When the Earth was formed, a super accurate machine would have to have been created to produce all of the inferior, less accurate, machines we have today.

The Gingery book series is a good example of making machines more accurate than the tools used to build them.
He describes building a lathe from scratch with very basic hand tools. Once you get so much built, the lathe actually builds it's self. :cool:

Mrbean.

I tend to disagree heartily with Mrbean and here's why:

We make cams from our own designs. In doing so, we see continuous and often predictable profile degradation/deviation trends as you go from design data to the machining master.

There is then more change generated as you go from master to finished cam. We even see differences in the profile when we grind the cam with a "small" (13")versus a big (18") diameter wheel.

Only via empirically derived and applied compensation techniques can we "manage" this in spite of the fact that we've rebuilt and redesigned certain features of our grinder and/or our production processes - the machine simply can't do anything other than what the synergy of its parts will allow - good, bad or indifferent.

There is defininite analog "smoothing" and/or morping that takes place in the manufacturing process which results in profile deviation. Is it enough to matter???

Suffice it to say that it is manageable and/or containable once you learn how to compensate for it. Machine compliance and/or normal tolerances are merely a part of the reason why it is IMPOSSIBLE for even the best analog and some digital equipment to EXACTLY duplicate design intent.

Now, assume you're want to to make an EXACT 10 turns/inch ball screw. If you have ANY lead error in the lead screw of the lathe used to make the ball screw, the BEST you can hope for is to duplicate the lead error in your "master" lead screw (don't forget axial runout potential in the spindle bearings too which gets superimposed as well) - the more probable result will be the amplification/exacerbation of existing lead error - thus, the by-product WON'T/CAN'T be better than the original - ERROR begets error.

Having worked in the bearing industry (machine tools) and now in the cam industry, I know that the reduction of error is MANDATORY as you seek to achieve the production of more precise, error free product. The creation of better machines that do things with less composite error is critical to the creation of lower PPM defects.

Hi all,
I would seem to think that a leadscrew on a lathe would be fairly accurate and if used to manufacture/duplicate a ball screw would be accurate enough for me.

I wonder how much weighting people put on to the accuracy of thier ballscrews, in other words, why did you buy a ballscrew? Was it for the accuracy or was it for the efficiency of it? ie. it spins real easy and little power loss.

Regards M

When you work in areas where 0.000100"s of inches makes a difference, the use of precision ground ball screws is pretty much mandatory - a lead screw simply isn't in the same league

(PS there's already been a long exchange on cut vs ground lead screws vs ball screws - this shouldn't be an excuse to resurrect that thread)

When you're shooting BB's at a light bulb at a distance of 2 miles, the utmost in accuracy is essential if you want to hit the target.

If you'r using hand grenades, you can be within several yards and still accomplish your goals.

If you're using thermo nuclear weapons, +/- a mile is close enough.

How close does it (the ball screw) have to be to suit YOUR needs??? ONce you answer that question, you can choose between a rolled or ground or DIY version....

NC Cams. I have to say, I cannot argue with your statements in a pevious post. Any error in a machines accuracy will be reproduced in any part made on that machine. That is true.

So that begs the question. Where did the first ball screws come from?

If it was a more accurate type of "screw drive" that produced ballscrews, where did that come from? and how did they make that, and so it goes on.

At some point somebody must have made something more accurate, or we would still be using 3 sticks, some string and a flint, for lathe work.

If nobody can produce parts more accurate than what we have today, then the only way to go is backwards? This does not happen. So how do we get more accurate machines?

screw compensation in software

but for screw compensation in software you need hardware to run the software on. and as we all know for hardware you need component placing machines which are high-precision machines and which - yes, you guessed it - have ball screws on all 3 axes. ergo no software - no ballscrew! :D

In production processes there is always a trick to avoid the errors made by conventional machines. Like rolled ACME thread. Like how ball bearings are made. For production of a ballscrew the machine would be targeted on pitch accuracy and maintaining diameter. It is certainly not made on a lathe. Too much pitch inaccuracy, too much flex, insufficient support. In normal production you carve a product, including all the inaccuracies of the machine. Specialised production machines are made for the characteristics of the product and output easily a quality that you can never achieve with normal production equipment.

Carel

Screw compensation can NOT compensate for non-repetetitive runout that exists in bearings nor can it compensate for repetitive runout if the CNC doesn't know where it is. You have to ultimately eliminate the mechanical slop or "tell" the computer where the defects are so as to comp for them... it is doable with laser sighting these days but had to be dealt with via "perfect" machines in the early days.

This is why the machine making tools have SUPER precise and special qualified bearings. One particular machine tool maker I worked with went so far as to call out ABEC 7's for size tolerance but then they called for ABEC 9 radial runout. To achieve even fine axial runouts, they then spec'd out the use of grade 3 balls (perfect within 3 millionths) versus the grade 10's usually used in some M/T bearings.

They then called our for 100% inspection/qualification for axial runout over 17-20 revolutions. Why 17-20??? because it takes that many revs to get ALL possible axial runout patterns to repeat. Then, they ground the spindles after assembly and also ground the center OD's after the ground the ID's.

This was for a tool room spindle used to cut lead screws for their generic "machine tool" grade lathes...

Like I indicated earlier, you have to make the machine that's gonna make the machine to a higher degree of accuracy than the part you're trying to make.

Or as the old axiom states, two wrongs never can make a right....

screw compensation in software

Did you mean "you can compensate with the software" or did you mean "screw software compensation, it's bull****"? ;)

thanks for all the replies.

a mill with a warped bed can still grinda nice flat.

lets just say the machinist is the limiting factor cause in the end we all know machinery has been steadily improving ever since man has been making things.

A the other hand ofcourseeven a machinist can be limited by his machinery cause lathes are as far as i know never involved in commercial leadscrew manufacturing.

A friend been to a leadscrew manufacturer locally resently and found that the leadscrew are "tapped/die'd" in a lathelike way meaning a rod would be very slowly spun around while different cutters took different cuts seperated along the diameter of the rod in question.

the right toolin only in combination with the right craftsman is what enables you to work to tight tolerances.take scraping ,no more than a steady hand and a chizzle.

pff reminds of the chicken and the egg.

This is a seriously old thread.

Anyway, scraping is part of the exception rather than the rule.

If a good craftsman were given some high quality profile tooling (for ballscrews) to use on a large 20x80 Chinese lathe (very rigid but questionable in quality), the resulting ballscrew could not and would not be more accurate than the leadscrew on the lathe itself.

On the otherhand, if the craftsman were given a decent grinder (ABEC-7 spindle bearings for good surface finish, everything else mediocre), he could possibly grind a high precision screw by carefully monitoring screw rotation and screw traversal through rotary and linear encoders, respectively.

a mill with a warped bed can still grinda nice flat.
:confused:

You mean like a 1/2" flat on the end of a rod? Sure. Slap on a face mill and flatten an 8" x 16" piece of steel? I think not.

mind boggling.

anyone know were the motherscrew came from ?

The mothier screw. I'm laughing hard. Okay, I keep reading the general consensus of many here stating you can't make precision more precise than the machine doing the work, and due to the machine's many falicies, you will end up with a crap(pier) product than the mother machine. Yeah, I guess you're right. If. (and here is the real ringer) you accept the product that comes out of the machine and say, "It is done and as perfect as I can make it." Well, I guess if you're not a craftsman, then I guess you accept the crap(pier) product that came from your machine. That's it in a nutshell really, and it is the divide between the craftsmen and the Engineers. <sorry for the dig guys>

Now when you get the best a machine can do from the machine and you go no further with it, you pretty much are stuck with the dilemma of what NC Cams speaks. Crap(pier) parts than the oringinal machine. But way back when all folks had were FILES and those were hand made at that, machines were born. Somewhere along the line we went from files to ultra precision. Sorry, god didn't do it. No, what craftsmen did was made the part as well as they could, then made a mating part that fit and they lapped the irregularities out. Or they used a profilometer and a stone, or they found other ways to HAND work the crap into excellence. Hand work.

It always gets me when I read it is IMPOSSIBLE to make precision. Imagine needing to make a special device that measured angstroms with the crappy acme screws and slide ways of olden days. Folks nowadays would tell you impossible without air bearings and other fancy machine accesories. Sorry, it was done long ago with machines many would shudder at. Back then, they used dynomite to blow broken drills out of train crankshafts. Forget an EDM, it wasn't invented. Break a drill off in a locomotive crankshaft? Hell, take it out in the back 40, dribble in some dynomite past the flutes, and blow that sucker out. Go to the library and get a book on old machine practices. You will be amazed and no, you don't need a golden grail of a machine to produce precision. You just need to learn craftsmenship.

I read, or saw a picture somewhere, of a device for making an accurate helix through kinematic methods. The method involved using something similar to a lathe taper attachment, where the guide bar was set parallel to the hypotenuse of the right triangle representing the helix angle of the screw. A system of cables and weights kept the travelling shoe in contact with the guide bar.

Such a device would be as precise as the flat and straight guide bar could be made. There are proven methods to improve flats by manual methods, reducing the errors and improving on the machine that machined the original.

I have not really thought out whether such a contraption could be made to work, but it seems feasible. :)

You can make a more accurate thing, with a less accurate thing. You just need accurate position feedback, to "correct" the error. You need more precise feedback, and steadier tools, not more accurate tools.

And probably : scrapping a lot of parts is -part- of making ultra precision parts.

thx for digging in guys.

I think the best example of low precision parts being pieced together to create a high precision assembly would be your typical rotary table.

A phase II rotary table, made in China, is pretty cheaply made, but all the internal gearing, manages to be precise within several minutes of a degree. The same principles could probably be applied to grinding a high precision ballscrew using nothing other than budget gears and acme screws.

hmm this could make a nice feedback system when time is not an aspect of your succes.

a feedback for indexing a "high"precision screw that is.

see as long as you got quality men you can make quality tools.
wich again is quite comforting(and i might even be able to sleep again)

cheers Zumba

I read, or saw a picture somewhere, of a device for making an accurate helix through kinematic methods. The method involved using something similar to a lathe taper attachment, where the guide bar was set parallel to the hypotenuse of the right triangle representing the helix angle of the screw. A system of cables and weights kept the travelling shoe in contact with the guide bar.

Such a device would be as precise as the flat and straight guide bar could be made. There are proven methods to improve flats by manual methods, reducing the errors and improving on the machine that machined the original.

I have not really thought out whether such a contraption could be made to work, but it seems feasible. :)

Older gun barrel rifling machines also the common method used in home build rifling machines.

http://www.firearmsid.com/Feature%20Articles/RifledBarrelManuf/BarrelManufacture.htm
http://www.benchrest.com/forums/showthread.php?t=25813&pp=15
http://www.practicalmachinist.com/ubb/ultimatebb.php/topic/16/631.html

Master screws were (and sometimes still are) made, first by a fairly accurate "roughing" generation method, either on a lathe, or a grinder, or from a special purpose machine geared to produce a screw without any pre-existing model.

The screws are then hand-worked on one side of the thread face with laps using feedback from special-purpose measuring machines that can continually analyze pitch error. After several iterations, you now have a screw that is more accurate than laboratory gauge blocks.

Ironically, these all have a form similar to an Acme thread or a square thread, because you can't lap a ballscrew.

Then, these master screws are used as length references and go into master thread grinders. The master thread grinders are ultraprecision machines that require a whole building of infrastructure to support their operation. These master screw grinding machines can produce very, very accurate leadscrews fresh off the grinder of any form you desire. The first ones were completely manual machines, and after a period of time, slowly became analog (not cnc) corrected machines. You can go buy a CNC thread grinder that has been profiled and relies on the masters generated by manual methods for its accuracy.

Now, this "artifact" method of using master screws for accurate grinding has been eclipsed by machines that are guided by metrology frames, and use no form of screw whatsoever. Rather, they use capstan drives for linear axis positioning, interferometers for angular and linear feedback, and piezoelectric fast tool servos for final correction at the tool interface.

unsubscribe thread

Rail Extension Kit for Screw Drive Garage Door Opener


I posted this in another thread, but maybe folks here would be interested too.

I was wondering if anyone had ever tried using a screw drive from a garage door opener to drive an axis on a homebrew machine? I wonder how the acuracy compares to threaded rod or belt drive home systems. Not the highest, I am sure, but for woodworking machines it might be ok.

I was searching the sears.com web site looking for a part and saw this item -

Craftsman 8 ft. Rail Extension Kit for Craftsman Screw Drive Garage Door Openers $19.99