PDA

View Full Version : Do/Don't Ball Screws Have Wear?



Paul Alciatore
07-25-2017, 01:15 PM
While reading another thread where the advantages of ball screws was brought up, I have to wonder. I can see the advantage of zero backlash. And they may be more accurate, but there are some very accurate acme screws. But an acme screw will wear over time and use. So, wouldn't a ball screw, which relies on a number of small areas of contact between the balls and the screw, also have wear over time and with use? The balls would contact only a very narrow track on the screw and nut so wear may be more rapid. Won't it come to a point where it will have more wear at some points and backlash and need to be replaced, just like an acme screw? How does the operational life of a ball screw compare with that of a conventional acme screw?

Is it that the number of balls in contact with the screw and nut is large enough to provide a similar amount or even larger amount of contact area? Or are they just made better? Hardened? Better steel? Or what?

MaxHeadRoom
07-25-2017, 01:28 PM
Most ball screws are case hardened at least, as I mentioned in the other thread some are very hard, in this case the Hi-Win appeared to be tempered well below the surface.
Another thing to watch for is cold rolled versus precision ground, with the latter the most preferable, I have tried cold rolled, but was not impressed.
The pre-loading of the nut automatically maintains a constant opposing pressure to maintain zero backlash over time.
If kept clean and lubricated, the wear factor is extremely low over time.
Customarily every other ball in the nut is a fraction smaller than the rest in order to prevent any type of jamming effect.
Max.

MTNGUN
07-25-2017, 01:51 PM
One rolls, the other slides.

enginuity
07-25-2017, 01:57 PM
I find it helpful to think of preloaded ballscrews like a preloaded machine tool spindle with no play.

The elements of the bearing (or ballscrew) are loaded to the point that some deflection occurs in both the balls and the race (or balls and screw). But this stress point is carefully controlled to avoid premature failure. Very specialized high quality steel is used for ball bearings and the races.

Ballscrews do wear out - but their life is similar to roller bearings. Ballscrews have a finite life -like bearings, but it is often in the millions of cycles if properly designed and manufactured.

Ballscrews when worn also often sound like rolling element bearings when they wear out. They just keep getting louder and louder. Many times the worn screw will still exhibit zero backlash, but the sound of the barking dog or annoying slide whistle will force you to change it out.

skunkworks
07-25-2017, 02:05 PM
We have a 60's vintage cnc (nc) that has preloaded ball screws. No backlash still after all these years..

(probably over sized and slow running - but still)

http://electronicsam.com/images/KandT/conversion/xaxis/ballscrew.JPG

sam

J Tiers
07-25-2017, 03:01 PM
While reading another thread where the advantages of ball screws was brought up, I have to wonder. I can see the advantage of zero backlash. And they may be more accurate, but there are some very accurate acme screws. But an acme screw will wear over time and use. So, wouldn't a ball screw, which relies on a number of small areas of contact between the balls and the screw, also have wear over time and with use? The balls would contact only a very narrow track on the screw and nut so wear may be more rapid. Won't it come to a point where it will have more wear at some points and backlash and need to be replaced, just like an acme screw? How does the operational life of a ball screw compare with that of a conventional acme screw?

Is it that the number of balls in contact with the screw and nut is large enough to provide a similar amount or even larger amount of contact area? Or are they just made better? Hardened? Better steel? Or what?

Ball bearings, and presumably ball screws also, do not "wear out" at all, unless something gets into them to cause wear, grit etc.

They SHOULD fail by metal fatigue where chips spall off the surface of balls and screw, until it is too noisy, as mentioned. The particles could end up causing wear, but they don't come from wear themselves.

The metal fatigue comes from the balls and screw deforming under the preload plus working load.

danlb
07-25-2017, 03:39 PM
The metal fatigue comes from the balls and screw deforming under the preload plus working load.

In other words, yes, they have a limited life and thus they do wear out. :)

J Tiers
07-25-2017, 04:22 PM
In other words, yes, they have a limited life and thus they do wear out. :)

"wear out".... and "limited life" are not the same thing, you need to think about that a bit more before popping off.

A journal bearing "wears", the shaft rubs when starting (before the oil is holding it up), or if heavily loaded, and "wears" the bearing.... A sliding lathe carriage can rub and "wear" on the ways... they rub so as to remove material. That would be "wear".

The ball bearing does not "rub" the way journal or sliding bearings do*, so it does not "wear". The ball bearing companies make that distinction, are you that much smarter than they are?

* there may be a slight amount of "skidding" in some instances, and the non-point contact of the ball has to have a minute sliding action, but the failure mode is not from those, but generally from spalling off of pieces.

A.K. Boomer
07-25-2017, 05:02 PM
"wear out".... and "limited life" are not the same thing, you need to think about that a bit more before popping off.



Excuse me --- but really didn't think Dan was "popping off"

and in fact his answer is far more correct than that of yours, ball screws wear plain and simple, just the act of using a ball screw creates wear - ball screws are nothing more than un-caged ball bearings working in a harsh environment many subject to a "soup" of particle hash for oil that most of us call "lubrication"...

their is non-uniform "scuffing" action in various areas of the balls when engaging in various area's of the screw, just a given when you get to know terms like "elasticity of materials" and how it effects things like "engagement geometry's"

not to mention - did I mention un-caged? yeah that ensures spherical high load contact between certain balls that are trying to "outrun" certain others in any certain given circumstance such as engagement geometry deflections and such, that's where two or more balls start contacting each other under load, and round is round so the surface area has incredible unit pressures against it - not only that - they are traveling in opposite directions when they contact and at twice their normal outer parameter speed, the results are WEAR in the form of high unit pressure scuffing...

now add to that all the recirculating factors that keep the balls in constant high speed scuff motion whilst getting transferred from point A to point B and yeah your goddamn right ball screws just plain wear out, takes a while - but they do...

think about that next time you go popping off to someone...

MaxHeadRoom
07-25-2017, 05:08 PM
We have a 60's vintage cnc (nc) that has preloaded ball screws. No backlash still after all these years..

(probably over sized and slow running - but still)

sam

I can vouch there is many like it out there today still producing, the likes of old Mazak's etc., It is usually the DC servo's that are done first!
Max.

J Tiers
07-25-2017, 05:31 PM
Excuse me --- but really didn't think Dan was "popping off"

and in fact his answer is far more correct than that of yours, ball screws wear plain and simple, just the act of using a ball screw creates wear - ball screws are nothing more than un-caged ball bearings working in a harsh environment many subject to a "soup" of particle hash for oil that most of us call "lubrication"...

their is non-uniform "scuffing" action in various areas of the balls when engaging in various area's of the screw, just a given when you get to know terms like "elasticity of materials" and how it effects things like "engagement geometry's"

not to mention - did I mention un-caged? yeah that ensures spherical high load contact between certain balls that are trying to "outrun" certain others in any certain given circumstance such as engagement geometry deflections and such, that's where two or more balls start contacting each other under load, and round is round so the surface area has incredible unit pressures against it - not only that - they are traveling in opposite directions when they contact and at twice their normal outer parameter speed, the results are WEAR in the form of high unit pressure scuffing...

now add to that all the recirculating factors that keep the balls in constant high speed scuff motion whilst getting transferred from point A to point B and yeah your goddamn right ball screws just plain wear out, takes a while - but they do...

think about that next time you go popping off to someone...

We are going to disagree on that. Ball bearings "wear" like that too....if you leave them in dirt.

Quality machines do not have the ballscrews wallowing down in the mud. Crappy machines do. Don't let what a crappy machine does to a ballscrew or ball bearing define the ball bearing or screw.

Check what skunkworks posted to see what a decent machine does. 40+ years and no significant "wear".

A.K. Boomer
07-25-2017, 05:42 PM
JT - it does not take hashed out oil to create high loading spherical contact,

and there's no doubt they are superior and flat out amazing - try comparing a plain bearing to a ball or roller and you will get the same results, iv ran vehicles with over 325,000 miles on their wheel bearings, does not mean the wheel bearing were not "worn" --- of course they were, worn to a polish and although a little loose still hanging in there - if it's not broke don't fix it,
same with the old CNC's ---- and that little bit of assembly pre-load is probably still covering for some of the WEAR that has accumulated over the years....

The question you need to ask yourself is this; Do you believe I could have gotten 325,000 miles out of automotive bearings had they been "un-caged"

allow me to answer that for you ----- Hell No

CNC ball screws fall short of my automotive example make no mistake - they WEAR out also and in fact do it at an increased rate.

Boostinjdm
07-25-2017, 06:44 PM
Nevermind

Forrest Addy
07-25-2017, 07:13 PM
"Finite life" suggests planning for eventual replacement, which includes having spares on hand if you anticipate the machine will still in your possession. Finite life also implies the limit of tolerable deterioration. In the matter of ball screws the failure mode is different than for parts subject to sliding. The rolling elements and their paths are subject to repeated compression loads. The "wear" life may be imperfectly expressed as an exponential function of unit load and load cycles. Failure suggests a binary condition where something is OK until it fails. Deterioration suggest a gradual process over which the condition and performance degrades to the limit of operability where, hopefully, a repair or replacement can be executed. There's a whole branch of plant engineering that discusses the ways, means, and economics of equipment maintenance.

Failure mode in rolling element bearings is fatigue. After a finite number of load weighted rolling cycles, the rolling elements and/or their paths begin to fail along grain boundaries which leads to cracks which which propagate to join. Particles or even flakes spall out like a chuck holes in a paved road. The process is usually gradual and more or less predictable.

Pre-loaded ball screws when used in precision machine tools are quite predictable and may when error mapped be used for position sensing. Generally about the time you can hear more noise from the ballscrew nut than the usual faint whir it's time to stock spares and schedule an outage for replacement - or find a buyer. When the ball screw gets to the rumbling stage you're not far from major failure and replacing the ball screw - or the machine - is getting urgent.

Anyway, wear for rolling element components is seldom based on material loss but on rolling path fatigue. If the race under magnification looks like a bad road, more time in service isn't going to improve it.

darryl
07-25-2017, 07:37 PM
Over 1/2 million miles on my Land Cruiser wheel bearings, no replacements done yet. 24k on my friends Subaru and wheel bearing needed replacing- something else to factor into it. Considering that a ball screw probably is made to some specified degree of precision that's higher than what a wheel bearing would normally be, I'd say that a ball screw should have a very long life indeed. Barring of course the intrusion of grits, etc.

Dan_the_Chemist
07-25-2017, 07:46 PM
I have a 1952 David Brown 990 tractor. It has a ball screw arrangement between the steering column and the steering arm. I had to disassemble it last year to work on a cast steering yoke that had broken, and the ball screw and ball nut were still in good condition after 60 + years of yanking and cranking on this old tractor's steering...

A.K. Boomer
07-25-2017, 07:58 PM
Considering that a ball screw probably is made to some specified degree of precision that's higher than what a wheel bearing would normally be, I'd say that a ball screw should have a very long life indeed.

"Probably made" leaves allot wide open, for one keep in mind what I stated earlier, they are inferior in comparison in design from the start due to being "un-caged" if your land cruisers bearings were un-caged they would only have made it a fraction of what they have endured... ball screws may or may not be more precision depending --- but they are stone age archaic in design when compared to a common Yugo/Vega or Pinto wheel bearing...

there are exceptions to this rule if they make them, if you introduce ball screws that have dished nylon discs inserted in-between the balls and take up all space with just a little left for freedom of movement then you eliminate the ball on ball contact, you also eliminate the "double speed" factor, you now have a pretty stable system

this system is used in Kubota's track hoe's main swivel bearing - it's called a free cage.

garyhlucas
07-25-2017, 08:42 PM
There are two basic ways of achieving preload on a ball screw. The most common is loading the screws with oversize balls. It is not uncommon to have the screws reballed at some point where they load a set of slightly larger balls. The other method is using double nuts loaded against each other which means they can be adjusted for preload and wear to some degree. My THK nuts are rated for 6,000 kilometers of travel at rated load.

wierdscience
07-25-2017, 09:11 PM
Both systems can and do outlive their owners if designed correctly.Both are subject the the same failure modes as mentioned grit and dust.I salvaged the leadscrews and nuts out of an older model Okuma CNC mill.It had multi start Trapezoidal leadscrews 50mm in diameter and the nuts were 150 mm in length.Even though the machine's hour meter had rolled over past 10,000 hours,there was maybe .0001" backlash present in the screws.These were protected with bellows and fed lube oil on a timer driven pump.

oldtiffie
07-25-2017, 10:11 PM
https://en.wikipedia.org/wiki/Ball_screw

https://www.google.com.au/search?site=&source=hp&q=machine+ball+screw+accuracy&oq=machine+ball+screw+accuracy&gs_l=psy-ab.3...4668.37684.0.41967.29.29.0.0.0.0.410.5280.2-17j2j1.20.0....0...1.1.64.psy-ab..9.14.3779...0j0i131k1j0i10k1j0i13k1j0i13i30k1j 0i13i10i30k1j0i22i30k1j0i13i5i30k1j33i21k1j33i160k 1.be3gR247Ri8

https://www.google.com.au/search?q=machine+ball+screw+accuracy&source=lnms&tbm=isch&sa=X&ved=0ahUKEwjqoqnh7qXVAhULUrwKHSYRDYUQ_AUICigB&biw=1536&bih=719

Andre3127
07-25-2017, 11:08 PM
Every mechanical object with moving parts eventually wears out, unless we're talking hydrodynamic or hydrostatic bearings. It's inevitable.

Sent from my XT1053 using Tapatalk

A.K. Boomer
07-25-2017, 11:49 PM
Every mechanical object with moving parts eventually wears out, unless we're talking hydrodynamic or hydrostatic bearings. It's inevitable.

Sent from my XT1053 using Tapatalk

Hey - what can I say - finally someone with an understanding of how things actually work...

A.K. Boomer
07-26-2017, 12:10 AM
Even though the machine's hour meter had rolled over past 10,000 hours,there was maybe .0001" backlash present in the screws.

at an average rate of 50mph that's right up there with Darryls 1/2 million mile toyota land cruncher claim, and yet keep in mind there's no way he was averaging 50mph so you can probably about double that 10,000 hour figure,

also keep in mind whatever axis is being used on the machine most time's it's sitting completely idle as the hour meters continuing to run,,, not so with an odometer and wheel bearings... just saying, little perspective please and some respect for the almighty caged low quality bearing --- engineering is everything...

skunkworks
07-26-2017, 12:59 PM
No backlash compensation needed...


https://www.youtube.com/watch?v=FgOqEz5Tk-Y

skunkworks
07-26-2017, 01:06 PM
and that is moving around a couple ton...

ikdor
07-26-2017, 02:59 PM
Considering that a ball screw probably is made to some specified degree of precision that's higher than what a wheel bearing would normally be, I'd say that a ball screw should have a very long life indeed.

Given that I'm rather involved with making wheel bearings I object to that remark ;)
Apart from seal failure and hard cornering wheel bearing life is virtually infinite. Almost all fatigue comes from cornering. Now these bearings are ground to extremely tight tolerances whereas a lot of ballscrews are just rolled. Even the ones that are ground are not as accurate as it's really hard to grind an accurate helix as opposed to making a circular raceway.
Igor

A.K. Boomer
07-26-2017, 04:46 PM
Pro's and Cons

Like Ikdor states cornering is hard on conventional bearings and that is due to using a bearing in a way it's not really intended, kinda half splitting the load and trying to "shear" the balls in half with half races, creates allot of unit pressures on both the spheres and their races,,,

this is actually how a ball screw functions "normally"

But, ballscrews do have an advantage when being used this way because they can "stack" as many rows as they want to help "share" the load and conventional bearings can't,
But, there's only so many rows you can stack before it becomes irrelevant due to torsional deflection of the ball screw itself,
due to this effect the lead ballscrews from the driven side take the brunt of it all and as the screw progresses the load lessens after so many rows to the point of doing nothing at all but creating extra moving parts...

yet it is this stacking effect that contributes to the reason as to why they do so well, come up with a way to keep the balls from touching each other and you really would improve the longevity even more so, but from many of the examples given on some of the older machines that might be overkill...

there is no doubt the good quality ones do very well as is...

wierdscience
07-26-2017, 07:17 PM
and that is moving around a couple ton...

Nice!:cool: What's the machine? Horizontal mill?

skunkworks
07-26-2017, 08:12 PM
It is an old Kearney and Trecker Milwaukeematic IIIb.

(using it as I type...)

http://electronicsam.com/images/KandT/oldkandt.JPG

It is my favorite machine to run. It just works. strait and true. (only complaint is coolant control...)


https://www.youtube.com/watch?v=39q6kvrSBSk

wierdscience
07-26-2017, 08:28 PM
Cool,am I seeing right it uses straight shank tooling and a master collet arrangement of some sort?

oldtiffie
07-26-2017, 08:46 PM
As a ball screw will move under load it will be dependent on the magnitude of the load and its directions.

As a screw is used to "push" or "pull" a load the reaction/s of the screw (reaction/s) will be rotation, compression, pull/push.

https://www.google.com.au/search?q=dovetail+calculator&oq=dovetail+dimensions&gs_l=psy-ab.1.1.0i71k1l4.0.0.0.578652.0.0.0.0.0.0.0.0..0.0. ...0...1..64.psy-ab..0.0.0.oXcSFM2i3Ys

wombat2go
07-26-2017, 10:15 PM
I worked a bit ( a long time ago) on automotive driveline production machines, running 2 or 3 shifts,"365" days cycling every 15 seconds or so ( typical)
The outer loop displacement feedback was independent of the ballscrew.
The 4 quadrant motor control had some basic analog and uprocessor intelligence to halve its velocity iteratively at waypoints close to setpoints to
increase displacement accuracy as the ballscrew wore out and to reduce shock and wear on it and motor.
As the recirculating ballscrews and printed motors wore out, they were just replaced as spare parts.
Sometimes, I recall, we got some back and cut open to see.
They were not repairable parts.

skunkworks
07-27-2017, 03:44 PM
Cool,am I seeing right it uses straight shank tooling and a master collet arrangement of some sort?

Yes - strait shank. K&T used this system for a few years. The newer generation ( like the 800) used a smaller version of this style. (mechanically barcoded too)

This is our spare spindle guts..

http://electronicsam.com/images/KandT/conversion/spindle/spindle.JPG