[cncuser1]

I have a 3/8 ball screw, .125" lead, with about .007" backlash.
It uses .0627" balls.

Would it be possible to reload it to reduce the .007" backlash?

[Adobe Machine]

yes, but you would be suprised how little of oversize will reduce backlash. I would advise that you checkbefore re-load)
1) your fixed angular bearings ( search many , many discussions on this subject, especially by CNC Zones resident bearing engineer, NC Cams.Some are very technical and hard to digest the 1st reading, but essentially the gist of the white paper is cheap,improperly mounted, worn or improper preload equals backlash, not fixed by oversize ball nut bearings.
2) The ball nut mount : sloppy, loose improper mount

Both of the above can be checked with a good tenths dial indicator.

If, after careful inspection you do determine that it is in the ball nut, and the ball screw or shaft is not worn ( usually in the center ) then re-load with some increment of ten thousands ( as in .0003 to.0007).Some even re-load
with every other one an oversize. Better experiment with different oversize
1st, do not let the ball nut get in a bind, you could damage the ballscrew or nut.
Do not use a magnet to pick up the balls, they become magnetic and begin attracting metal slivers that really screw things up. as you can imagine.

Over size chrome steel balls,series 52100 can be viewed at www.salemball.com, click on chrome balls, click on inventory, go to "inches", should give you an idea of inventory in your needed size.

more than any thing good luck

ADOBE (OLD AS DIRT )

[cncuser1]

I ordered balls, from Salem, oversizes by .0003" and .0007", and .0014".

I'll report back with results

[Adobe Machine]

kool..just one more thing...seclude yourself ..have a real clean area with lots of light...no phones, wives, kids or dogs ( usless its a real neat dog )you will have it done and working in no time..

down load : www.rockfordballscrew.com/manual.htm....costs nothing..has the best pictures and instructions of " how to" re-load ball nuts I have found.
Good Luck

Adobe (old as dirt )

[cncuser1]

I developed my own technique:
I back off the nut on the screw so that when looking into end the ball nut the opening of the retun tube is visible. ( the nut has two holes in it for each return tube. the hole is obviously visible on the outside of thhe nut. It is also visible also from the inside of the nut too. With the tube inplace perform the stuffing operation)

one ball at a time stuff the balls into the return tube through the open end of the nut.

Every once and a while screw the nut fully onto the screw to test the action. Then back off the nut again from the scew, like before, so the opening of the return tubeis just visible from the open end of the nut, stuff the balls one by one into the tube. I do not use grease while stufffing.

I use tweezers with a small indentation in each jaw to hold the ball. I place the ball, then push it into place with thhe tweezer.... worked pretty good.

[cncuser1]

Well I reloaded two ballscrews
I used .0642" balls
I have about .003" backlash in both.

I was hoping for better, I don't know If i will continue or move on.

[Adobe Machine]

wow, you used the .0014 oversize and reduced backlash by .004...:
Just a couple of questions...was this a new or used shaft/ball nut combo ?
Have you checked the thrust on the ball shaft with a dial indicator ?

I have no experiance with this small of shaft, but have a friend that designs, repairs manufactures small equipment that uses the smaller ball screws. Will call him Monday and find what his input is on your problem.. Post as much info on the ball screws as you can so I can ask him intelligent questions.( he owes me one any way )

Adobe (old as dirt )

[NC Cams]

DON'T FORGET TO CHECK THE GIBB ADJUSMENTS. Lots of slop can occur here and no amount of nut preloading will eliminate it.

and

DON'T FORGET TO CHECK THE PRELOADED SUPPORT BEARINGS. Wear or lack of preload/rigidity in these bearings will cause LOTS of initial take-up lash, especially at direction changes. To eliminate slop here, you want HIGH preload with HIGH CONTACT ANGLE ball bearings mounted in DF configuration.

Once you get the ball nuts to the point of having a mechanical feel when they turn, you're probably going to have to look elsewheres for any further backlash elimination!!!

THERE IS NO SILVER BULLET FIX!!! It is a systematic and careful CUMULATIVE elimination of slop that you have to eliminate so as to get low/minimal backlash in an axis of travel.

NOTE: you didn't say if you had rolled or ground screws. Some rolled threads have a true "radiused" ball raceway. However, precision ground raceways are often ground in a "gothic arch". Thus, a rolled raceway looks like a "U" whereas a gothic arch is more like a "V" albeit with a slight radius on the sides of the V to better conform to the ball.

Since the U shaped groove has more horizontal travel potential at the ball contact point than a V shaped one, you might not be able to get the axial stiffness (lack of slop) in a rolled screw that you can in a ground screw due to the forming geometries used for the raceway.

[cncuser1]

Just a couple of questions...was this a new or used shaft/ball nut combo ?
Have you checked the thrust on the ball shaft with a dial indicator ?

.......Post as much info on the ball screws as you can so I can ask him intelligent questions.( he owes me one any way )

The two sets are the same, about 10-12 years old.
They were not used often, but were improperly disassembled a few times and some balls were lost BUT I took a good look(10x magnifier) at the wear surfaces and no serious defects. In fact the black oxide finiish was still visible. Backlash is the same at centre ( where it is likely to have been used the most) as well as at the end of the screw ( where it is likely to have been used the least)

Black oxide to me means rolled not ground screws. I have no info or part numbers, I am taking this mmachine over from Dad. He completed it in the 50's and the ballscrews were added much later in a re-build.


I measured the backlash in the following 2 methods

method 1: lock down gib, verly lightly turn handlle and read off amount of turn before ANY restistance is encountered...about .004"

method 2: unlock gibs, dial indicate from saddle to table. Push and pull on tablle with the weight of my body. Take care to not twist table and therefore measure play in gib. Take care not to back drive ballscrew. Both axiis measured about .004"

I tried to measure the backlash of the screw/nut assemblly on thhe bench but couldn't get everything to stay still enough.

I have posted elsewhere about the axial play in my needle thrust bearings, I don't think this is the problem, but may just remove them and re-measure the assebmled machine ( minus bearings) for backlash

One night this week I'll post some pictures here of the relevant items, I just have to pull the machine apart...again.

[Adobe Machine]

cool,will go from there...now I guess that this machine was designed with needle thrust bearings in combination with possibly a simple radial bearing ?Possibly NC Cams could lend his knowledge here, but I do not think you could take up all axial thrust with needle bearings and thrust washers with out damaging the bearings or washers, this is why they have angular bearings mounted in DF ( to you and I thats face to face ) with fitted spacers for pre-load...could you draw a sketch of the support bearing arrangement for us to look at ?

Understand that measuring shaft movement is kinda difficult, but what I do is find the end of the shaft that was used for the live center on the lathe ( you should see a 60 deg. hole in the end of the shaft) find a ball bearing that will fit in the hole & I use marine trailler wheel bearing grease to hold the bearing in place...that will give the end of your dial guage test indicator a good surface to ride on.Roll the table manually to 3/4 of guage travel and reverse , if the guage moves then you have found the source of the rest of your back lash.

I do encourage you to post a simple drawing if the shaft support bearing arrangement..

Wow, Madagascar...really an intriguing place eh, only in the movies ?

Adobe ( old as dirt )

[NC Cams]

I suspect you have a needle bearing, followed by a bronze bushing to suppor the end of the screw followed by another needle (if you're lucky).

Needle bearings are not usually preloaded, per se, as they are typically designed to be drop in interchangeable with conventional thrust washers. Folks sometimes try to do it with belleville washers but you can overcome the deflection curve of these quite readily.

THis is why angular contact ball bearings are used - first they can be preloaded and still roll because the balls make "point contact" with the raceway.

EDIT: try replacing the thrust bearing with a solid washer that is essentially tight. Then push shove on the table to see if you have gibb/table slop. At that point, your slop should be confined to the ball nut and/or the giggs. Once you fix slop here, THEN go to the ball screw support bearings.

YOU CAN"T ELIMINATE BACKLASH/SLOP UNTIL YOU METHODICALLY QUANTIFY ALL THE PLACES WHERE IT IS COMING FROM. Thus, FIND THE ROOT CAUSE FIRST, then develop a sollution strategy, piece by piece, section by section, part by part.

END EDIT

You also have to consider the bearing geometry. Needle bearings make line contact and also tend to want to skid because the OD of the needle is trying to roll at a different surface speed than the ID of the rollor (think about the difference in circumference at the ID vs OD of a needle bearing and it will make perfect sense). EIther the ID or the OD wants to slip if you preload them - NOT GOOD.

In light of the fact that this machine has a needle thrust beariing or bearings, I'd be inclined to speculate that it has a fixed distance/clearance at the thrust end, or if you're luckly, there is a nut that can be adjusted to "preload" the washer and take up clearance. In such cases, you can't make them tight enough without experiencing binding in some way or another - that's what I've seen in the cases where I tried to fix needle thrust/bushed ball screw ends.

Either way, you probably won't/can't come up with a low friction/high stiffness/low clearance axial thrust absorber using needle thrust bearings. Besides, you shouldn't preload them due to the slip/slide problem associated with the skid potential discussed above.

Ball screw bearings are designed that way for a reason. Folks try to get around it by all sorts of gyrations but, eventually, the endemic limitations of the use of the less than optimum bearing rears its ugly head.

You either have to live with what you have or do a major redesign to try to package something more appropriate for/to the task.

[Adobe Machine]

Thanks NC, knew there was some thing that prevented needle bearings from being good thrust. That was a good explination of slip /skid to illustrate."Tighten up" just never worked with needle thrust bearing, no matter how well they were lubricated.

Had a lot of "thrust bearings" )(needle) fail in High Performance out boards and out drives when lots of torque and horse power was applied ..argued endlessly with a domestic out board- inbord manufacture about their control of thrust..told their chief engineer he was still in the 1950's with his "new" design, he invited me out of his office, but they almost went broke fixing thousands of outdrives still in warrenty . The "other" marine manufacture solved their problem with precision matched angular thrust bearings..

Adobe (old as dirt )

[Adobe Machine]

Contacted my friend today. He is a M.E. involved in precision optical inspection equipment. Their tolerances are mind boggling ..000005 (+-).000002! Every thing is done in a controlled atmosphere room.
They use lots of small ball screw applications, all are ground and come to them at no more than .0002 back lash..He did say that in some instances the ball screws are reloaded to minimise back lash, but most accuracy is controlled by optic and other means for instant correction.
His impression to me that even a rolled 3/8 ballscrew that had .007 or more back lash was most likley not intended for any precision work.He also raised the possibility of mixing a metric ball screw with inch balls ?
In all instances they used precision angular bearings..

Adobe (old as dirt )

[NC Cams]

MOST ball bearings (even metric ones) still use inch sized balls. It seems that this was done eons ago so that load capacity equivalency could be maintained from maker to maker within each bearing size.

Yes, there are metric balls used but this is the exception rather than the rule.

If you think it is hard to find an oversized inch ball with micron size differentials, wait until you see how easy it is to find metric ones with comparable differentials.

Yes gang, ball screws are not merly ball screws. Although rolled threads are adequate for a lot of applictions, when you start looking for near zero backlash potential, you're probably going to have to spring for ground screws.

The Japanese have a phrase for this: shoganai (show- guh-nye). Translation: that's simply the way it has to be.

[Adobe Machine]

NC Cams: There were a lot of ball screws built in Europe that had metric ball screws, that in fact used metric balls, although most likley not a lot of them came to the US except in speciality equipment ( Telescopes, high end optic trackers)

Salem ball co and others keep a good inventory of metric balls.

A good rolled ball screw, specially where you can use a douple nut can be more than sufficient for the Hobbiest and small shop, tolerances less than
.0001, WITH THE CORRECT BEARING SET UP is obtainable.Just a lot of people can not afford high end ground ball screws, and most when delivered to a hobby consumer still have .002 to .003 backlash, unless you have paid "extra" for a "0" backlash ball nut re-load. My feeling is, invest in good bearings and rolled screws with double nuts..and then invest in a good set of duplex bearings to eliminate radial and axel play.

Helped a friend ( and former competitor) install a kit, including rolled ball screws with double nuts and Barden bearings mounted DB, in a BP clone that had already been CNC'd. He did have the table and ways ground and scraped , and cleaned /redid a good portion of the oiling system. After 7 weeks of 2 shift work, an accuracy check showed this machine was still at less than .0001., even in some hard corners.His total cost is $3100.00,
( not including refreshments) but he could not replace this nice of a macine for that amount.

In this thread started by CNCUSER1, there is still a possibility of a a mount bearing problem or possibly who ever retrofited the machine 10-12 yrs ago got a hold of some European ball screws that needed metric balls, or like most things, a combo of all of the above.

Adobe (old as dirt )

[NC Cams]

Adobe: your needle thrust inboard/outboard posting in #12 explains why thrust bearings in transmissions are/should be designed with spatial consideration for either needle or bushed thrust washers used interchangeably in critcal thrust postions.

This way, if they OOPS in figuring the use of one or the other, they have the option of substituting the other in case they boost up torque or power flow and don't want to redesign the case for the 0.020" thinner/thicker bearing that they need to install the RIGHT bearing.

Sadly, lots of people don't give TOTAL thought to the simple geometric limitations of bearings. Moreover, they don't even take the time to READ/UNDERSTAND what load capacity means for a bearing. Sadly, it is explained but few really take the time to learn.

It took racers quite a while to learn what happens when you put needle thrust bearings behind cam sprockets and they fail - surprisingly, some even learned that a bronze thrust washer works MUCH better there and elsewheres.

Oh well. That's why there are winners and losers. It is also why some are quite proficient at holding down their finishing positions in spite of MA$$IVE efforts to the contrary.

[Adobe Machine]

Do you mean those funny little bearings in my oil pan, that got into my High volumn pump and .....just another window in my K.B. I gotta tig up..me and that bearing engineer had a real discussion !
How are they doing with needle bearing cam bearings ? ?

Adobe ( old as dirt )

[NC Cams]

Needle cam bearings: Depends on the type.

The real early 'drop in style' like Crower tried to pioneer (rollers contained in groove cut into journal with split busing outer rings) have long since fallen by the wayside.

The true press in place caged needles that the blocks are recut for are quite common in Pro Stock drag racing and Nextel/Bush/Crafstman truck racing. I dunno about the sportsman classes as this ends up being big buck due to the trick cams you have to use in concert with them.

Only trick with needles is to properly harden and grind the cam journals (not all are done properly!!! some are real lame) and also to grind the journals to the proper size and tolerance (some are grossly improperly sized).

[cncuser1]

One night this week I'll post some pictures here of the relevant items, I just have to pull the machine apart...again.

I pulled my Y axis apart and have attached some pics.
The close up of the nut is self explanitory

The other two pictures show the screw&nut. Moving to the left of the nut:
(midway , to the left of the nut, is a piece of hardware, it is not relevant)
- Is a shoulder machined onto the screw, a hardened washer (.0625" thick)is press fit up to the shoulder.
- next to the left is another hardened washer ( .032" thick), then the bearing pictured in a previous post, then another hardened washer ( .032" thick).
- Next is a 1/2" steel plate
- next to the left is another hardened washer ( .032" thick), then the bearing pictured in a previous post, then another hardened washer ( .032" thick).
- Next is a pulley keyed to the screw shaft
then a hand keyed too the shaft
then finally a 1/4 nut clampimg everything up

So the nut clamps everything between it and the machined shoulder.

In my humble opinion, when two fingers lightly ride on the nandle OD and can turn the handle .003", the problem is not in the clamping.

I tested the assembly by holding it all down with toolmakers clamps to a 1/2" brass plate, to try to simulate the backlash, it did not appear on the bench set-up!

The only guess i can make is that the backlash comes from a springyness within the nut, nut an actual gap.

I just ordered slightly larger balls from salem

Recap: stock .0627" balls gave .007" backlash
oversize .0642" give .003" backlash
I just ordered .0660" ( but they didn't have enough for both of my lead screws, I'll cross that bridge when i come to it)

I'll report back with results

M

[Adobe Machine]

Thanks..one other point you might look at is the flange attaching the ball nut to the table pinned or set screwed ?I had one drive me to tears, just could not find the backlash, and had the nut stuffed so hard that it was getting hard to turn, finally I had one of my better machinests look at it ..took about 45 seconds and he turned to me and said "Boss, can you not see the ball nut moving on the flange ?" Not only had I missed it, but the allen set screw was lost or long gone..Drilled for a hard pin and installed the set screw with blue lock-tight..backlash gone, had to restuff the ball nut with smaller balls.

Is there a way to set up a dial indicator on the end of the shaft when assembled ?..

Keep us informed.

N.C. Cams: Just courious, will he be able to preload the arrangment above to eliminate all back lash under cutting conditions ? He is getting closer, but with out having to redo the whole system, would you have any economical suggestions ?

Adobe ( Old as dirt )