Arrow 1000 - Extreme Backlash

[t_hum]

We have an '98 Arrow 1000 with an A2100 control that has been giving us fits in producing accurate parts. Some parts will drift out of tolerance by up to .030" in the X axis and .002" in the Y. The Z also is out .005". This machine is new to us but the previous owner was not too terribly concerned with regular maintenance it appears. When aligning the machine after each startup we have been testing the machines zero by indicating to the table. The 0 drifts on average .010" in the X. The Y and Z also drift but are usually below .001. Also, when the machine stops after rapid moves in the X or Y we are usually greeted by a loud bang Ive been assuming is the table reaching the end of the backlash.

To diagnose the source of the issues we performed two procedures:

1. Indicated from the spindle to the table using a Haimer 3d sensor. We then pushed and pulled on the table using a 2x4 in both the x and y. Using a moderate amount of force the X indicated +.0085 in one direction and -.0025 in the other. The Y moved +.0005 in one direction and -.002 in the other. Z moved +.005.

2. We then indicated from the machine base to the end of the ballscrews as well as from the base to the table. The screws and table move the same distances noted above. The screws will rattle when moved by hand radially.

We assume then that the thrust bearings are shot on this machine and need to be replaced. We have 2 new complete sets for the X axis.

So our questions are these:
Are we correct in the assumption that the thrust bearings are the culprit?
Are they difficult to replace and do we require the service manual to do this repair?
How much preload is needed on the thrust bearings?
Any other areas we should take a look at for possible sources of backlash?

Your help is really appreciated. This issue has been driving us nuts and it makes making parts nearly impossible with any sort of accuracy.

Thanks.

[Rick Kight]

t_hum
As far as the x axis goes

Are we correct in the assumption that the thrust bearings are the culprit? yes most likely as they are not sealed and coolant / dirt gets into them. We have added a grease fittind in the bearing block to help keep stuff out of them
Are they difficult to replace and do we require the service manual to do this repair? easy to replace manual not needed
How much preload is needed on the thrust bearings?They are a matched set you install faceto face or back to back preload is non adjustable just install and tighten the nut
Any other areas we should take a look at for possible sources of backlash? Thrust brgs are the most likely we have had issues with our ball screws/ nuts BUT we run very abasive material and our machines are getting old, we have had a few of ours rebuilt

The other two axis are better protected and we have had little to no issues with them

Rick

[slowlearner]

I have had a coupling work loose before but I'm sure this was checked first.

[t_hum]

I did take a quick look at the coupling. Everything seemed fine at first glance. I was between part runs and didn't have a ton of time so Ill have to take a close look before we replace the bearings.

I was under the impression that if the coupler is loose that can lead to obviously accuracy issues with the encoder giving a different reading than the table is at. However the end play on the ballscrew makes me think that the main issue is still the thrust bearings. Maybe both the coupler and the bearings are contributing. Thanks for the tip.

Has anyone had the preload back off over time? Im wondering if the Y and Z, which arent nearly as far off as the X, could have lost some of their preload.

[t_hum]

An update for those interested...

Both sets of bearings on the ballscrew were completely shot. Both balls and cups were severely pitted and generally loose.

On top of that the coupler on the X axis had lost its pin between the V groove and set screw at some point. The coupler has obviously rocked back and forth for some time. The coupler set screws are now approximately 30 degrees off from each other as the coupler tube has been twisted from either a bad crash or just repetitive motion. A new pin and coupler tube have been ordered.

[mike_Kilroy]

did u consider fixing coupler issue by simply buying a compression squeeze coupling from someone like r+w or gerwah instead and ignore the v groove?

[t_hum]

I hadn't considered a different style. I assumed the v groove would be necessary to transfer the high torques. Is there a specific coupler you have experience with?

[mike_Kilroy]

We have experience with both R+W and Gerwah (now called only Ringfedder). Both have bellows style compression couplings.

We have not done so yet ourselves, but we are about to replace one of the obsolete Vickers motors on one of these machine's Z axis and supply maybe a

BKL/80/19/(19?) this is rated 80nm for around a 12nm rated (24 max) rated motor with a 19mm shaft with the v groove. I put (19?) as I do not have ballscrew diameter handy. price on such a coupling is about $ 200.00

specs here: http://www.rw-america.com/bellows_co...ling-bkl-t.php

Of course we sell these feel free to PM us!

[mike_Kilroy]

.... I assumed the v groove would be necessary to transfer the high torques....

I just reread this and thought it may be worth a comment to see what others may come back with...

I am an electrical engineer with mechanical servo experience. I have been told over and over and over again by mech wizards that keys (v groove included) do NOT help transfer torque; they just line stuff up. it has been beat into me that to use a key to transfer torque is a no 1 no no; keys used to transfer torque will simply wear out and become bigger and no longer transfer torque. It has been beat into me that shafts transfer torque around their paremeter; a key is just a small sharp edge when u get down to it, and so cannot really transfer any torque - more than once or twice.

have i been taught by the mechanical wizards wrong or do keys REALLY transfer torque?

[t_hum]

Ive always used keys in applications that require high torque transfer. Spindles use a drive dog for a reason I guess. But I do understand there are pros and cons to each type of fitting.

This link does a better job than I ever could...
Sorting out shaft connections | Machine Design

[scudzuki]

I just reread this and thought it may be worth a comment to see what others may come back with...

I am an electrical engineer with mechanical servo experience. I have been told over and over and over again by mech wizards that keys (v groove included) do NOT help transfer torque; they just line stuff up. it has been beat into me that to use a key to transfer torque is a no 1 no no; keys used to transfer torque will simply wear out and become bigger and no longer transfer torque. It has been beat into me that shafts transfer torque around their paremeter; a key is just a small sharp edge when u get down to it, and so cannot really transfer any torque - more than once or twice.

have i been taught by the mechanical wizards wrong or do keys REALLY transfer torque?

Well that is an interesting perspective.
Splines are certainly better than a drive key but I can think of plenty of applications that use square or woodruff keys to transmit rotational forces.

The first one I can think of is my first go kart. The centrifugal clutch and the 1" horizontal output shaft of the 8 HP Kohler engine were aligned by a 1/4" square key and the clutch was held on by a 3/8 bolt at the end. There was a setscrew over the key to prevent it from falling out, but the key was the only element transferring torque. That key and the mating keyways held up pretty well and I rode the crap out of that thing. That's just one of many examples I could illustrate.

Joe

[mike_Kilroy]

ya, i think that keys have been used Incorrectly or cuz they didnt have the choices have today?) for many many years to transfer torque.... but ALL the mech design engineers i know at cincinnati milacron and other machine tool companies & GM today insist they a wrong way to do it when there are squeeze type compression couplings available today. all clearly state that a key used to transfer torque will eventually wear and leave backlash and ultimately fail. all state that compression couplings will never fail is tight and sized correctly. T_hums machine design article pretty much confirms this too. the only times we put keys in our good servo couplings we sell today is for either 1) an old fashion engineer/customer who does not realize that keys is wrong way to transfer torque, or 2) someone who needs to lign shafts up for like homing. just what i hear from those who say they know....