What's new
What's new

How to apply preload to ball screw thrust bearings

ranchak

Cast Iron
Joined
Dec 4, 2007
Location
Southern California
I need to replace the thrust bearing in my Mighty Viper VMC-950, they are in a triplex set arrangement, angular contact bearings. The bearing manufacturer states that the preload is 4500 N, the starting torque is .33 N-m. 4500N is 1011.64 lbs. There are no spacers between the bearings and there appears to be a only a washer for the nut to run down against. So how do I apply 1011.64 lbs of preload?

I did find this thread:

http://www.practicalmachinist.com/vb/general/calculating-bearing-preload-163624/index2.html

Where in post #22 Rich explains what I think I should be doing.

It's funny I asked the bearing manufacturer how to set the preload and their engineer said that I should apply 1011.64lbs of torque to the nut. I think someone applied to much torque to his nuts and he either can't think straight or he isn't much of an engineer.
 
most of the time the preload will be mechanically set with spacers inside the assembly. Torque to 50-150 ft. lbs ( dependent on screw size, etc.) but treat it like a lugnut on your car with similar torques.... but check it out and make sure some junior woodchuck candidate did not go in hacking stuff up messing with mechanical preload..... your thrust bearing housing should spin freely on the screw with no visible runout when tight...... it should feel about like a cnc mill spindle does......
 
I think the guy meant force not torque since "1011.64lbs of torque" is meaningless.
Kind of missing a unit here (distance).
The post you point to also seems to toss screw/nut efficiency out the window along with drag loss on a typical bearing nut.
I'll leave solving the math to you as far as torque at the nut.
Generally here you would buy preground preloads and you just want to ensure that you apply more force than the ground in preload to tighten every thing up.
Past this you can go way over and nothing happens.
Bob
 
On my Makino VMC, ballscrew pretension was measured by stretching the ballscrew. Indicators were placed on the ends of the ballscrew, zeroed, then the preload nut was tightened until the relative change of the indicator readings corresponded to the number in the manual (.0014", IIRC). On a ~16" travel ballscrew, this took a surprisingly large torque, easily 100+ ft-lbs.

As you and others noted, there's a unit problem with the numbers you've been given.

Ballscrew stretch is usually set by the ballscrew manufacturer and machine tool builder, not the bearing manufacturer. You might check with one of those companies for more information.
 
A preloaded A/C bearing is axially ground such that if you were to place the assembled bearing on a flat surface, the inner race would sit slightly higher than the outer race. Under full preload (4500N), both races would touch the flat surface.

You don't need any spacers for these types of bearings. The inner races come in contact with each other as you tighten the locknut and won't allow you to "over-preload". What's the thread diameter and pitch of the locknut?

Guessing this is for the Z-axis?
 
The ball screw in question is the Y axis, I haven't disassembled the ball screw yet. I do know that the diameter of the threads is 30mm. I believe this is the same bearing used in Fadal 4020s, maybe Bob in NM can shed some light on this. The bearing number is: NSK 30TAC62BSUC10PN7B or ********* BSB030062.
 
On my Makino VMC, ballscrew pretension was measured by stretching the ballscrew. Indicators were placed on the ends of the ballscrew, zeroed, then the preload nut was tightened until the relative change of the indicator readings corresponded to the number in the manual (.0014", IIRC). On a ~16" travel ballscrew, this took a surprisingly large torque, easily 100+ ft-lbs.

As you and others noted, there's a unit problem with the numbers you've been given.

Ballscrew stretch is usually set by the ballscrew manufacturer and machine tool builder, not the bearing manufacturer. You might check with one of those companies for more information.

This is how it's properly done from my experience.
 
I've never seen the screws mounts on this machine so pardon the ignorance.
The ball screw in question is the Y axis, ......... The bearing number is: NSK 30TAC62BSUC10PN7B or ********* BSB030062.
And it is in a triplex mount?
Most "Y's" will be dual or quad.
That seems strange, what is on the other end?
Is this a fixed-free, fixed-floating or fixed-fixed setup?
Maybe fixed-fixed and the third is to help handle the screw "stretch" loading?
Bob
 
On my Makino VMC, ballscrew pretension was measured by stretching the ballscrew. Indicators were placed on the ends of the ballscrew, zeroed, then the preload nut was tightened until the relative change of the indicator readings corresponded to the number in the manual (.0014", IIRC). On a ~16" travel ballscrew, this took a surprisingly large torque, easily 100+ ft-lbs.

As you and others noted, there's a unit problem with the numbers you've been given.

Ballscrew stretch is usually set by the ballscrew manufacturer and machine tool builder, not the bearing manufacturer. You might check with one of those companies for more information.

That seems like a shitload of preload.... My machines spec out .0004-.0005 but they aren't makinos.
e
 
Old Cincinnati VMC here, .0015" on a 50" travel, 2" diameter screw.

I can see .0015 over 50" travel. His makino was only 16" travel. My machine has 16" of X travel and they only call out .0005 max preload. Too much you will fry the bearings.
 
Yes, but when cold it will still have the same load on the bearings, won't it? this is a low speed machine, not likely to heat the screw.
 
I know on my machine when they tried to tighten the pretension to fight thermal growth the bearings went out about 3-4 months later. Machine only has 1000 ipm rapids and slow accell. A lot of weight on the ball screws though.
 
That's fast compared to this one, 400 IPM rapids. 3000 lb table capacity, not super heavy. The 26" Y has no preload, bearings at only one end in fact.
 
When its cold the screw would shrink, putting more pressure on the bearing since its pulling the nut towards the other side?
 
I've never seen the screws mounts on this machine so pardon the ignorance.

And it is in a triplex mount?
Most "Y's" will be dual or quad.
That seems strange, what is on the other end?
Is this a fixed-free, fixed-floating or fixed-fixed setup?
Maybe fixed-fixed and the third is to help handle the screw "stretch" loading?
Bob

Hi Bob,
Often, the arrangement with the X and Y axis is that the Non-drive end of the ball-screw floats in fresh air. The screw goes from the motor coupling, through a radial and thrust bearing arrangement, through the ball-nut and then nothing.

Regards,

Bill
 
If you're just concerned about preloading across a single triplex set of TAC62 ballscrew support bearings, as others have mentioned, the preload is ground into the races so you just tighten things until the races are bottomed out. Note, installation technique depends on if it's a DBD or DFD arrangement, etc. My screw, being pre-tensioned, was a DFD arrangement (preload applied across outer races), since this reduces the sensitivity of the bearing stack to misalignment of the housings. Luckily you can mix and match bearings as long as you buy SU ground bearings. This document has some helpful info: http://www.bearing.co.il/Angular Contact Ball Bearings Combination 1.pdf

That seems like a shitload of preload.... My machines spec out .0004-.0005 but they aren't makinos.
e

You're right, it is. I checked the numbers, for the Z axis it's .0008" stretch, for the X and Y axes it's .0012" stretch. I was going off memory and it's been a couple years since I did it. My bad. I think my machine is ~45mm ballscrews, but then again we've seen how reliable my memory is. It's old-school; box-ways and 600ipm rapids. Absolute stretch number is dependent on overall ballscrew length between bearings; maybe your machine has a shorter distance between bearings?

Am I correct assuming ambient temperature shouldn't matter for ballscrew preload, since the screw has the same CTE as the cast iron frame? Thus, unless the ballscrew somehow gets chilled relative to the base casting, the screw should never tension itself beyond the initial preload applied. As the screw warms up due to machine operation, this pre-tension should relax.
 
Am I correct assuming ambient temperature shouldn't matter for ballscrew preload, since the screw has the same CTE as the cast iron frame? Thus, unless the ballscrew somehow gets chilled relative to the base casting, the screw should never tension itself beyond the initial preload applied. As the screw warms up due to machine operation, this pre-tension should relax.

That is very close to what was presented in training I took at Makino in the late 80s.

The Makino maintenance manuals of that generation had a "standard" stretch value for each axis. When you bought a screw from Makino, it came with a cert that told the stretch value for that particular screw to achieve accurate pitch. If one did not stretch the screw when installing it then the positioning over the full stroke would be short by an amount close to the stretch value.
 
Hi Bob,
Often, the arrangement with the X and Y axis is that the Non-drive end of the ball-screw floats in fresh air. The screw goes from the motor coupling, through a radial and thrust bearing arrangement, through the ball-nut and then nothing.

Regards,

Bill

I have triplex bearings at both ends of my ball screw, the end does not "float".
 








 
Back
Top