Angular Contact Bearings - any experts out there?

[slow-poke]

I'm on a quest to get the backlash on my machine under 0.001" or as close as reasonably possible.

This is an older 8x30 knee mill with new Chinese rolled 1605DFU ballscrews and double nut.

With the ballscrew/double nut out of the machine and in my vise, I could feel some backlash between the two nuts, I was able to squeeze a 0.002" feeler gauge between the nuts. By adding a single 0.001" shim between the nuts the backlash is now about 0.001" so I think I have the ballscrew nut about as good as I can get it (it is what it is). That leaves the bearings......

With the original 6001Z bearings in the BK12 blocks and everything assembled, actual backlash is 0.004", so about 0.003" axial movement in the bearings. I ordered a random set of 7001AC bearings. These are not a matched set. so as I understand it I will need to shim them to get close to zero. I measured all five bearings and in all cases the inner race is wider than the outer race by between 0.0006" and 0.0022" Seems to me like it would make more sense to ship these with narrower inner flanges so tightening the retaining nut would allow the end user to torque these to zero or near zero backlash. Does that make sense?

So at this point, with what I have, I think my only option is to add a shim between the outer rings?

If I try that I'm thinking of using say a 0.005" shim, and then in theory I can take out the remaining 0.001" or 0.002" by torquing the retaining nut, does that make sense?

Any wild guesses on how much torque will be required on the retaining nut to take up the remaining slack?

I realize ordering a matched set of ACB's might be a more sensible approach, but I would like to give the shim approach a try.

Comments, suggestions, expertise is most welcome as this is all new to me.

 

[Cadillac]

Acouple questions you had first set you order a set of 7001ac bearings then later you said you measured all five bearings and the inner race was .0006 proud. Are we dealing with two angular contact bearings or sealed bearings stacked?

If angular contact the outter race is preload to get the desired play. Usually there are two spacers one for inner and one for outter. In a Bridgeport they are exactly the same. My guess on torque is it would be a trial and error. Use the proper torque for the nut size with your desired shim thickness and compensate accordingly.

 

[slow-poke]

Acouple questions you had first set you order a set of 7001ac bearings then later you said you measured all five bearings and the inner race was .0006 proud. Are we dealing with two angular contact bearings or sealed bearings stacked?

If angular contact the outter race is preload to get the desired play. Usually there are two spacers one for inner and one for outter. In a Bridgeport they are exactly the same. My guess on torque is it would be a trial and error. Use the proper torque for the nut size with your desired shim thickness and compensate accordingly.

These are random 7001AC (angular contact bearings), my understanding is that when you order a matched pair, they grind them so that when torqued together the inner and outer widths become matched when the axial displacement approaches zero.

The way I see it I could torque these like crazy (via the retaining nut that only applies pressure to the inner race) and the outer races would still have up to 0.004" slop. By adding the shim to just the outer race, in theory I should then be able to tighten the retaining nut until the two inner rings just touch and perhaps a bit more to provide a little thrust.

 

[Cadillac]

Yeah you’re on the right track. A surface grinder would be of help. I wouldnt overtorque to much.

 

[David2011]

Are you making the mill into a CNC machine? If not, that tiny amount of backlash is nothing. I set the dials on my mill to be zeroed when pushing and just work from the appropriate end of the bed.

 

[Ischgl99]

Are you testing the end play with the bearing outer rings clamped together in the housing, or free on the bench?

 

[slow-poke]

Are you making the mill into a CNC machine? If not, that tiny amount of backlash is nothing. I set the dials on my mill to be zeroed when pushing and just work from the appropriate end of the bed.

CNC conversion, I'm using LinuxCNC closed loop with linear scales, and with 0.004" of backlash, the servos hunt at the destination oscillating by the amount of backlash.

 

[slow-poke]

Are you testing the end play with the bearing outer rings clamped together in the housing, or free on the bench?

End play measured with everything installed on the machine, so the outer rings are now clamped together with a 5mil outer ring shim, perhaps I need a thicker shim, or perhaps more torque on the retaining nut, I'm still getting about 0.003" of axial displacement.

 

[Huub Buis]

Contact angle bearings require an aligned setup or they don't last long. So the bushings must be absolutely parallel (even thickness). I don't have a surface grinder so I make the bushings on the lathe.

I make turn the outer bushing, face the side and part it off. Then I do the same with the inner bushing. Then a face a bar and glue these bushings on the faced surface (use the faced side of the bushing) using super glue. During gluing, I press the bushings against the faced surface using the tail stock with a rubber sheet between the bushing and tail stock. Then I face the 2 bushings so they are of equal thickness and parallel.
I mount the bearings with the bushings and measure the play or preload (no smooth turning of the shaft). Now I repeat the hole process but after facing both bushings to equal thickness, I remove 1 bushing (tap with a soft hammer) and do a shallow cut on the other bushing. Now the bushings differ in thickness and that should reduce the play or preload. It takes a few trials to get a slight preload.

Maybe you can make the preload adjustable the same way it is done on lathes and milling machines using a double or slitted (preferable) nut.

 

[Ischgl99]

End play measured with everything installed on the machine, so the outer rings are now clamped together with a 5mil outer ring shim, perhaps I need a thicker shim, or perhaps more torque on the retaining nut, I'm still getting about 0.003" of axial displacement.

0.0006 to 0.0022” width variance difference between the inner ring and outer ring is really high, that should be much tighter. What is the full part number of the bearings? Are the bearings mounted face to face, or back to back? If these are universal ground bearings, there should be no preload or end play, it seems like there might be two different grinds getting mixed together.