Ballscrew end bearing advice needed

[picclock]

I'm doing a mill conversion to CNC, using 20mm ballscrews.

The original acme screws have dual thrust ball races. AFAIK, most people fitting ballscrews opt for deep groove radial bearings with the outer bearing case spaced and the inner bearing centre under a compression load to 'eliminate' backlash.

Is there any reason why I should not use the original thrust races (XSY 51201) ? Radial support needed is minimal and the axial loading is far superior with thrust races.

If i use the deep groove solution do I need any radial support at all ? Just clamping the bearing centres, outers spaced, with a defined preload of around 25kg will allow the bearings to find there own centre (self aligning ??). Not sure how to measure or set the preload force value - perhaps by compressing a spring to a length known at 25kg, and measuring the torque needed on the ballscrew shaft nut.   Also bearing specs I have read do not specify static axial loads.

Motor - ballscrew combination thrust produced is 157kg (calculated from http://www.cncroutersource.com/linear-force-from-torque-calculator.html )

Intended bearings are 6002RS, with a rated radial load of 585kg, and calculated axial load 292kg, derived from 50% of radial load as per SKF data.

Any assistance much appreciated.

Best Regards

picclock

[BillTodd]

Typically, the larger cnc mills operate with the screw under a slight tension (literally stretched between multiple sets of angular contact bearings).

I dont see anything wrong with using the standard thrust bearings in a pair at one end to take the fore and aft thrust. The other end could be held in slight tension by another trust bearing or  just supported by a ball race.

The problem I found , was keeping the screw straight (a particular problem using the roller nuts as it tended to force the two nut apart). The table end bearing holders were not especially flat or parallel and tended to bend the screw as the preload was applied to the bearing set (I drilled and tapped the end of the screw to be able to load the thrust pair). I found I have to open up the mounting holes the get some giggle room for alignment.

Bill

[picclock]

@BillTodd
I think you have confirmed my thoughts on the matter. For the X axis I was just going to support the free end with a single radial bearing, rather than tension the screw. Not sure if tensioning offers any real advantages. For Y and Z I will leave the free end unsupported.

I had sort of planned to have a manual handles, disconnected with spring loaded dog clutches,  for the axes as a just in case feature, but the reality is that I have a pendant and precision control so I can't forsee any need to. Possibly I might include a fixed end nut which I can drive with my battery drill and a socket for inevitable power down maintenance. 

>> I found I have to open up the mounting holes the get some giggle room for alignment.

Planning to leave the bearing outer recess oversize and let it find its own position. The inners will be aligned by the screw shaft, and as they will be clamped with 20+kgs of axial force I can't imagine the outer bearing ring/inner thrust race will be too likely to move.

Thanks for your advice

Best Regards

picclock

[awemawson]

Don't forget when driving using the pendant you don't get the 'feel' that you get with hand wheels. I.e. There is no feedback to your brain if the cut is a bit heavy.

This is especially true with large servo driven axis and high HP spindles where the system is quite capable of shoving your nice new shiny lathe tool or milling cutter deep into the work!