Engineering principle behind this type connection

[Jimsehr]

I changed the tool a smaller dia because I thought 1/4 dia was too large for a 3/8 dia ball. Too much wasted material. Think 1/8 dia tool would be about right.
I don’t know how far the ball shape needs to be. I would need to see a finished part . I also think a 5c collet that extends out of the spindle farther would help on making these. I forget what Hardinge calls that type collet. Or a Hardinge extended length emergency collet will do the same. You can get these collets
from Hardinge.
Just making a one off takes about 3 or 4 passes with the tool and takes about half a minute each.

 

[jwmelvin]

I still don’t quite get your wobble. With no screw and no load, am I right there is no wobble? The axial interface is the outer rim of the rotating piece, pressing down against the stationary piece? And the radial interface is the outer circumference of the rotating piece, or a central bushing?

Maybe rethink the design so the load is carried directly through the interface. That is, before loaded, the two parts are touching. So then load does not cause a change in clearance but instead increases the pressure at the already-contacting interface.

Sketch a cross section of your design. It doesn’t sound optimal.

 

[Jimsehr]

I did not have any chatter problems at all. I did not get a chance to try a form too. But that might be quicker.
How close do size of parts need to be ?
I still don’t see any need for a bearing.
Send me a pic of what finished parts look like and tolerance needed.

Jimsehr

I checked Hardinge long nose 5c collets and see their long nose ones cost $151 ea but you can bore a 1 inch extended length emergency collet for $ 64 dollars.
But you have to drill and bore them to size.

One other thing once chatter starts it is harder to control. May be nit picking but I would move the center of the ball turner closer to the center of the lathe cross slide.
Plus I would not bore the base to hold the rotating disk cutting tool. I would just put the rotating disk flat on the flat base. It would be easy to set 2 radius dia locating holes on one disk. And it would have less chance of chatter by not having a sliding tool to adjust for different size radius. Also for production I would change the thickness design of the rotating disk. I think it would be able to cut a thousand balls before needing to change rotating disk.

Jimsehr