Machining the ballscrew and setting up the 4th axis to machine hex flats

Need to machine down one end of the SFU3205 (32mm dia, 5mm pitch, single ballnut) ballscrew to fit the thrust bearing (25mm through bore) and put some flats on it so I can drive it with a 17mm socket and cordless screwdriver or impact driver. Firstly, the bearing fit and the major diameter for the hex.

Got the ballscrew set up in the Tree CNC lathe, which has a 2" through bore in the spindle. This is hardened Chinesium steel, so could be hard on the insert. Using a Korloy trigon insert that should be reasonably robust. Furthermore, it's double sided, so I get 6 tips per insert and even if I nadger a few of them, it's no big deal. These inserts only cost me ~£3 each - 50p a corner.

Off we go. 

Ooof. I was a bit ambitious / aggressive with the speeds and feeds (1mm DOC, 0.1mm/rev and 100m/min). Although these values accord with the manufacturer's recommendation, they probably didn't anticipate dealing with hardened steel with interrupted cuts. Either way, the tip fractured, the axial cutting force went up and the stock slipped back in the chuck. The chuck key is cleverly designed to make it difficult to use an extension to overtighten - and it was about as tight as it would go. It wasn't a matter of not tightening it up properly.

Reduced to 0.25mm DOC, 0.05mm/rev and the same 100m/min constant surface speed:

That worked out better:

And a manually machined a 2mm groove for a circlip:

I managed to not cock up the position of the groove:

Machining hex flats on the ballscrew:

I recently upgraded the CNC12 software used on the Tree lathe but it seems I didn't bother to set up the 4th axis. NVM - I still had the previous installations to refer back to and was able to read off the appropriate parameters. There's actually a blog page from back in July 2020 that shows the correct values when I was setting up the Yaskawa Sigma servo drive that drives the 4th axis. It uses a belt reduction and a harmonic drive, so the overall reduction ratio is a calculated value built in to the servo drive setup.

And yes, the 4th axis has a sufficiently large diameter through bore to take the 32mm ballscrew. I'll just machine those flats manually, rather than bugger about generating toolpaths in Fusion. Currently I have a 19.3mm diameter boss end (fnaaaar!) and require 17mm a/f, ie touch off and descend 1.15mm, then rotate 60 degrees, repeat 5 more times etc. 

The circumscribed circle of a regular hexagon with a 17mm inscribed diameter is approximately 19.63mm or exactly 34 / SQRT(3). I've deliberately reduced that OD by 0.33mm so that I don't have sharp edges - sockets don't have them.

50mm face mill with APMT inserts:

Looks good. 2 passes, total 1.15mm stepdown:

Socket is a good loose fit:

Flats are 17mm a/f as good as dammit:

It's still pretty hard, even several mm down. It's obvs through hardened, not just surface hardened. Took a bit of work with a fine file to debur the edges but it's looking good now.

All done and back in its bag for now. 

I need to machine up the thrust brackets for this bearing and the ballnut now.