Having just finished machining a pair of Thompson ball screw ends to accept bearings and a lock nut, I was wondering if thier was a better way. The screws are case hardened to I think around 65 rockwell. Does anyone know how to anneal or soften the ends of the screw to make it easier to machine?

I get a rag really wet and wrap it around the part that doesn't get annealed. This helps to not soften the part that doesn't get machined. Then I take my acetylene torch and get the part to be annealed bright red, hold of a few seconds and then let cool slowly.

HIH

Steve

Where I used to work, we made up ballscrews all the time. ( Machine tool rebuilder ).

Our method was to cut the screw stock off to the length of thread required. We would then bore a hole in each end, (the centers are soft).

We made up the bearing journal/drive coupling ends, as a separate pieces, with a tongue that would go in the bored hole, as tight a fit as could be pushed together. We pressed the end-pieces in, and MIG welded them around the joint.

Usually the end pieces were cut with .025 or so stock left on, so we could finish machine the ends after welding to get it all concentric.

Phil

I just snag off the casehardened outer layer with an angle grinderif I don't have many to do. It is of course softer underneath.

I would strongly suggest grinding the case off rather than taking a chance of warping the shaft. If this doesn't fit your envisioned modification, then I think the info in post #2 is good information for softening it up. Rotate it frequently as it cools to keep the heat dissapation even.

Scott

I take a slightly different approach. Being mean, I buy only the length of ballscrew needed for thread :-) I face the ends (hold the bs in an purpose-made alloy split collet in the 3-jaw) with a carbide tipped tool. Didn't need to soften or grind anything. Then center, drill, bore, ream to take an adapter. I used 8mm reamer for a 15mm nominal BS. At bottom of reamed hole I have about 10 mm of 6 by 1mm thread. Adapter has 8 mm portion and 6 by 1mm thread used to draw adapter into bs. Final fix is with Loctite (tm). Just finished the x-axis for a Senior mill using this approach. I used 500mm of bs for a leadscrew with total length of over 850mm, including portion to couple to stepper motor.

Keep the comments coming, I like to hear what everyone else is doing. The set I turned I used a rubber flex collet to hold the screw. I faced the end with a carbide insert and centerdrilled it. I put a live center in and turned it down. I t was tought to turn though. I used 2, 3 sided inserts up turning and had to adjust the tail stock to remove the taper. in the turned shoulder caused be it springing away from the cutter. In the end they came out good but it was a lot of work.

For small stuff, yateleybob's method probably works much better than welding.

Yeah annealing, inserts, grinding off case and even hard machining all work.
The really interesting part is workholding. On some screws, particularly rolled ones,
the outer lands have a burr from the rolling process. They are not accurate datums for the raceways, so holding in a chuck or collet can create big errors. If possible, I check the lands with an indicator on the carriage with it set for threading at the ballscrew pitch. Then shim and adjust as needed. Someone else might have a better way...

Yeah annealing, inserts, grinding off case and even hard machining all work.
The really interesting part is workholding. On some screws, particularly rolled ones,
the outer lands have a burr from the rolling process. They are not accurate datums for the raceways, so holding in a chuck or collet can create big errors. If possible, I check the lands with an indicator on the carriage with it set for threading at the ballscrew pitch. Then shim and adjust as needed. Someone else might have a better way...

[I used some heavy aluminum wire, wrapped in the track, to provide a consistent surface for gripping in the lathe. It fell nicely into the smooth part of the screw, and seemed to hold it on center. I also used Les' method of removing the case-hardened layer on the end before I did that; using a large grinding wheel and revolving the part, it was possible to keep it more or less round. The bottoms of the lands give a good indication of where it's at. I made some gizmos for holding a couple of ball-nuts together with wave-washers in between, so as to eliminate backlash. It seemed to work, although if screwed too tightly together they add too much resistance.]

Andrew Werby
www.computersculpture.com

Andrew's technique looks good (thanks!) and I'll "borrow" it for the Senior Y axis. The rolled screws I've laid hands on (German, via a UK company) seem to be burr free, judging by clocking the raceways with it held in the collet. BTW the cylindrical non-threaded nuts supplied by this firm (they are ISEL nuts) have a backlash adjustment system built in. All it needs is a 6mm or similar grub screw.

Wow, lots of good methods were presented!

For me, using a tool post grinder in the lathe works well.
Must go slow (grinding) but once thru the case I had
nearly the finished diameter I needed for the bearings.

For a few screws it worked ok and only required 1 setup.

Pres