Making a long threaded screw

Steve, It's been about a quarter century since I've done this (CNC spoiled) but I'll give it my best shot. First off set your compound slide at 30 degrees and square up your tool and cut the first section of thread. In all likelihood your lathe has a threading dial on it and you engage the feed at the same number at every pass. If so, then what you need to do after you rechuck the part is to engage the feed on the same number, feed up close to a cut thread and shut the spindle off leaving the feed engaged. If your lathe doesn't have a brake be careful as the carriage will continue to feed as long as the spindle is rotating. Now use a combination of cross slide travel and compound slide travel to nest the tool into the previously cut thread. Note the numbers on your compound feed dial as they will be different than your first thread section. Depending on how fussy you want to be you can use an eye loupe or some Dykem, and rotate the spindle slightly by hand to make sure the tool is lined up perfectly with the previously cut section. Once you are satisfied that the tool is lined up with the thread, back the tool out, disengage the feed, back the carriage up, start the spindle and begin threading as normal by picking up the same number on the threading dial.

HTH Dan

Given the proper setup, it's possible, but would be a serious challenge, especially if you are new at machining. You'd need a follow rest, possibly a steady rest, be able to feed the material through the headstock, and be able to pick up threads properly. Having considerable skill and experience would be the most valuable asset, along with soft jaws. Depending on how you pursued the project, you might be limited to length, but you could cut one longer than the center length of the machine.

Harold

I should have mentioned that the end product would likely be of questionable quality.

Harold

Hey! Not true, at all!

It's something of a chore to re-index a piece of threading work, but it's not that bad. Takes patience and time, that's all.

Consider: Mark a thin scribe line down the length of the work to show the rotational indexing position. Position your tool in the thread very near the chuck, and lock the half-nuts. Then unchuck, and re-index the work in rotation by the scribe, and longitudinally by leaving a small amount of the already-threaded portion remaining beyond the chuck jaws, and using the (never moved) bit to find the correct longitudinal position in the chuck.

Keep lost motion in mind, and snug the tool to the "cutting" face of the thread (by moving the work to the tool) before each re-chuck.

LLoyd

LOL. The real long answer is, the task will be so time-consuming, and the outcome so uncertain, that you are better off doing it another way. Borrow a larger lathe. Purchase the lead screw and fit to your machine. etc.

Short answer though is correct. Umm, not really. No. Also you did not specify any kind of lead accuracy tolerance on your requirement. A lot depends on that. If the tolerance is "it's gotta be the right number of tpi," then the answer could be yes.

If you need ten microinch lead tolerance over the entire thing, then definitely no.

The truth lies somewhere in between.

Jim

Yes, it is... but it does require a lot of skill, care, and experience. In lieu of a thread mill of sufficient size (and I've had to do this with a thread mill on rare occasion)... it may be your only option.

Can you use the lathe to turn the stock, and use a die to cut your threads, or is this for a thread you don't have a die for? Novice question on my own part as well.

Well, that gives me two more questions for the group.

One - Assuming some experience on the part of the operator in cutting threads and a good lathe, how much less accurate is using a die than cutting with the lathe?

Two - How could one measure the thread to thread distance accurately to see how much error was created when reindexing after moving? Looks like thread micrometers only measure the diameter.

Steve.

A die won't give a precision lead to the part. Because this is a leadscrew for the lathe, that's not the best way to make the part.

Leadscrews not only have to have the correct thread geometry, but they are effectively length standards in and of themselves.

A die-cut leadscrew would work, but it would impart all the lead error into whatever parts were then fabricated when using it on the lathe.

Jim

Make a rig with two nuts, possibly (probably?) each made with some way of getting rid of play/backlash. Fix them so they don't rotate but allow some motion along the length of the screw. Start with them both on one side of your "join", measure the distance between (marks on) them, then move them so one is past the join and measure again.

It would depend on the quality of the die, compared to the accuracy of re-aligning the stock for the later cut(s). I'd like to think that a good die is going to give you the quality you'd need?

I'm interested in the answers you'll get to this. What kind of accuracy are you looking for?

Ah, I missed the part about it being a lead screw for a lathe.

So...how does one cut a clean leadscrew when one doesn't have a clean leadscrew to start from? Someone made the first one...

If I can see or feel the break in the thread where the stock was moved, there really isn't any point in measuring it. A measurement would need to detect .01 or less.

Steve.

Assuming you don't have every precision measuring tool known to man... use what you've got of known (or best) accuracy... in this case, the lead screw and a dial indicator.

Since a thread is nothing more than a v-groove, it shouldn't be all that difficult to devise a measuring tool that uses a small round feeler ( flat sides and round bottom would be easier to use with the indicator) riding in the cut groove. You are looking for horizontal displacement as you ride from "old" cut to "new" cut. With patience and a ten-thousandths indicator, you are probably going to get closer in that short distance than the lead screw is in a foot.....

Screw threads have been fabricated for a long time, using manual methods like chasing. The real trick to to create the metrology needed to check them, and refine them. Basically you are talking jo blocks and maudsley.

Jim

I would guess (note the word "guess") that the better way would be to turn a couple of threads onto the end of the stock and then stick it into a precision nut which is the work holder for a grinding machine and grind the thread. The grinding wheel is always going to be a constant distance from the nut so just apply constant pressure to continuously turn the stock and work it into the nut and a screw of unlimited length could be done accurately. Maybe I can make an attachment for my lathe!

Steve.

So if I clamp a nut to my steel surface plate and then mount the dial indicator nearby so it rides in the thread groove I should be able to rotate the screw with consistent pressure against the nut to remove backlash and detect variations. Cool.

Steve.

Well, yeah, but the problem it that is "after the fact." OK for inspection, but not setup. ... and bear in mind you are going to have some "averaging" going on over the number of threads that engage in the nut.... maybe not the best approach... but that *is,* roughly, the concept.