I'll be turning a piece of low carbon steel rod, nominal 7mm (0.2756") diameter, 0.003" per foot straightness tolerance. The stick out from the ER25 lathe chuck will be 2-7/8" which includes enough to part off the finished piece. Even though material removal is minimal, 0.003" off the diameter, I'm pretty sure I'll need to employ a tailstock live or dead center. I have very little experience with live or dead centers even though I own both
.
I know I'll need to spot drill the rod to accommodate the center, should I do that with the rod stick out set to 2-7/8", or chuck it very close, spot drill, then re-chuck to 2-7/8"?
And, live or dead center?
Thank you.
My first thought, reading this post, is that if you're asking these questions, you NEED some additional test material, as you're gonna scrap some parts Cutting these test parts will be necessary. It would be easier to work down if you were cutting these out of 5/16 or 3/8 rod, where you had some room to take multiple cuts. To hit these with that small of a cut, first, you're going to validate and quantify the lathe setup, the tool setup, the cutting tool it's self, and generally EVERYTHING. Then you'll start making "real" parts. Some of that can be whittled down in time, as you get to know that lathe, but the only way to know how the lathe acts in a given operation is experience and repetition. Taking three thousandths off of the diameter of something is not trivial, as it means you've got one shot at a cut that's one and a half thousandths deep. This eliminates most of the practical ways to "land on a dimension", as the first time you "touch off" on the work, you might well already be over your depth of cut. You've got to nail it on the first go.
This test piece includes testing your centers. If you've got a live center (with bearings), it WILL have some quantifiable runout. If it's low enough to hit your tolerance, great. They're fast, simple, and easy. If that live center runs out too much, that's when the dead center (non-rotating) comes in. Because they don't turn, they have ZERO runout. But they themselves have a learning curve. You've got to have a good high pressure center lube, you've got to get it just tight enough, otherwise it's not doing it's job, but NOT too tight, because it'll overheat and burn up in no time...
Drill the center with the piece pretty much fully inside of the collet. Then for the cut, bring it out to dimension plus room to work, as you said. Find "zero", dial in the tool to the necessary cut, WRITE DOWN your two dial readings (cross slide and compound), and cut a piece. (Don't adjust the dials, you'll loose more accuracy than what you're trying to attain, save that for production work. Use the numbers that are there). Measure the diameter at each end. Does the difference between the two ends meet tolerance? (That is the DIFFERENCE, the taper TOLERANCE Don't bugger with diameter yet, as adjusting taper WILL affect the diameter as well... Only taper at first.). If it is in tolerance, then you're good. If no, then on something that small, you're "probably" looking at simply shifting the tailstock to match the job, even if it gets shifted back where it was after to match your "normal" work. Make the correction, then cut another one. Once that's dialed in over the required length, do the same, except check the ends AND the middle of the part. You might be OK with that, but if your tool is not cutting freely at that small depth of cut, you might also find a "belly" in the part, where the center portion is fatter than either of the two iends. IF there's a belly, that WILL be part deflection, and is going to require a more freely cutting tool bit (or a follow rest) to correct. And honestly, it sounds like it might be getting a little crowed in there for a follow rest on such a small part.
Then, and ONLY then, consider the diameter. (DO NOT change multiple variables at one time, unless they're WAY off the map.... You'll chase the lathe in circles all around the shop.). So, once that little part is cutting straight (within tolerance of straight), THEN, look at the diameter. Take your written numbers, and dial in or out from where you were, and cut a diameter. Measure that diameter, move in or out from there until you can NAIL your final part in one pass. Pay attention to the dials. Make sure the compound is where it's supposed to be (perhaps "preset" that so it's dead nutz on a number so that you can see that it has not been disturbed). On the cross slide, pay CLOSE attention. One thousandth per graduation is probably a coarse resolution, you're more'n likely going to be splitting marks at least in half, if not quarters to hit the sweet spot.
IT sounds like a bunch, but really, after the first time or two, it goes pretty quick. Just work systematically through it.
Pro tip- If you're doing "test parts" to validate a setup, keep a sharpie or paint pen nearby. You MAY get your useful parts out of the test material, but every one that falls outside of dimension will be visually indistinguishable from those that do fall in spec. MARK the out of spec ones clearly. There's nothing quite like handing somebody a bad part, while the good one is sitting right on the workbench....
.