A couple of years ago I bought a South Bend 9” C model lathe with the hopes of being able to manufacture small parts for my vintage motorcycles. I was working full time at the time so didn’t have a lot of opportunity to get it right. Now I’m fully retired and trying to get the best out of my lathe.
A good solid face plant directly into the learning curve is kind of to be expected. It's initially quite steep. There's seven thousand things you need to know before you ever get started, and no way to remember it all before you have the chance to actually do anything to actually have a feel or sense of what's going on. They kind of compound on themselves. Your limited past experience may be very valuable, but it might be a bit of a double edged sword too. Smaller lathes do tend to need a little more "coddling" than commercial ones will. You probably "know" that, but don't loose sight of that when you get frustrated. It'll come.
Ive tried various materials, various tool types, I’ve spent hours making sure tool height are dead nuts on center, I’ve cut dry and with cutting fluid, I’ve taken large depth cuts and shallow too. Still not getting there.
I believe the feed rate it too high so today I’m going to change gearing to get it as slow as possible.
This is going to be huge for you. If you're changing change gears to change the feed, DOCUMENT IT. Feed rate, Tool, Material, and if you want, speed. You'll get a feel for it and give up the list in short order, but starting off? There's a LOT of combinations to remember.... Write down everything you find that works. One thing about low horsepower lathes is that you very often want to compromise one thing to benefit another. You might want a deeper cut at a stupid slow feed, or you might prefer the same material removal rate with a shallower cut and a faster feed rate.
Just for reference, (similar lathe, my ground tools, some "import" carbide stuff- When I stop buggering with the shifters, the lathe almost always ends up at the left side of the gear box, and I "toggle" between A and B which is 0.0027 per rev and 0.0053 per rev. Those two are very versatile for me. Until I get into tougher stuff that I probably shouldn't. I get hydraulic cylinder rods pretty easy. Fancy(er) steel. Those seem to typically appreciate a slower feed rate and a deeper depth of cut, I tend to end up between 0.0016 per rev and either of the two spots marked 0.0019 per rev. (And I'm sure that the few ten millionth's of rounding error makes all the difference). Again, that's my tools doing the cutting, but maybe that helps you get a ballpark as you get going.
I don’t want to be the bad workman blames his tools guy but I’ve thought of maybe buying a new PM lathe and start from a known quantity. I’ve read lots on these forums the debates on new import vs used equipment which always comes with the caution to “make sure you have someone who knows what they’re doing evaluate any used Machines” So it occurred to me my SB may be worn out or seriously out of adjustment.
Well, your lathe could be out of adjustment, or out of whack, but break it down. The spindle should be tight, kinda sorta. An indicator on the spindle nose and a bar through it, you should be able to lift the chuck end of the spindle a thousandth or two (One is the minimum, it's plain bearing. You've gotta leave room for the hydrodynamic wedge to build up and center it when it's running). You don't want it "too" sloppy, but although it might show in the work some, you should do pretty OK with a bit more. The cross slide and the compound dovetails should be brought up just so. The compound should be out of the way, and for "normal" cutting it should be backed up so that it's fully supported at the business end, don't leave it (and thereby the tool post) hanging out like a diving board without a reason. At that point, you could have bed ways twisted like a barber pole, the tailstock adjusted as far off of center as it will go, and dials that somebody scabbed in off of a metric lathe that don't match the lead screws. You'll cut off sized tapered parts all day long. But you should be able to do that with a decent finish.
I can run a level over the ways but don’t know what is an acceptable tolerance.
Save any "leveling" for down the road. One step at a time. That's NOT your issue YET. For now, acceptable tolerance is "looks about straight". If you really want to do that preemptively, outsmart your carpenter's level. Lay it over the tip of the Vs on the bed at the headstock end and the tailstock end. DON'T TAKE IT TOO SERIOUS.... Carpenters levels have a huge "tolerance" built into them. Start with it up against the headstock. Figur out EXACTLY where the bubble is. It need not even be in the middle. Just eyeball a line width (or half a line width) of accuracy as to where the bubble actually is.. Do not tip, turn, or rotate the level in any way. Lift it from where it is, and set it on the tailstock end. If you're within a line width of the same place (even estimating when it's not on a line), if you're within a line width, you're good. You can do better (maybe, if it's crooked), but for now, that's way more than good enough. Probably not even necessary at this stage.
im struggling to get good surface finish and good dimensional stability and repeatability.
Or, this piece breaks out separately as well. Leave out the surface finish for a bit. Dimensional stability and repeatability needs to be worked on. With any lathe, there will be "some" deflection on every cut, so you're going to have to plan your way in. If you want o turn down two "things" to 1.942 inches, you're not going to dial in a number and get there. You're going to want to leave some room so you can take a couple or three "equal passes" at it, five, ten, fifteen thousandths each. Whatever the lathe is comfortable with in the material, how much you have to work with, etc. I quote "equal passes" because they're not equal. So the two hypothetical "things" here that you want identical, initially you can pretty much have a free for all down to some set number larger than your target. For giggles, let's say that "loose target" is 2.542, as that leaves a very convenient number. When you hit your 2.542 dimension and land just shy, or blow it out of the water... There's still 0.060 to come off of the diameter. .0.30 on the radius. Three at 0.010 depth of cut on the dial (0.020 on diameter should nail it, but each pass gets adjusted a couple of thousandths to keep you on track. for three equal passes, so that the last one, right to dimension, lands on target with the deflection accommodated accounted for. Very small cuts to sneak up on dimension are not all that predictable. You end up trying to take one or two thousandths, the tool just won't dig in, and when it does it loads up and takes three or five.... Well ground HSS tools (or any way you can get a dead sharp tool with very low tool pressure) really helps with that issue.
If you've got a tool that cuts freely, (ish), I'd toy with that. The hitting dimensions part. I would not recommend expensive materials, rather something inexpensive that you can go ahead and burn it up. Doesn't really matter that much what it is. Just big enough that you can spend some time on it without getting it so thin that the part's own deflection becomes another variable. If you don't have any tools that you feel cut somewhat freely, maybe hold off on this step, because excess tool pressure, while you can do it, it just causes unnecessarily large amounts of deflection, which just makes the learning curve that much steeper.
Similarly I don’t know anything about spindle wear or carriage wear and backlash. Where do I start? Am I barking up the wrong tree?
You're barking up the right tree, you just havn't chased a squirrel up it yet.
Take it one piece at a time. You're doing the exact same thing that I did when I drug my 9A home. That is, trying to take on everything at once. Plug it in and make parts. You'd be better served (as I would have) to plug it in and learn to make chips first.
To make a decent looking cut, assuming the lathe doesn't look like it fell off of somebody's tailgate while they were driving down the freeway, you only really need to qualify Qualify the spindle bearings (literally a dial indicator measurement) and verify that the dovetails on the cross slide and compound are tight enough to not wiggle. Then your lathe, regardless of other (reasonable) conditions, should make good cuts.
And the dimensions. Assuming that the carriage lock locks when you want it, and the cross/compound slides are dialed in tight enoug that the dials stay put while you're cutting , you should be able to hit dimensions with some practice.
And the leveling/tapering issue. Cutting tapers in particular. It's a geometrical rabbit hole, it can be a bit tricky to wrap your head around at first, but it can be considered, for all intents and purposes to be a whole separate category from the other two. You can do all the things, measure it perfect, but you're never going to get it "done" until you can measure some cuts. That takes a servicable, even if not perfect) finish on a cut, and some degree of confidence in your getting multiple cuts to come out "dead nutz" the same, so that you can confidently say that the descrepency is in the lathe bed, tail stock adjustment, etc. Outside of a quick check for gross errors, step one and step two have to come first. .
