I have an old clockmakers lathe, which has been useful, but I have done nothing yet which requires working to a tolerance.
The lathe is old (1905) and so I can be fairly certain the bed is worn (its a "double inverted V" ^- -^ as opposed to the Myfords which seem to be flat for some reason). I also expect the the wear is localised (its a standard 3-1/2 inch x 36 inch lathe) close to the headstock.
I can't work out how to measure the wear on the bed, so I though a useful qualitative test might be to take a full length light cut off a 60cm silver steel bar, turned between centres and then repeat with a second identical bar.
If I then clamp the two bars together I'll be able to see the profile of the bed wear by placing a light behind and looking at any gap.
The lathe isn't worth the hassle of getting the bed re-ground and the thought was that I could then setup the head and tailstock over the best section of the bed, because I don't have any projects in mind where the workpiece is longer than about six inches.
Provided the lathe's initial setup is OK are there any flaws in my thinking?
The double inverted Vee bed is superior to a flat bed as regards accuracy of alighnment of the headstock, tailstock and carriage. Even when slightly worn these main components should remain as well aligned as original althought the carriage height will drop with wear but can still turn true of course. Wear is far less on the tailstock as it is on its own vees and moved very little compared to the carriage so even a worn lathe with an inverted vee bed should have the tailstock and headstock aligned and the vertical centre heights still the same. Why not uy a test bar and put it between centres to use with a dti?
I don't see any advantage in using silver steel ? The final finish will be turned in any case and if you just want something to take a skim cut then you can turn down any pair of old bits of steel to have the same initial diameter before your skim ?
Err yes. Major ones. You aren't going to be able to turn a 2 foot long piece of bar to any sort of accuracy unless it's like 4" in diameter and even if you could, silver steel is hardly the best choice. It doesn't cut that nicely compared to many other materials. Even if you had the right material and a thick enough bar it still tells you three fifths of sod all about bed wear.
Bed wear is in the vertical plane. It makes very little difference to cutting diameter in the horizontal one and what difference it does make is a function of the bar diameter. The smaller the bar the bigger the effect of a vertical displacement of the cutting tool. However, even with a very small diameter bar you aren't going to see any noticeable change in the cutting diameter from a few thou vertical wear on the bed. The most likely thing you are going to notice from a test bar is not vertical bed wear but wear or play in the horizontal plane in the gib strips between the saddle and the side of the bed.
And finally, all the above notwithstanding, you don't do this sort of test by cutting a bar in the first place. You do it by putting a ground test bar between centres and running a dial gauge along it, but even if you do that the biggest thing that affects lathe accuracy is wear in the cross screw and the dial gauge won't tell you about that anyway.
So we are boiling down to the conclusion that your proposed test is A) impossible to do accurately and B) pointless. There's an old maxim. A good turner can do accurate work on a bad lathe and a bad turner can't do accurate work on a good lathe. The operator is actually more important than the machine. Once you know your machine's foibles you can compensate for them in one way or another. My old Colchester Student is so shagged it ought to have been pensioned off years ago. There's 2mm of slop in the cross screw and the gib strips get tight in some places and loose in others. I can still work it to under half a thou when push comes to shove though. I keep an allen key in the bolt that adjusts the split cross screw nut and tighten and loosen that as required to tension the play out of the toolpost. In extremis I'll even lean on the saddle a bit to tweak the cutting size when I'm working down to the last few tenths. I'd rather have a good machine that works repeatably but I can manage with a clapped out one by knowing how to use it. That's the most important thing you have to do with yours.
I did think about using the DTI, but was worried about the ball on the DTI and its position on the curve of the bar. If there is wear in both vertical and horizontal planes, then the DTI will read a mix of both won't it?
My thought about using the two bars was really about copying the shape of the bed to the workpiece, but as Dave Baker points out, wrong on all counts!
It sounds like I'd need to mount the cutter vertically, use a light cut and a bar of appropriate diameter and material. Dave again suggests its not worth doing anyway, so I'll concentrate on improving the rigidity of the saddle and toolpost for now!
Nope- you'd just run the DTI along the top of the test bar rather than along the side of it.
Dave again suggests its not
It isn't compulsory to have a ball end on a DTI. You can make a flat end easily enough. Anyway it's a minimal error. You'd need many thou of vertical movement to get even a fraction of a thou horizontal movement as the DTI tip moves away from the exact centreline of the test bar. It's a mathematical problem that can be solved by calculating chords of circles. I wrote one in a spreadsheet many years ago. Here's what it tells me.
For a test bar of diameter 19mm here are the horizontal errors for a given vertical movement of the cutting tip from centre height.
In other words three fifths of FA. 20 thou vertical error to get even half a thou horizontal error. You wouldn't even measure that much error with a DTI because of the radius of the ball end but that means solving two chord problems simultaneously and that's beyond me to be honest. Like I said previously, vertical wear in a lathe bed makes no difference worth worrying about to cutting accuracy.
Things you want to worry about are getting the headstock perfectly aligned to the bed and the tailstock perfectly centred. Once you've done those two things you can work to very fine limits on a very shagged lathe. I routinely work to about 2 or 3 tenths on mine when I'm turning valve seat inserts to be a perfect fit in the counterbore in the cylinder head. Similarly when I'm making bronze valve guides I try to get within a few tenths on the o/d before polishing the last bit off with emery before fitting the guides. You can't do that just by leaving the tool on a given setting. There's too much wear in my cross screw for that and it moves around a fair bit every time I snick the feed lever into place. You have to get within a couple of thou, measure and then sneak up on the final size with dust cuts but it can be done.
Good practice is to get a bit of bar made of something that cuts really cleanly and play with it. Aim for an exact size and see how close you can get and what it takes to achieve accuracy. Tool tip geometry, tip sharpness, flex in the bar and all sorts of other things make far more difference to final size of cut than wear in the lathe. Bronze or heat treated aluminium cut nicely as does leaded mild steel or EN36. Many other steels just tear too much and aren't nice to play with.
Thanks Dave - this is good stuff that is hard to pick up in the books. I have books that tell me how to set up a lathe, but they don't tell what not to worry about!
Fantastic bit of advice Dave, This should be etched in stainless and hung above the lathe... Next time you curse the machine for eating that last thou !!! LOL
PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.