Taper Turning using set-over attachment in tailstock

I wish to cut a short and accurate internal 1:16 taper (3/4" per foot) with my topslide and my plan was to se a commonly available set-over centre device placed in the tailstock, setting it over an amount according to the 1:16 ratio (actually half the amount) using a DTI gauge for accuracy, and placing a ground test bar between centres to use as an angular template for setting the topslide angle using again a DTI gauge to travel up and down the test bar until no variation. The set over amount for the tailstock attachment is a function of the length between centres, and my question is , is it best practice to take the "length" as the distance between the tips of the centres, or the length of the test bar, or some other length that takes into account the movement of the conical positioning of the centres inside the ends of the test bar?

Edited By Adam Harris on 25/08/2019 23:47:42

Edited By Adam Harris on 25/08/2019 23:48:53

No need for the set-over device as that simply adds another layer of potential error.

Start with a true running bar exactly on the lateh centre-line. A good test bar in the headstock centre if you have one will do fine. Alternatively turn something accurately between centres. Needs to be comfortably longer than the top slide travel.

Use the top-slide graduations to set it to approximately the right angle and mount an accurate DTI in the toolpost so its probe runs along the centre line of the bar. Now you can use the sine of the taper angle to calculate the amount of DTI travel when the top slide is moved a suitable distance. I'd use something giving convenient tosplide dial readings close to the central third of the slide travel. You need to keep a decent amount of the slides engaged for shake free movement. Adjust the slide angle until you get the right shift in DTI reading. You really need a pusher screw system for controlled adjustment of the angle.

Hafta calculate the DTI shift via the angle or Pythagoras because the slide is moving down the taper side (i.e hypotenuse of the triangle). If you are using a taper turning attachment with the saddle moving along the bed (i.e base of the triangle) you can set things up directly in inches per foot. Much easier.

Clive

The exact geometry is always suspect unless you use a 'ball ended' centre, when you can take the centre of the ball as a datum. Certainly it's unlikely that you can set it accurately this way.

Better to set the bar by moving the saddle (not topslide) by, say 4", and adjjusting until the reading changes by 1/4" over the 4".

Then set the topslide until it reads the same along the bar.

Neil

You can get it in the right ball park using trig or CAD as I tend to do now and then use the method Neil describs again with the difference over a set length worked out by trig.

Bar length would be the one to base your initial calculation on if using an offset tail stock ctr..

Quick way.

Put a suitably stout stub of metal in the chuck with around 3 inches of stick out.

Turn to a good finish.

Mount the DTI on the top slide and adjust the angle until you have 0.125" of DTI travel in 2.004" of top slide travel.

One thou error in top slide movement corresponds to tapers of 3/4" in either 11.9946" (under) or 12.0066" over. Unless your top slide is in poor condition you should do better than ± 0.5 thou in travel. Heck if you are worried use the DTI to verify what dial readings correspond to exactly 2.004".

Faffing around with the extra bar you will probably stack up similar errors anyway.

Ball ended centre argument is irrelevant in a stationary set up. Proper way for a set up is to use vertical pin joints like half a universal joint but that is getting silly.

If you do use the set over bar an move the saddle as Neil suggests work over 12" so you are using the specified slope values and set two bed stops to define the travel to exactly 12". Not so easy to define travel exactly on most lathes if you don't have a DRO. Can you set the travel to less than 5 thou error. If not best go the turned stub route.

Clive

PS Always verify your error budgets.

Edited By Clive Foster on 26/08/2019 08:39:45

Thanks everyone for your advice - very much appreciated! I don't have a DRO and I don't have saddle stops on both sides (only 1 on the chuck side) and the graduations on the saddle handwheel are only 0.2mm and visually rather close (and operationally rather clunky) whereas the topslide has easier visually spaced graduation marks of 0.025mm. I could use the extended arm of a long digital calipers jury rigged to the bed for saddle movement but I think much simpler to take Clive's method of just using the topslide along with my metric DTI (0.01mm graduation marks) and ground testbar, taking the pythagorean calculation of the hypoteneuse travel of 12.005858 inches should put 0.375 inches perpendicular travel toward the test bar, ie a ratio of 0.03123475. Thus in metric terms, 57mm of topside travel along on the hypoteneuse should result in 1.780 mm DTI travel - Bob's your uncle!

Edited By Adam Harris on 26/08/2019 15:50:43

Edited By Adam Harris on 26/08/2019 15:53:22

I can start the initial tuning at 17mm topslide travel for 0.530 , then 25mm (DTI 0.780) and 33mm (1.030) but I can see that gently tapping the topslide bezel around is going to be a test of skill, and then locking up the nuts without changing anything...Actual length of internal taper I need to cut is about 68mm

Edited By Adam Harris on 26/08/2019 16:14:54

You only need one bed stop and a bit of packing of a known width, 1-2-3 block would be ideal but you can soon face some scrap to a set length. Just drop that in between saddle and stop to take a measurement and again with saddle against stop.

For fine adjustment of topslide clamp a threaded block to the crossslide and use a screw in that to push the topslide to required angle.

Thanks Jason. I will follow your suggestion to clamp a threaded block to the cross slide if i can

Best to arrange two screws pushing on either side of the cross slide for your adjuster. Positive move both ways. Nice fine thread is best.

In my view this sort of thing is worth taking time out from what you are doing to make a reasonably proper job of it rather than a quick and dirty lash up. Proper job always works better and faster. Odds are if you need to do this once you will need to do it again so make life easy. Its not exactly complicated to make.

Same with two bed stops. Jason is dead right that a specific spacer works fine but its inflexible and tends to get snaffled for material stock when you need a bit "that size". I have a gash won't wring anymore set of gauge blocks for such spacers so no snaffling. If I didn't the second bed stop would have got made for both lathes within a week of installation.

Going back to setting the taper I now recall that the last time I did such a job using the top slide I set up directly on the parallel turned material blank. Worked well. But it must be 35 years since I've had to do without proper taper turning attachment.

Clive

Should add that the push pull or just push for that matter only really works for topslides that have a fixed pivot point not one you can slide along the cross slide so no good for the modern 250, 280, 290 type machines

I hesitate to run counter to Clive Foster in the debate about ball-centres, but ... If you use conical centres, rotating the work or not, the exact angle to calculate when setting up requires that both conical ends sink exactly the same amount into their respective centres. And the the centres themselves are drilled using the same size of centre. Not really a problem, I suppose, as long as the hardness of each end, both points and bar, is the same. Until wear takes place, of course, on the minute slivers of steel in contact.

But a pair of ball centres does away with such problems. And blueing the balls for each end (not needed to be exactly the same diameter) and pressing them lightly in the relevant centre cone each end, will allow you to measure with some precision the relevant length for your calculation. This last is not possible using cones which bear on part of the mating cone on one side, and a different part on the other side, each end.

Its not as though these ball end things are difficult to make - although I do not understand why they are not available off the shelf. (Or are they?)

Cheers, Tim

Thanks everyone - I have managed to do the angle adjustments by judicious light tapping with a very light hammer. On a 57mm travel I am only out by 0.01mm which i think is good enough. The difficulty is really the effect of tightening down of the lock nuts, but after removing and thoroughly cleaning and then replacing nuts and washers, it all worked much better! If I am out by even as much as 0.015mm (about 0.0006" on a 68mm taper I am ok I think because it is a bronze propeller hub, fitting onto a stainless shaft and there will be a little room for the softer bronze to smooth out the tiny errors. Thanks everyone!  I may have a go after dinner at getting the error down to zero. 

Edited By Adam Harris on 26/08/2019 19:56:20

Edited By Adam Harris on 26/08/2019 19:57:33

Edited By Adam Harris on 26/08/2019 19:58:01

Jason, the lathe is an Emcomat 20 so yes topslide revolves around fixed pivot point

Why not finalise the Top Slide set over by turning a male taper? You can make any mistakes there, without endangering the final workpiece.

When that is right, rotate the toolpost, or change the holder, to the boring tool and then bore the female bore in the prop. hub.

Howard

For things like prop drivers and collets or a tapered crankshaft I try to do both at the same setting then even if you are off by a tiny amount from the desired angle both tapers will match. You just need to plan your work so that the two parts are ready to have their tapers turned at about the same time. Depending on the setup it can sometimes help to cut teh female taper with teh lateh running in reverse or going into a small hole that opens out.

I'm sure you know that the measuring equipment must touch exactly on the centreline, and this also applies to the tool height, internal and external.