Taper question

[AdeV]

My lathe is turning a taper...



The piece is ~7.5" diameter, and has had a few thou turned off it, as you can see. The length being measured is around 1.5". It is tapered 0.003" (thinner near the headstock)...

I tried cutting using both the cross-slide - which I thought may be slightly off-line, hence the taper - but the final cut was taken using the carriage on power feed - cutting away from the headstock - which had the same result (so, as it transpires, it looks like I got my cross-slide straightened up nicely  )

What's out of line to cause that? The spindle? I'm not sure it can be anything else, unless I'm doing something stupid somewhere? The taper attachment is not in use... Also, although I'm using a 4-jaw chuck, the disc was dialled in to within a few tenths, maybe 1/2 thou once I'd cinched all the jaws down hard.

[CrewCab]

If it's an error in the turning then headstock bearings seem the first thing to investigate, however, have you measured the work piece at both ends with a mike or vernier, just wondered if it could be a bed error assuming the gauge is mounted on the carriage? ................    then again I could be woofing up the incorrect tall leafy thing  

CC

[ieezitin]

Adve.

Your in the sweet spot for bed ware there being close to the chuck, your saddle does a lot of movement at that point! One thing to check.

Your chuck is it seated firmly on the nose? Are there any burrs and or crud on the threads?. Stick a clock on it and see if there is movement and or check the threads.

If you have a test bar slam in a dead center and rig up to tailstock and dial in and or make cuts if necessary if its a consumable bar, this will tell you once you have truth if you cut again and there is a variance then the bearings may be failing.

Are you using excessive oil in the spindle bearings?

So many variables that you need to knock down one by one. it’s a PITA in your situation but a process of elimination from here out.

Hope this helps.                  Anthony.

[Darren]

Somebody please feel free to correct me if I'm wrong, but even if the headstock was slightly out of alignment it would still cut straight. Not straight with the reas of the bar in the chuck but the cut area wouldn't be tapered.

Same goes for not being true in the four jaw chuck.

 

[ieezitin]

Darren.

Hope this drawing works. I have tried to answer your question as I read it. Please feel free to ask or tell me if I have your question wrong.

Adev.    Your chuck jaws could be going a scew over that length too. You may not have a problem other than the jaws in the chuck.

All the best .   Anthony.

[andyf]

I'm not sure about that, Darren. Agreed that if the chuck is holding the work on the skew, but the headstock is in line with the lathe bed, and the bed isn't warped, the result will be a perfect cylinder, because the axis of rotation is parallel to the bed.
But if the headstock, and hence the spindle, aren't in line with the bed, the axis of rotation of work held in the chuck won't be parallel with the bed, and a tapered cut will result. It's the same as using a set over tailstock to cut a taper on work held between centres.
Ade, I'd suggest checking with Rollie's Dad's Method (Google it). It can't distinguish between an out of line headstock and a warped bed, but the latter is unlikely to be so severe as to give the 2 thou taper in about 1" of length shown in your pics.

Andy

[AdeV]

Thanks for the replies chaps - too late & internet is too slow to answer them all individually now, so I'll do a couple:

Anthony - as far as I am aware, the chuck is mounted OK. I've not disturbed it since I bought the lathe a few months ago. I've not checked the runout on the chuck body, will do that tomorrow. However, the work hasn't been moved since I took that cut, so I can't see how the chuck would cause a taper? Even if it's mounted cockeyed, surely the lathe would cut what seemed to it to be a perfect cylinder - it'd only be when you compared that to the ends (or previously turned work) that the lack of concentricity/angle would show up?

I've only checked with the DTI on the top of the work so far, it may be that there's wear in the bed causing the saddle to drop. A long test bar ought to show this. I think I have enough kit here to actually turn between centres now, so maybe that's something to try. But not before....


...Andy - thanks for the Rollie's Dad Method reference; first link was a PDF which explains it in nice clear terms. I didn't immediately "get" why it works, but I see it now. I'll try that first, before I go putting drive plates on the lathe, etc..

I see you also answered, in better terms than I, why I don't think it's the chuck at fault...

[andyf]

Hi Ade,

It always takes me a while to get my head round it, every time I read it.

"RDM" probably pre-dates Rollie's Dad, but it's a handy title. There's a walkthrough with pics on Gadgetbuilder's site at http://www.gadgetbuilder.com/Lathe_Align.html . The headstock shimming he describes is for a mini-lathe, but the principles are all there.

Andy 

[AdeV]

Hmmm, interesting....

The horizontal misalignment is about 0.0005" (1/2 thou); and that could have been introduced by the pressure of the clock's plunger on my 24" 0.625 dia bar...

Vertically, the misalignment is some -0.0340"! i.e. -34 thou.... (over about 24"). That's quite a lot... and I guess is plenty enough to cause the taper I'm seeing. FWIW, that -0.034" is with the plunger clock pressing down on the bar, so there may be a slight over-read due to that. Even if it's a 30 thou droop, it's significant.

So... either the lathe is banana shaped, or the headstock is "nodding" slightly towards the tailstock... I shall try shimming the tailstock & report back for further advice. Meanwhile, I have to go get the cat fixed....

[Dean W]

Ade, try cutting a test bar about 4-5" long, a couple of inches in dia, then measure the ends.   When you cut your bar, take material off it's entire diameter for its full length.  Then take about three cuts for your final finish without putting on any cut, and run at the finest feed.  That will let the tool cut out most of the flex.  This will give you a fairly quick and hard result.  I'm betting it will show that you have nothing like the amount of taper you thought you had in your first post.

Running your DI on the top of your work piece doesn't really tell you much about any actual taper your machine may be cutting.  Try running it along the side where the tool cuts. 

Many older lathes will have a low area in the bed near the chuck.  It might sound bad to hear that, but it doesn't make a huge difference until it gets pretty bad.  The tip of the tool dropping down a couple of thou doesn't make near the difference to the diameter at each end of a piece as it would if it were actually moving toward or away from the center line of the spindle, as it effectively would if the headstock were askew.

This problem won't be because of the chuck, or how you mounted the work piece.  It doesn't matter to your work piece if it is mounted crooked in the chuck, as long as you are taking material off for the length of the piece.  It could be mounted practically sideways, and if the lathe is cutting right, the work piece diameters on the ends will end up true.

Dean

[kwackers]

Is your bed twisted? I can easily get that sort of taper by jacking the bed - doesn't take much.

Assuming no bed wear (or little) then considering you've not had the lathe long it sounds like a possibility? The manual for my lathe goes into checking for cutting taper and explains how to adjust the jacking screws to correct for this.

When I first got my lathe it cut taper and I was fairly suspicious of what the manual claimed - given the size of the bed it didn't seem likely that you could twist it enough. The twist was easily detectable with a engineers level, I adjusted it out and then machined a test piece as per the manual and did a few final tweaks. Result, no taper.

The manual suggests machining a length of bar (fully supported by your chuck) at least 4 inches long. The centre of the bar is machined down leaving two 1 inch islands at each end. These are cut on power feed with a final cut of around 1 or 2 thou and then measured with a micrometer.