Lathe unwanted taper

Yes chuck side is turning larger.

That's right, and getting taller, thats why I left it on a hoist!

Jokes apart, Jason copied that pic from one of my previous threads...it is now safely bolted to a bench

Thanks for the offer Tim, much appreciated, but as Ndiy said last time I checked Malta was still surrounded by the sea

On the subject of levels, I managed to knock out a simple spirit level which is pretty accurate and repeatable, used scrap pieces I had and a glass tube. Bowed the tube slightly with a set screw, works a treat and cost me nothing!

I appreciate that the problem is well on the way to being solved, and a little more shimming and tightening should see everything running true, but I wonder about the comments regarding workpiece flex.

Surely if the fat end of the taper is at the headstock, then it's not a flexing issue; if the workpiece was flexing away from the cutting tool, then the taper would be of largest diameter at the tailstock end, (or am I missing the blindingly obvious again)

Bill

Well spotted! I'm one of several who missed the blindly obvious...

Depends, If the tool has back rake with a soft metal like aluminium it might be pulling the job toward the tool rather than pushing it away. Depends too on if the tool is slightly above or below centre.

Dont think it was material flex, I had a dial rigged and pushed the bar by hand to get a feel of how much deflection I would get, the cuts I was taking were in the order of 0.02mm, very shallow. As for tooling I used both positive and neutral back rake angle tooling, with no noticable difference.

This morning purchased some mild steel round stock and other material for some projects I have in mind, if I find time this afternoon I'll turn a test piece and check again

I was sceptical about this (that a 1/2 kg downward force would cause the bar to deflect by 0.2 mm, or .008" ) but whilst the common reaction on t'internet is to fire off an opinion, correct or not, I decided to conduct a real world test.

I have not got a piece of 20 mm aluminium bar that long in stock, but I do have 3/4", i.e. 19.10 mm.

I chucked this up in my Boxford, set up a Kennedy DTI that measures .0005" increments at 8" from the chuck, and applied a 1/2 kg piece of brass rod on top of the ali at 8" from the chuck.

Result? Less than .001" deflection, and this is 3/4" (19.1 mm) bar, so deflection of a 20 mm bar would be less.

In addition, Chris taking a light lengthwise cut will not result in all the force of the cutting tool deflecting the bar.

Some (or most) of the force of the cutting tool is in a lengthways direction, so would not deflect the bar sideways by all of the force of the cutter.

Noting the above, I would say that despite not using a tail centre, the taper turning is NOT caused by the bar being unsupported at the tailstock end.

I'm not saying it should not be supported, merely that I don't think this is the cause.

I would go with Jimmy's theory, that the headstock is likely not in line with the bed, or Hopper's theory that the lathe bed is not level (ie, it has a twist)

Edited By David Standing 1 on 10/11/2017 13:55:38

This morning I had a go turning another test piece from the mild steel I bought yesterday. Turned a 1" O/D piece over 150mm. First off I turned 2" from one end to remove the black scale and turned it around in the chuck to have a proper seating, then I tried to machine the rest of the bar (about 150mm) but it chattered badly, even with a light cut, had no such issues with aluminium.

Probably aluminium being soft complies with the tool and shears without deflecting much, steel being harder deflects, or so it seems to me....

Anyway I supported the free end with a revolving centre and took multiple cuts, the results are the same as those I had with aluminium, or a bit better difference of about 0.01 - 0.005mm over 150mm, not bad I guess.

The only issue I'm noticing is that when I tighten the tailstock barrel lock, the test piece deflects and then when I machined it I got a taper (0.05mm difference wider at the chuck end) If I just snug the lock it will be fine.

Also, checked the bed for level, the twist it originally had is much reduced now, initially the level was reading out of scale in opposite directions, now it's still reading some twist but much reduced (a normal spirit level will read the same at both ends)

Chuck side reading close to 0 after test cut

Tailstock side reading 0 after test cut

Deflection when tightening the tailstock barrel lock, get greater the more I tighten up to 0.05mm max

Level at tailstock end

Level at chuck end, a bit off but I think it's good enough, it starts to get tricky to get the bubble centered at both ends!

If you're measuring the work "diameter" variation by running a DTI along the work then all bets are off. You need to measure the diameter with a micrometer to get a proper picture.

Andrew

Try another chuck first? or turn between centres?

Edited By john carruthers on 11/11/2017 10:28:00

Absolutely!

Changing the chuck will make no difference. The centre line of the spindle rotation will be unchanged and that is what the cutting tool must be aligned to.

Martin C

Is that a carbide tool in the picture? Possibly explains your chatter problem, at least to some extent. These small lathes do better with a sharp HSS knife tool bit, with a radius on the end of just a few thou, especially on light cuts. If you still get chatter, you might try having the test piece sticking out of the chuck by 4" rather than 6". (Assuming here that you have checked headstock bearings for tightness and all gib adjustments also.)

+1 on use a mike not a DTI to measure your taper.

Using the tailstock centre tells you nothing about the alignment between bed and headstock, only about the offset of the tailstock, which is adjustable. It shouild be adjusted to turn parallel when the tailstock base and barrel are firmly locked in position.

Edited By Hopper on 11/11/2017 11:47:39

Chris B.

Have a look at this **LINK** site by the much respected Harold Hall on the subject of getting a lathe to turn parallel.

Eric

Haven't got time now to type the way I set up the Myford without any fancy gear, but have a look at Rollies Dad's Method of Lathe Alignment which should take you to a pdf.

When you've got the lathe to turn true along the bed with a chuck held work-piece, then and only then, can you set about adjusting the tailstock.

Bill

The diameter is the same at both ends according to the micrometer, I get what you mean, re dia measuring by dti, as it will follow the taper if there is one.

The tool I'm using is HSS one of the ready made ones from chronos, I just rounded off the edge. I checked the headstock with a dti on the spindle nose and face and every thing looks fine.

When I turned a couple of inches unsupported it did not chatter at all, only when I put 6" unsupported did it chatter. Probably should have turned a shorter test bar.

Chris

Not meant to be a criticism, but if that micrometer is sitting there unsupported (even if it is resting on the frame), then you probably have it cranked up way too tight. It is a micrometer, not a G clamp.

I can also say, having levelled my machines with a proper Rabone Chesterman machine level, it is astonishing how much a fraction turn of a bed adjusting screw can deflect the bubble.

If I wanted an indicative readout from a machine level, I would only use a proper machine level, not a home made one. A normal spirit level won't give the readings you want either, as you have discovered.

The fact that you have already detected twist in the bed suggests you should be absolutely sure you have removed all this first, BEFORE attempting any test cutting.

Edited By David Standing 1 on 11/11/2017 13:29:00

It's ok David, I'm holding it by the ratchet (tried to leave my fingers out of the picture as I didn't do nail polish!) I'm sure it wasn't tight.

A proper engineer's level is not an option, don't have one and I'd rather spend the money on other tools which I really need. I one I mocked up is by no means anywhere near a proper level, but it gives a far better picture than a normal level and allows for some adjustment to be made, where as with a normal level I was blind.

To be honest I'm pretty much almost satisfied the way it's running now, it's turning a set diameter over a length with no significant taper. The thing which is bothering me is the tailstock deflecting the work when tightening the spindle/barrel lock.

Will check the posted links re tailstock alignment and see if it fares a bit better

Thanks to all contributors for you help

If you want a 'target' to aim for schlesinger limit for spindle alignment is 0.02mm over 300mm.

Tailstocks generally don't stay aligned for a long period of use and should be reset before critical jobs.

Neil

Tried the rollies dad method and managed to get the horizontal alignment right, but vertical alignment was another story. Instructions say to lift both corners either tailstock or headstock side, but which ever side I lift no change occurs, must be doing something wrong.

Chris, Firstly I should stress that I'm not a trained engineer, well I am, but in telephony not mechanical. Everything I've learned has been self taught, so read anything I write with an appropriate pinch of salt.

I'd suggest that the reason you've struggled with the vertical alignment is to do with rigidity of the workbench and lack of mass in the headstock, plus it would only really work for lathes with 6 or more mounting bolts, 4 at the headstock and 2 at the tailstock on a very rigid bench.
If your lathe is significantly out of vertical alignment, I wonder if having a word with Warco might be a good idea, as that really should be running more or less true when it left their depot. I'm assuming here that the headstock was correctly bolted down when it left the factory, without any flecks of paint underneath (Think Hubble Telescope)

I think you understand the principal of putting a twist in the bed to make it turn parallel with a chuck mounted workpiece. The vertical alignment would require the whole of the bed bending a tiny amount with respect to the centre of rotation of the headstock spindle, and I can't imagine that this should be required in a lathe of this size.

The caveat to that is, have you actually put a bend in the bed whilst adjusting the twist in it. If so, don't worry as I'm sure it will be within the elastic limit of the cast iron. That is, in the ideal world, if all four bolts at the headstock are firmly fastened and cannot move at all in relation to the ground, then everything about the headstock is fixed and all adjustments at the tailstock end are in relation to it.

"Levelling the lathe" (with a capital L) refers not just to getting it level from headstock to tailstock, indeed, that is the least important, though obviously it should be as "level" as possible. The main one is fore and aft, which should be the same at both ends of the bed, hence the "twisting" scenario to make it so.

I think you should be able to see that, if the headstock casting is truly immovable, then if the tailstock bolts were pulling that end into the bench as the wood squashes, the the bed would deflect downwards at that end. Obviously a similar scenario if you put a couple of washers under the tail end of the bed before tightening down the headstock, then the tail end of the bed would deflect upwards, with respect to the headstock.

I'm assuming that since you have neither a calibrated test bar or precision level, you will need to sort it out by turning and measuring.

I'd suggest making sure that in your case the "level" between head and tailstock in neutral, and that any differential shimming only adds a twisting, rather than a bending moment as well, in order to keep the vertical alignment correct.

Here's the way I set up my Myford, which may or may not be approved of by the many apprentice trained engineers on here.

Take a decent length of steel bar, similar to what you've been using and maybe a foot or so long, clean one end and mount in the chuck. thin down the middle section of the bar by a couple of mm or so, similarly the far end as well; no need for accuracy. The intention being, that you well be left with two collars standing proud of the main bar as far apart as the length of bar will allow. (make sure you leave enough room at the right hand end beyond the collar to allow a drive dog to be fitted later.)
Now sharpen and hone a HSS tool to the best edge you can get. Hopefully by now if you've been honing for long enough, the formative test bar will have cooled down, so take the very finest of cuts to both collars.

Measure them with a micrometer, ideally a "tenths" one - mine are all imperial. If they aren't exactly the same diameter, which is most likely, this should tell you which way you need to apply a twist to the bed.
Once you've managed to get two identical collars, you know the two ends of the bed are "Level" with each other.

Only now can you turn your attention to the tailstock.

Centre drill both ends of the bar. Either fit, and check, a centre in the headstock taper, or chuck a bit of steel in the 3 jaw and turn a taper on the end to make a centre which you know is true - it must be you've just made it.

Fit a drive dog and turn the bar between centres; with your nice sharp tool, again with the finest of cuts true up and measure the two collars. Keep minimal projection on the tailstock barrel during the operation.This will tell you whether the tailstock needs to move towards or away from the front of the lathe.

Once you have the tailstock correct, true up the end of the bar that you held in the chuck prevously, as it will now be slightly out compared to the centre drillings.

You now have a lathe that's reasonably well set up, and a re-useable test bar as well, so keep it in a padded box with some rust resistant paper. OK so it's not as good as a calibrated precision ground bar, but if you could afford one, you'd have already bought it.

Bill

Part 2 to follow