Thread: Lathe alignment - cutting concave - why?

Hi there, When I first read your thread heading I was thinking you meant while facing. So it's levelled and the spindle test bar shows it's reasonably aligned. You mention it's alignment in the vertical (measured a droop of 0.03mm due perhaps to gravity) plane, did you also measure alignment in the horizontal plane? Other reasons could be perhaps some crud, swarf etc under the saddle? That's all my basic experience can come up with. It is interesting how it does this in the middle instead of a taper.. Edit: Have you tried with a different piece of material, or different size? 7 uM is not a big number for a these type of lathes I suspect. Cheers, Simon

Did you cut the bar to create the sample areas or are you just measuring a bar thats been put in. If you turned it, change the speed on the lathe and see if it makes a difference.

Originally Posted by simonl

Hi there, When I first read your thread heading I was thinking you meant while facing. So it's levelled and the spindle test bar shows it's reasonably aligned. You mention it's alignment in the vertical (measured a droop of 0.03mm due perhaps to gravity) plane, did you also measure alignment in the horizontal plane?

Yes, reading with a dial indicator in the tool post. Along the horizontal plane I get 0 run out along the length of the ground test bar.

Originally Posted by simonl

Other reasons could be perhaps some crud, swarf etc under the saddle? That's all my basic experience can come up with. It is interesting how it does this in the middle instead of a taper.. Edit: Have you tried with a different piece of material, or different size? 7 uM is not a big number for a these type of lathes I suspect. Cheers, Simon

I am not thinking it is swarf under the saddle, as it is only doing this in the mid point of the AL bar. The saddle also has wipers to keep all the crud out. I have tried different materials and recorded largely the same result.

Originally Posted by scottyd

Did you cut the bar to create the sample areas or are you just measuring a bar thats been put in. If you turned it, change the speed on the lathe and see if it makes a difference.

I have an MT5 precision (Chinese) test bar and I have also cut multiple test bars in AL and steel at different speeds. I have seen this concave result in each test. I cut a full length of 45mm AL and I see this concave result. Then I cut reliefs in the AL bar, then cut small collars taking a 0.05mm cut each time and I get this concave result from the cuts. The end collars (headstock and tailstock ends) read 0 with a Mitutoyo micrometer that reads down to 0.001mm. The center collar however reads -0.007mm.

You forgot to mention how long your piece of alum is...

In any case it does not matter. Those results are easily acceptable for such a lathe... When you realise a 0.007mm on diameter is only 0.0035mm sideways movement..

And we all know much much 0.005mm is...

Originally Posted by .RC.

You forgot to mention how long your piece of alum is...

In any case it does not matter. Those results are easily acceptable for such a lathe... When you realise a 0.007mm on diameter is only 0.0035mm sideways movement..

And we all know much much 0.005mm is...

RC, the length of the AL bar is ~250mm.

I will try to cut a shorter test bar, say 150mm to see if I end up with a 0.007mm taper. I really wanted to get things perfect.

Do you have any tweaks I could try to see if I could get rid of the concave section? Does the sag of 0.03mm along the test bar (300mm x 50mm excluding MT5 section) sound right?

Originally Posted by variant22

I will try to cut a shorter test bar, say 150mm to see if I end up with a 0.007mm taper. I really wanted to get things perfect.

Then you will probably need to buy a better lathe. 0.007 in 250 mm is well within spec.

If it's indicating correctly off the test bar, but cuts differently, its possibly the cutting pressure causing the bed to flex. If you have a precision level you should be able to confirm that if you piggy-back that on the saddle when you cut the bar. Compare your lathe to a heavy toolroom lathe and you'll see the difference in bed design.

I'd also suggest using steel rather than aluminium for your testing, and just take fluff off with the final pass.

You obviously need to have your garage temperature and humidity regulated to improve you inaccuracy of your precision manufactured Chinese lathe.
I'm amazed you got it that good.

Originally Posted by Clubman7

You obviously need to have your garage temperature and humidity regulated to improve you inaccuracy of your precision manufactured Chinese lathe.
I'm amazed you got it that good.

Thanks Clubman7. I am amazed by your insight. I see however that geography is not one of your strong points.

Originally Posted by Pete F

Then you will probably need to buy a better lathe. 0.007 in 250 mm is well within spec.

If it's indicating correctly off the test bar, but cuts differently, its possibly the cutting pressure causing the bed to flex. If you have a precision level you should be able to confirm that if you piggy-back that on the saddle when you cut the bar. Compare your lathe to a heavy toolroom lathe and you'll see the difference in bed design.

I'd also suggest using steel rather than aluminium for your testing, and just take fluff off with the final pass.

Pete F, thanks for the comments. I will try piggy-backing my Stiefelmayer level.

Interesting comment regarding cutting steel. I did cut a steel test bar a few weeks back when I was leveling things. I will cut another one and see how it fares.

I find it strange how it can cut convex. A perfect zero at both ends, but yet convex in the middle. I guess I could tweak the level to see if I can get rid of the convex cut.

Anyone reading this that needs a starting point, I recommend watching Tom Liptons (oxtool) lathe leveling videos on Youtube. By far the best leveling and headstock alignment videos I have seen:
Leveling and Setup of the Metalworking lathe Part 1 - YouTube
Leveling and Setup of the metalworking lathe P2 - YouTube

If it's cutting concave, then it's quite the opposite.

That's a lot of run out for that sized bar.

Maybe the bed is bowed.

Rob

Originally Posted by Stustoys

Are you using carbide or HSS?

Carbide. CCGT, AL grade - the shiney ones! (Korloy CCGT060204-AK)

I have some freshly ground HSS on hand however that I can give a whirl tomorrow.

Originally Posted by variant22

Yes, reading with a dial indicator in the tool post. Along the horizontal plane I get 0 run out along the length of the ground test bar.

Try the same test with the indicator on the back side of the test piece. I.e on the opposite side of the tool. Where you would mount a rear mounted tool. That will often show up deflection / spindle bearing clearance.

I see you are micrometer measuring your turned journals, have you run an idicator over those journals and are they zero - zero, despite what your micrometer is telling you? You clocked a test bar, but did you clock your test piece? In a pure world, if you are micing 0-0 at the head stock and tail stock, an idicator should follow zero - zero. You can verifiy that by checking the back of the test piece.

Regards Phil.

So it cuts slightly an hour glass shape (not concave as in facing).

The lathe bed may not have been checked and levelled with a precision level, but instead you may have used the spool cutting test to untwist the lathe bed.

With a cutting test, you adjust two diameters along a ceratain distance (in your case 250mm) to be identical. That would work with an ideal, wear and error free lathe whose bedways are perfectly straight, the rear and front way exactly in the same plane over the whole length. Many lathe do however have some wear on the bedways, usually in the most used region close to the chuck. And this wear is not always the same on the front and rear bedway. There may also be some bow in a new lathe. You still can shim the bed such as to get the two diameters on the spool to be exactly the same. But this will only be true for these two points along the bed. Any point closer or farther than the 250mm may be off.

Think of it like this: Imagine your bed is twisted. This means, that the center height of the tool tip will be either too low close to the chuck and too high at the tailstock end. Or vise versa. When the tool tip is exactly at center height, it will turn the smallest diameter. And a larger diameter when too high or too low. If turning a rod it will be hour glass shaped.

If the problem is caused by bed twist, you may untwist it by levelling it with a precision level. But if the problem is caused by a non-linear bed error such as wear or bow (maybe because the bed was not seasoned long enough before machining and between machining and grinding), all you can do is distribute the error along the length of the bed, so that it is nowhere extremely large or disturbing.

Some first questions would have to be:
- by which method did you level the lathe bed?
- is the lathe new or near new, or could there be wear?
- before measuring, are you letting the temperature of your workpiece equalize long enough?
- are you using automatic feed and a very sharp, freshly ground and honed tool for the test?

Originally Posted by nearnexus

If it's cutting concave, then it's quite the opposite.

Rob

Quite right Rob ... if the tool started on centre.

I took it that the lathe is new or relatively new, and bed wear wouldn't be an issue.

My little 260 bends like a banana with a precision level sitting on the carriage and running it along the bed, and that's a brand new lathe! I was shocked when I first saw it, I'm not sure how much it affects the diameter, but a precision level certainly shows the bed moving with changes in weight around it.