Lathe spindle alignment

I decided that a bit of measurement was in order.

I put my test bar in the spindle and marked rings at 50mm intervals and lines lengthwise at roughly 90 degrees. the bar is consistent diameter through the measured area within a couple of microns.

The picture below is the test set up.

I took care to set up the indicator (a 2 micron jobbie from China) in line with a centre in the spindle first.

Here is a snapshot of the results. Note that I repeated this twice with the test bar in different positions and each time ran through the measurements twice - the results were not identical, but were reasonably consistent.

you will see that the spindle appears to move away and then towards the indicator with a maximum error (on average) of 83 microns 150mm down the bar. I can only assume that this is due to a twisted bed, something I'm not in a position to fix at the moment.

This does seem to show that the average error is now under 20 microns (< one thou) for the first 100mm, which is reasonable.

I'd be interested to get input from anyone who is better equipped to interpret this data - I certainly am not!

I will be on the lookout for a bit bit of steel at some point - there is a scrapyard in ripon and I think this is an application which is fine for some IDontKnowWhatItIsIum as long as it's sizeable and heavy.

Iain

.

Good thinking

I’m happy to be wrong, but … My first instinct is that the bar should be sagging under its own weight, and therefore your figures need to be inverted.

The ‘bump’ at the beginning is a bit odd, but it’s small enough to possibly be an indicator error

… alternatively; if it’s real, then I would be checking the fit of the Morse Taper rather carefully.

MichaelG.

.

Edit: __ Apologies: What I wrote may be nonsense … I’ve just realised that you were using a lever indicator on the [presumed] centre-height of the bar, not a plunger measuring vertically.

Your arrangement could complicate the geometry somewhat. 

Edited By Michael Gilligan on 12/06/2023 18:05:49

Anybody familiar with Rollies dad method of averaged Horizontal and Vertical measurement bar does not need to be straight or runout on chuck does not matter, I found this interesting.
Rollie's Dad's Method of Lathe Alignment-PDF

The bump in the measurements could indicate wear (not worn) of the lathe bed. It can also indicate that the gibs are to tight or a combination of both.
Measure once moving the carriage from left to right and once moving from right to left. If you find similar results, the gibs are not to tight.

20 microns/100 mm isn't bad for a "stock" mini lathe. Beware, it would mean 0.04 mm in turned diameter. If you are happy with these results, than all is OK.

If you have a chuck mounted, the first 40 mm of the lathe bed aren't used. So it is to be expected that this part of the lathe bed is in mint condition.

This is how I measure the headstock alignment with the chuck in place:
Measuring headstock alignment

Edited By Huub on 12/06/2023 19:10:04

Hi, Huu,

My 0 starts about 50mm from the spindle as that's about as far as the cross slide will go. The lathe has a single prismatic feature so should not be too subject to wear. I have, however, just tuned up the clamps to the bottom of the bed (I suppose they're gibs) and dropped the lift from 0.04mm to about 0.02. Any tighter and the cross slide won't move.

Chris. I think I need the lathe back together to use the famous Rollies Dad method, which I think boils down to using averages.

Michael. there will be some droop, but this should not have much affect as long as I'm close to the centre of the bar (which is effectively vertical over a short distance).

And I've measured in both directions and get the same results.

More tomorrow!

Iain

Keep up the good work, Iain

MichaelG.

What you have done is essentially the same as the infamous Rollies Dads Method, ie taking a bar that may or may not be dead straight and dead true and taking readings with the bar rotated through 360 degrees and taking an average. Job's good.

The acid test will be doing a turning test, where cutting forces acting on the machine can give a different result from static measurements. Recommended turning test bar is steel 1" diameter sticking out of the chuck by 4".

You are right, static measuring and actually turning can/will give different results.

I prefer making a " Rollies Dads" test bar because it is easy and cheap to make and because it is running true it makes adjusting a lot easier.

Another cause for the bump in the measurements could be some "damage" on the trapezial part ( single prismatic feature) of the lathe bed. If something drops on this part, it will cause a dent and also high spots near the dent. You can grind away these high spots using a small stone.

Without having waded through the whole thread, will climb onto one of my hobby horses.

If the lathe is not secured to the bench, the bed maywell be twisted. The cassting will have stress relieved itself and twisted once it had been machined.

As already mentioned "Rollie's Dad's Method" is the way to remove twist from the bed.

(Read Ian Bradley's "The Amateur's Workshop", or his " Myford 7 Series Manual" which describe the techniques already mentioned. )

Removing the headstock, will have made this task more difficult, since you will be unsure whetehr you are aligning the headstock or removing twist.

Moral:

Always gor for the simplest things first. The Headstock might have beem aligned, and the bed just needed the twist removing.

But, you are where you are.

I would suggest, having bolted dfown the lathe, trying to take twist out of the bed by using a sensitive level on the cross slide,, at each end of the bed, by travesring the Saddle from one end to the other.

Ian Bradley will tell you which side of the tailstock mounting foot should be shimmed / adjusted. (I prefer using screw adjustment to be more precise. )..

Once that is done, as accurately as possible, then you can use the Alignment bar and start aligning the Headstock.

Good Luck!

Howard

Hi, Huub.

Ways are stoned, but that part of the bed looks OK anyway.

Howard. All this makes sense, but the bench is way out of true and sorting that out to the point where I'm near where an engineers level would indicate would be a goodly amount of work.

However, the thread has persuaded me that this would be a good thing to do and I will add it to my todo list. Probably associated with adding in a heavy baseplate as suggested earlier.

Iain

You are better of with the same amount of material as a short fat square tube of steel or a short fat block of some mineral or concrete.

Edited By Niels Abildgaard on 13/06/2023 15:26:53

Ian, could you use these results to move the headstock and retest? Also, a set of tests at the top of the bar rather than the side would check the other variable. I don't know just how sensitive a level would be needed to find any twist in the length of the bed. I have two Moore and Wright levels, I will try testing them for sensitivity and if one looks promising, I could lend it to you.

Edited By old mart on 13/06/2023 16:26:02

Edited By old mart on 13/06/2023 16:29:14

Edited By old mart on 13/06/2023 16:33:46

To experiment with bed twist on a small lathe just put the top bar of a large F clamp in the bed slot to the right of the tailstock or similar contrivance to arrange a large horizontal lever that twists the bed. Hang a brick on the lever while observing your indicator.

Bazyles method would not need the level and is a good idea. I added a block of aluminium to raise up the headstock of the Atlas12 x 24 lathe to allow room for adjustable "levelling" at the tailstock end. So far the lathe bed has not needed any adjustment and is very accurate with the tailstock end loose.

An Engineers Level (0.02mm / 1000mm if I recall) is one of the many things I've bought because, 'some day I'll need it'. Perhaps this is now!

I'm out tomorrow and need to finish a project, so it may be a few days before I get back to measuring.

I've now got the lathe back together (not without a couple of scares!) and cut a test bar. However, I'm not sure I particularly trust what I've written down, so I will have another look later in the week and report back.

Thanks all.

Iain

You will have to sit your level on a suitable parallel to clear the prismatic bed, and shim the headstock end temporarily to get the bed close enough to see the bubble. You might find a twist, or the bed might be perfect.

Iain, You have a PM.

The important thing is that the level gives the same reading at both ends of the bed, to indicate a lack of twist.

It is of much less importance if the lathe leans slighty towards or away from you.

Chances are that a lathe sitting on an uneven bench will show twist. (Probably, once the casting was machined, it will have started to stress relieve itself, and twisted anyway )

A used machine might well have been fixed down so as to , inadvertently, twist the bed, and it has stayed, at least partly that way afterwards.

Hobby machines are designed and built down to a price, so the casting is unlikely to have been stress relieved before machining. If a new lathe, out of the box, can be out of line when checked, stress relieving will not have been high on the list of priorities.in that factory.

Things would be different with a high quality high precision industrial machine; and reflected in the price!l

Having put the lathe back together, I've turned down a piece of (what I thought was ) EN1A and then drill and started to bore. The outside is kind of OK, but the inside is a mess. Screeching and an almost knurled effect.

I've tried another piece of steel with similar results. In some ways this is worse, The cutter is being seriously deflected. I suspect the second piece of steel (origin unknown) may be tougher than my little lathe is happy with (boring anyway). The first deflects less but resonates badly.

I would like to suspect the steel rather than what I've done with the headstock and re-assembly, but wanted to seek some more expert opinions before I strip the lathe back down again...

Help!

Iain

If the OD turns smoothly, and parallell, the lathe would seem to be OK, so no need to strip it.

Why risk upsetting things now that you have got it aligned.

Since the problem seems to be confined to the boring operation, it could be the tool that is the problem.

I would check: mounting (Boring bars are rarely as rigid as ordinary turning tools, being longer, and often thinner)

Rake angle?

If the boring bar is on centre height, are the clearance angles correct?

Is the speed that you have previously used for this tool appropriate? Chatter usually means a lower speed is needed, especially wherev the tool is not rigid.

Using a boring bar, the depth of cut is likely to be smaller to cope the reduced rigidity.

The "knurled" finish suhhests that the tool is vibrating badly, digging in and out of nthe cut.

Howard.