Accuracy or lack off - what to expect

[Admiral_dk]

Hi All

I'm frustrated with my tools .... My lathe has a Z-Axis accuracy that I'm happy with + it can be locked, but the Y-Axis has me very frustrated. If I put a mike on the Y-Axis and try to push or pull as hard as I can with my bare hands - I can move the Y-Axis some 0.4mm according to the DRO 

As this is in Diameter mode, it equals to a movement off 0.2mm. ~ 0.08" and it really explains why I can do deep cuts and have a very nice finish and then try to take a very thin finishing cut that is all over the shop => in a horrible finish .... 

I tried a number of things and reduced the problem some - but I have a Acme-Thread on the Y-Axis where the nut can move 0.09mm ~ 0.035" on the axle (that can't be removed).

I had a look at some suppliers of Acme and Ball-Bearing Spindles and they will not promise any better than 0.05mm ~ 0.02" accuracy on an Acme Spindle .... But as we all know - Accuracy and Backlash isn't the same. If I had zero backlash (I wish), I could be rather sure that the diameter I set the Tool on the DRO, would result in a diameter on the subject that will be extremely close to the diameter I want (if I haven't made any mistakes).

What kind of Backlash do you find acceptable / expect ?

I can see that my new Mill have Acme-Nuts that are sawn 2/3 through and two small screws to adjust the backlash more or less away - works very well on X, but not so well on Y, and Z is so heavy I do not expect to see any here. I haven't tried to adjust Y on it yet, as it is a pain in the proverbial to get access to do so.

Is a similar thing a kind of cure for the lathe or what do you think ?

Per

[Charles Lamont]

Firstly, there seems to be something wrong with your decimals. 0.2mm is about 0.008", not 0.08".

Secondly, that is not an unreasonable amount of backlash for a machine in reasonable order. Feedscrews wear in the most regularly used portion. My elderly mill has about that much on both x and y, and although it has anti-backlash adjustable nuts, that is about the best I can do without making the screws too tight at the extremes of travel.

Thirdly, given that with normal screws and slides, some backlash is inevitable, learn to live with it. Make sure the slides are well adjusted. Feed in one direction, and if you have to go the other way, wind back well beyond the backlash and then approach again. Lock the stationary slides when taking a cut.

If you need extra precision, use a dial gauge (and slip gauges, AKA Jo-blocks) to check the actual movement.     

[steamer]

I agree with Charles.    Remember that there is a very old saying...." Everything is made of rubber"....  A machine tool will deflect under load, and as such you have to work with that deflection.

Instead of taking a bunch of heavy cuts, and then a light one....try taking even steps  to your diameter.

So lets say you want to remove 3mm from the diameter of a nice stout work piece...you could take 2.5mm, and then sneak up on the last 0.5....OK...but during the first cut you put some spring tension in, and as you try to take the " light" cuts, some of that tension is released as you cut....making the cut deeper than you intend.   This often results in overshooting the dimension, and is frustrating.

Try the following
with a 25 mm or so piece of bar in the lathe, ( use a free cutting steel) and a sharp suitably sharpened lathe tool  ( positive top rake)

on a 3mm cut......
Cut 2mm on diameter.
measure the result in several places, and write down your results
Go back to your start and take a dead pass.....meaning do not dial in any more cut....just leave it where it is
Observe, and measure the cuts again

That is a "dead pass" or a spring pass, and is an indication of of how much deflection is in the system during the cut.

Now take a 0.5 mm cut, and a dead pass

How much spring did it have with that DOC?....

Now ...how much is left?....( should be something like 0.5 mm...maybe a little less....)

Dial in the difference, and take the cut.....measure where you are.

Now we know, from the previous pass, about how much we will take with a dead pass....you should be near the difference between that cut and the dead pass.

The point is ...to take equal depth cuts   once you start getting close...not to take smaller and smaller cuts each pass....as with that method the amount of spring in the system is constantly changing.  So it ends up doing something you don't expect.

Now....if I'm roughing....
I may take a .125" DOC per pass until I get to within 0.125" of the size.. I'll let it cool for a half hour or so.

Then I'll start taking 0.025" cuts, in equal steps, from there with some dead passes.....Then when I go for the last pass, I have some confidence that the system will behave like the  previous 4 or 5 passes.    Just divide the remainder up into equal depths of cut....That way when you get to the last past, you can dial in the remainder, and you'll probably nail it.

If you're running coolant, you can attack all the way down to size...taking equal cuts...until you get to size.  Same thing but usually you can be far more aggressive.

A smaller tip radius on the tool, and more top rake will usually result in less cut pressure, and less spring....on light lathes, tool geometry is very important, and usually, our light lathes will not tolerate negative rake carbide tooling very well.    HSS works well there.....horses for courses...

Dave

[steamer]

Oh and my newly rebuilt SB....which is 8 years old now, has about .020" backlash in the cross slide....I suspect some of that is the screw, as the nut is " new"...er ah 8 years old, while the cross slide leadscrew is original and from 1952....but no matter.....I can hit .0005" on diameter pretty easily.....

but I'll tell you ....put a dial indicator on the spindle and tension the belt....and wow!....it moves a lot!!!

Big lathes just dont move as much...because they are so much more heavily built....

Dave

[Jasonb]

What lathe is it?

What cutting tools are you using, an insert may work well on a reasonably deep cut but not so well if only very shallow in which case try to make the final cut of reasonable depth.

I have a lot more backlash than that and get an acceptable finish and can work to 10ths of a thou if needed.

[Jo]

The old trick with worn nuts is to only cut in one way. Adding a Lock on the saddle is simple enough to do.

But I get the idea you are saying the nut itself is moving??? Is that correct? If so you would need to push it to one end of the movement and then find a means of keeping it there  Could you provide photos of the problem?

What kind of Backlash do you find acceptable / expect ?

I have DROs to my two larger lathes I have not noticed any Backlash on either But some people can make do with unacceptable levels of backlash that cannot be gotten rid of and still knock out working models - they have learnt to work around the failings of their machines ¹  .

Jo


P.S. 1. As you get older and have sufficient finances behind you, you may not want to continue to be one of the "Make do and Mend" generation  . 

[Charles Lamont]

Looking at this again, your poor finish on finishing cuts is probably nothing to do with backlash. There is backlash of several thou  on the cross-slide of my Myford, but I can take a final sizing pass 1/4 thou deep (~ 0.005mm) or less (without locking the slide). But I only use carbide tools for cast iron. For practically everything else I use HSS tools, and have learned how grind and hone them, and set them accurately on centre height. Cutting lubricant is another thing to look at.   

[Ramon Wilson]

Hi Per,

Backlash as such is something that all standard leadscrew and nut machines suffer with.

Forgive me stating the obvious but this is usually due to wear either on the screw or within the nut - usually the latter but mostly a combination of the two. If the leadscrew can move laterally to any great degree in it's bracket or the nut within it's mounting however that is a somewhat different situation. There has to be a degree of clearance in the leadscrew bracket but this can usually be adjusted to a minimum condusive to the friction of overtightening. The nut should not move in it's housing however. Insufficiently tightened gib strips will exacerbate the problem too.

Backlash due to wear can always be taken out by making sure the slide is always being pushed but if the wear is in the mounting of nut and/or leadscrew bracket you have no way of ensuring the tool tip is where you think it might be. With a heavy cut the forces involved will assist in pushing back against such potential movement but on a light cut the 'slack' of the system could affect the tool position on a light cut and hence variation in results

First off for me would be to check those areas first and, if possible, reduce lateral movement to a minimum. Any backlash in the screw thread itself being taken care of by the usual method

Making sure the gibs are adjusted to give a firm hold and using a lock (though if there is slack in the system a lock can sometimes have varying results in it's clamping effect) are also factors to consider though I rarely use the lock on the cross-slide whilst actually turning. Possibly a typo on Jo's part but a lock on the saddle will have no effect on the Y axis.

Without checking I'm not sure what the exact backlash is on my Myford but at 40 plus years old its not anything that causes problems - never really think about it and no, I don't have DRO on the lathe. As I sit here I would guess about 8-10 thou, possibly less

Hope that adds a little to previous comments - good luck with it

Our regards for a better New Year to you and John

Tug

PS I would agree with Charles on using HSS particularly where finish is concerned but then I'm a bit long in the tooth to change spots now 

[Admiral_dk]

Ups - yes I wrote the wrong decimal points on the Imperial measurements.

What really annoys me, is those cuts where I leave / do not touch the depth of the cut. It might start out very well and just make a very nice surface for a while and then it starts to cut deeper and I can see this on the DRO too. I have tried to adjust the gibs to the point where I need to put force into turning the handle (not violence) and the same thing happens if just a smitten better.

The reason I know how well it (most of the time) does roughing cuts, is that I have a good amount off 32mm aluminium I got for free years ago, and that I often need a part that is 10-20mm. So I do a few where I remove 2.5-3mm. per cut and they are usually with a nice surface. Getting closer I have tried to remove 0.6mm. per pass according to the DRO only turning the handle in one direction and then I see the tool moving on the DRO and a horrible surface on the part. I have tried to do smaller depth per cut, but the amount of change when it decides to misbehave is not constant with same depth and it might do most of the cut at same depth (making me thinking I got it right this time) and increase the depth by anything between 0.002-0.2mm. DRO readout    This is without me touching anything but having a finger close to the Z-Feed half nut lever.

My problem is that it's never the same result - if it was - I could learn how to optimise the way I use the lathe.

I'm sure I could be better at grinding HHS and that is probably why I usually get much better results with inserts -> I have recently bought new holders and plates I haven't tried yet as I wanted to solve the backlash / movement problem first.

[Jo]

Silly question: how much play is there between the saddle and the bed? Is the saddle crabbing?

Jo

[Admiral_dk]

I believe that this is a valid question Jo - but I must be sure that I understand you right.
You suggest that the saddle isn't moving in a straight line, if I understand you right ?
How would you measure it - both play and "crabbing" ?

The lathe has one "prism" as part of the bed, that should keep the saddle running straight and I can't feel any (not very scientific) movement if I try to move it. I can feel and see the movement of the cross slide.

[Jo]

I believe that this is a valid question Jo - but I must be sure that I understand you right.
You suggest that the saddle isn't moving in a straight line, if I understand you right ?
How would you measure it - both play and "crabbing" ?

Play can be measured with your dial on the front/back of the saddle and you try pulling it left and right.

The crabbing is where the saddle when moving to the left the front of the 'loose' saddle is pushed to the right then moves to the left for no apparent reason and the lathe tool digs in  The amount of 'looseness' is not visible but only noticeable due to odd dig-ins of the tool bit 

Jo

[Ramon Wilson]

Per - If the cross slide resistance is as you say - tightish to turn then something else is definitely moving if the tool is being pulled into the work piece as you describe. This could be as Jo suggests a movement in the saddle on the bed or elsewhere - top slide or tool post for example - where exactly though remains to be found. A tool pulling itself into the work piece suggests such freedom or a push off in the first instance followed by another pass which then digs in, pulling the tool in.

There's movement there somewhere but difficult to say from here without seeing it in action Video perhaps?

Tug

[Laurentic]

Per - I was thinking along the same lines as Jo and her crabbing saddle, something sounds like it is not quite right with your lathe. 

Normally I know there is backlash or 'lost motion' in my lathe cross slide and ensure it is taken up for the direction I am applying cuts and all is well; to see a tool moving on the DRO and then see a horrible surface finish indicates to me that something else is perhaps not 'tight' and allowing a varying movement, but again strange as you would expect that on all cuts not just light ones, unless heavier cuts force whatever is moving to move itself "back in its box" as we say. 

At the risk of teaching granny to suck eggs, have you been round all aspects of the saddle, cross slide and top slide and ensured all is tight?

Chris

PS - while I was writing it seems Jo and Tug posted along the same lines!

[Admiral_dk]

I have just been trying to measure the saddle with a "clock" with the magnetic holder on the bed, I can lift the back (part furthest away from me) 0.02mm. and the "front" between 0.03-0.08mm.up from the bed depending on where i place the saddle between the chuck and tailstock. The back sounds fine to me, but should I expect the front to be smaller - it is an old lathe.

I think I might have read your reply slightly wrong Jo - or more correct, I started measuring before I saw your reply. I read you as trying to twist the saddle - is this right ?