lathe query

[japfanParticipant @japfan]

Hello all, new to his forum,

 

several of my books on model engineering state that with a good lathe the headstock is set up so that the crosslide will cut in a very faint concave ( depressed) attitude across the work – I still don’t understand why this should be necessary, can anybody explain the reason for this phenomenon, and any advantage it has?

 

many thanks

Dave c

 

 

[japfanParticipant @japfan]

[Ex contributorParticipant @mgnbuk]

I was taught that it was to allow two faced-off parts to sit together without rocking – two concave parts contact around the outer edge & sit flat together, whereas two convex parts contact in the centre & rock.

 

Regards,

 

Nigel B.

[John HaineParticipant @johnhaine32865]

Tolerance, in a word.

[WolfieParticipant @wolfie]

How does that work then? Is the headstock tilted by half a degree or something? That implies that if you go all the way across or even just further than the centre that you would have a sloping face?

 

And welcome Dave, these guys are great at answering all sorts of silly questions, really helpful.

Edited By Wolfie on 11/01/2012 09:24:05

[blowlampParticipant @blowlamp]

No, it’s nothing like half a degree and it’s the cross slide, not the headstock that is inclined.

It’s more like a thousand of an inch in four or 5 inches, but usually less in better lathes and is done for the reasons stated by Nigel B.

 

If the maker of the lathe could somehow achieve an exact 90 degree relationship between the bed and cross slide, then in time, some wear would result in uncertainty as to which way the lathe would now be facing-up i.e. convex or concave. So the maker just builds in a small safety margin to ensure it’s always concave.

 

 

Martin.

 

[TerrydParticipant @terryd72465]

Hi Japfan,

 

Just think about it, would you rather a bottle with a concave bottom face or a convex one? As Martin says, it is only a tiny amount by the setting of the cross slide. If the headstock was offset your ‘parallel’ tuning would be tapered

 

Best regards

 

T

[japfanParticipant @japfan]

many thanks all, NOW I understand

 

 

cheers

 

Dave

[John McNamaraParticipant @johnmcnamara74883]

There is a little more to this story machine alignment is an art.

 

If the headstock was angled slightly so that it created a slightly convex surface when facing. The axis of the spindle is pointing away from the operator, and is not parallel to the bed.

 

This means when turning a shaft using the same settings, you will cut a taper that is bigger at the tailstock end. certainly not Ideal; particularly when fitting another member over the shaft. It will not fit when you test it, but the shaft is undersize further towards the headstock.

 

The natural tendency of the work to spring away from the tool will exacerbate this. making the shaft even bigger at the tailstock end…..Assuming you have not used a tailstock centre and the work is just held in the chuck.

 

I did quite a lot of reading on this before I worked on my own lathe… a lot of scraping and grinding then the saddle was rebedded using epoxy bearing material; a years work. From reading various texts there was consensus that for “Tool room quality” the headstock should point towards the operator by a maximum of .0005″ inches over one foot. (300mm) and a minimum of zero.

 

It the headstock spindle axis should also be truly horizontal or point slightly up but not down. Gravity is always trying to pull the unsupported end of the work down so slightly up is good.

 

Measured zero compared to a workpiece fitted in the chuck, will and never can be zero due to gravity pulling the work down at the unsupported end.

 

In most of the accounts and books on alignment for tool room lathes I have read the ideal was not zero but .0005.” inches to counter gravity and deflection of the work by cutting forces. Depending on the text a few used .0003″ for one or more of the settings.

 

(So called “Production lathe’s” are allowed to have double the error of “Tool room quality” lathes)

 

The same tolerance was applied to facing.

 

The headstock is not the only factor here the cross slide must also be truly 90 degrees in relation to the bed. Until this is achieved it is pointless adjusting the headstock.

 

I reality nothing is perfect, alignment is a set of compromises, each error must compliment the others.

 

Connolys book Machine tool reconditioning is a very good reference.

 

Somewhare I have a link to a government standard for machine tools, Just cant find it hmm…. Maybe another MEW member can post one.

 

Cheers

John McNamara

Edited By John McNamara on 12/01/2012 13:43:30

[blowlampParticipant @blowlamp]

John.

 

Sorry to take issue with you, but the cross slide is not set at a true 90 degrees to the bed.

 

It is set at a very slight angle so that it generates a slightly concave face to the component and has nothing to do with headstock alignment per se.

 

The lathe bed is the reference for the alignment of the other major components and the fastest way to get confusing errors, is to use as a reference, something that’s been referenced to something else first.

 

 

Martin.

Edited By blowlamp on 12/01/2012 14:21:49

Edited By blowlamp on 12/01/2012 14:26:04

[John McNamaraParticipant @johnmcnamara74883]

Hi Blowlamp

All measurements are taken against the bed.

However in order to meet tool room spec both the cross slide and the headstock are interrelated.

Hence my comment “alignment is a set of compromises, each error must compliment the others.” I wanted to set the headstock point towards me. Within the limits of my measuring equipment a square I wanted the saddle to be square to the bed.

It makes sense to me to use the headstock setting pointing 5 tenths towards the operator over 300mm as it counters the deflection of the cutting forces, and reduces the possibility of making the tailstock end of a bar bigger than the chuck end when plain turning.

When I set my lathe up I did the saddle first by lapping the cross slide until there was no binding with even hand wheel pressure throughout the travel (It was very good to start with so only a few hours work). obviously it mast also travel a straight line. in this case there was slight wear in the centre and maybe .001 to remove from the less used end that caused a slight bind when fully retracted towards the operator as you would expect. I can still see the manufacturers machining marks even after the tidy up.

The bed was a different story the lathe had started life as a copy lathe The hydraulic copy attachment does not move the cross slide hence little cross slide wear. and no leadscrew wear!, feed is a separate shaft. I don’t think it ever cut a thread. The bed and saddle were worn hence the rebuild.

 

Continued….

Edited By John McNamara on 12/01/2012 16:52:07

[John McNamaraParticipant @johnmcnamara74883]

After the cross slide was attended to the saddle was rebedded using Epoxy bearing material…. Search for MOGLICE and you will get the idea. However i used my own mix (Posted on this site under epoxy bearing material.)

You have attach adjusting jacks to the saddle, working on the inverted V’s of the bed. I used the M6 wiper cover holes to attach a 12mm plate that holds the lock nutted jack screws so did not have to drill, a fair bit of work to make, these were used to set the saddle in the exact final position and height it will finally rest in. once all the testing is done you lift it off, apply the bearing material then set it back on the bed (which has been coated with release otherwise it will be stuck fast.) That was 4 years ago now, still perfect.

To set up the saddle I used an 18″ reference grade square used from the front and the back of the bed, to reduce the error. I measured running the cross slide in an out with a dial indicator attached against the edge of the square. The saddle was set as high as possible bringing the leadscrew back up into proper line and the feed pinion into full engagement; no more slop in the handwheel.

I also set up a 300mm ground steel test bar in the headstock 3 jaw, (it does not have to be dead centre) then using an indicator (It must be a good one to .0001 accuracy and reliable not “sticky”) mapped the min and max at 180 degrees every 20mm along its length; using the saddle which I had to manually push along the bed as it was on jack screws) You end up with a set of pairs of measurements. If you chart the difference between each pair of measurements, Excel makes it easy to visualise, you will get a very accurate map of where your headstock is pointing in relation to the bed. a tiny shim was all that was needed to position the headstock.

It was only after I got the required accuracy of both measurements (together with the rise of the spindle in the vertical plane) that the saddle was set.

In Connolly’s book he mentions to methodologies for setting up a lathe; headstock first or saddle first. I guess there are differences of opinion.

This was one of the most satisfying jobs I have ever done.

Cheers

John

[blowlampParticipant @blowlamp]

It sounds like you achieved your aims John, but I wouldn’t necessarily recommend some of your methods to others.

 

The accepted practice to ensure that a slight hollow is applied to a faced component, is to angle the cross slide in relation to the bed in the way I have already mentioned. This also ensures that an item that is faced-up between centres, will also be a little hollow too.

 

The normal factory practice is to use an accurate jig which clamps down securely to the lathe ways and has a reference face at 90 degrees to the bed so that the cross slide movement can be clocked against it.

 

It’s done like this in case the headstock axis turns out to be at, or very close to, zero horizontal error in relation to the bed.

 

All the above assumes the bed is unworn and is set up without twist.

 

Martin.

Edited By blowlamp on 12/01/2012 18:28:41

[Ex contributorParticipant @mgnbuk]

After
the cross slide was attended to the saddle was rebedded using Epoxy
bearing material…. Search for MOGLICE and you will get the idea.
However i used my own mix (Posted on this site under epoxy bearing
material.)

 

Good luck using that method – I hope that it stays attached to the saddle. Most of the machines using this method that the company I used to work for came across during refurbishment had come adrift from the casting. It looked fine – just a “quick scrape” to get it bedded to the reground bed – but tapping with a file handle would inevitably produce to dull, hollow sound indicating that the coating was no longer bonded to the casting. On some machines the coating could be removed with a wallpaper stripping scraper – it just shelled off. The problem appeared to be oil from the central lube system forcing itself between the coating & the casting where the feed holes were drilled through both. On the only one I remember being attached & re-usable (a Scharmann horizontal borer IIRC) the other “downside” comes in – the stuff is very hard on scrapers due to the fillers in it. We had to get 2 boxes of 10 Sandvik carbide scraper tips for the fitter – while he was using one box, the other box was away for diamond lapping

 

We replaced the coatings with PTFE based tapes. We used a variety – Guidecoat and Rulon mainly, but also a hard white plastic sheet called BiPlast (had to be careful with that one – it wouldn’t run without oil & had to have a high surface loading or the friction was very high). We occasionally came across this type of coating that had come un-bonded, but pretty well every Moglice coated machine had to be re-done due to oil induced releasing of the coating.

 

And don’t believe everything you read in the manufacturer’s bumf about “setting up a saddle on a bed with sealing strips & injecting the coating …..” If you have an hour or two sometime I could tell you about the trials & tribulations of a now long gone UK milling machine maker trying to series produce CNC VMCs by injecting such a coating – but I am not about to do that on an open forum !

 

Regards,

 

Nigel B.

[Sub MandrelParticipant @submandrel]

Accuracy in machine tools : how to measure and maintain it:

a
practical manual for production engineers, foremen, fitters, and
inspectors interested in the use and maintenance of machine tools

 

I just took a look for this book by Georg Schelsinger on Google Books, its the origin and explanation of how deliberate errors in directions that get cancelled out in use are used to minimise the errors in the most accurate lathes. The ‘Schlesinger Limits’ are tolerance suitable for different types of use.

 

Unfortunately it looks like the book isn’t out of US copyright yet (1942).

 

More digging, the tests are earlier than that book. The German Wikipedia has this:

 

http://de.wikipedia.org/wiki/Georg_Schlesinger

Google translated it for me, but it doesn’t say anything about the limits, although its the right man.

 

 

Neil

Edited By Stub Mandrel on 12/01/2012 21:28:41

[SpringbokParticipant @springbok]

Yes I agree with blowlamp

 

Bob

[Author]

Posts