Lathe facing convex or concave

[MICHAEL WILLIAMSParticipant @michaelwilliams41215]

(1) One strange thing about manual lathes is that they are not usually easily adjustable for alignment .

Some little watchmakers lathes have swivelling headstocks and swivelling cross slides – adjustable in seconds . So if (eg) facing is not turning out as expected quick tweak and all is well .

Perhaps the basic workshop lathe needs to be very solid so that means hard to adjust but there might be merit in having a variable angle quick adjust cross slide .

I’m sure there are workshop lathes out there somewhere with this facility but I’ve never seen one .

(2) One place I knew had facing lathes with self adjusting spindle speed – as radius of point of turning decreased spindle speed increased proportionally .

Frightening beast because it was by no means a small machine and speed were very high near centre .

MikeW

[Phil WhitleyParticipant @philwhitley94135]

At the risk of starting an international incident, a lathe should face at exactly 90 deg to it's axis, and the face turned should be perfectly flat. How close the manufacturers get to this will depend on whether it is a Harbor Freight or a Hardinge. When we face bar off for appearance, we do not lock the saddle, when we are going for more accuracy we lock the saddle, perhaps check and adjust the cross slide gibs (gybs) a little on the tight side. When we have done this we have taken the machine tolerance and clearances as far back to the manufacturers tolerances plus the wear, that we can. We then face off at the smallest tolerance that particular machine can achieve, without going to huge lengths to achieve the last few (some would say irrelavant) tenths of a thou. It is pontless to continue this type of argument unless someone can come up with a reasoned argument as to WHY a lathe should face slightly convex (or concave). In a wearing situation the faced part will present its outer edge for wear first if concave, and it's centre if convex, but in the long run, it will make no difference. A flat surface will present all to a wear face, and thus wear less. The ideal of precision machine tool manufacture is to make a mchine that is PRECISE, and able to maintain that precision by adjustment as it wears. How near manufacturers get to this is generally reflected in the price of their products.

Phil

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Phil Whitley on 25/02/2014 12:28:52:

… It is pontless to continue this type of argument unless someone can come up with a reasoned argument as to WHY a lathe should face slightly convex (or concave).

.

Phil,

I am aware of NO credible reason why a lathe should be set up to face Convex [and Bazyle has already admitted to his typo].

The preference for slight concavity is purely a matter of expediency … Flat is "perfection", which is unattainable.

WoodTurners and Potters have understood this for centuries; why do we find it so difficult?

MichaelG.

[MICHAEL WILLIAMSParticipant @michaelwilliams41215]

Perhaps it’s because the Earth isn’t exactly flat but sightly concave to prevent people falling off ?

MikeW

Edited By MICHAEL WILLIAMS on 25/02/2014 13:01:19

[jason udallParticipant @jasonudall57142]

..then you add a compound slide….

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Bazyle on 24/02/2014 20:23:50:

The 'what did you do today' thread is getting a bit waylaid with discussion of whether a lathe should face flat or convex. So I thought discussion could continue here as it may be of interest to others doing a search in the future.

.

Good idea … Shame it didn't work out.

[Neil WyattModerator @neilwyatt]

The 'credible reason' was given by a poster above:

"I guess that is why many books on the subject suggest we err slightly away from perfect in the direction that is most likely to help us produce good work."

That is the principle behind the Schlesinger limits. One other sage poster observed changes in temperature throw everything out at these levels of precision. Setting up to chase tenths without a temperature controlled toolroom is running after rainbows.

Neil

P.S. It was Edgar T. Westbury, who seems to think the machinist makes a bigger difference to the end result than the machine. Looking at my own work, I fear he's right.

[John McNamaraParticipant @johnmcnamara74883]

Hi All

I dug up a couple of images taken when I mapped the headstock alignment of my lathe

The bar is 350mm long x 25mm thick ground finish

Note the three jaw chuck. A 4 jaw would be little better you can align it to the axis near the jaws but it will almost certainly point off axis as you measure along the bar. So a reading was taken at the high and low point by rotating the chuck and the difference was calculated by subtracting one from the other. I used an excel spread sheet to plot the error. as shown below. I took a reading every 5mm The indicators are mounted on a special measuring jig that rides one V way via 4 dome end ground set screws and one screw on the flat unused portion of the tailstock flat way. It was fabricated from scrap. held down by a lump of die steel

The microscope was used to check the straightness of the bed using a stretched music wire (exactly) aligned parallel and attached to the bed at each end I have a photo of the adjusters I made somewhere! The wire position was read by positioning the wire in the centre of a small slot using a .0001 micrometre head.

I measured the X and Y plane at the same time. for the vertical axis you need to allow for sag, a bar of that size will sag a little over .001" inches over 300mm

You can perform the same measurements without a jig just using the saddle as the sliding member. In this case I wanted to check the bed without the averaging effect of the saddle in order to plan how I was going to recondition the machine. But that is another story.

Regards
John

Edited By John McNamara on 25/02/2014 15:02:43

Edited By John McNamara on 25/02/2014 15:06:59

[Gordon WParticipant @gordonw]

Nothing can be made "exact" A lathe is set up to face flat, any error, and they must be some, is on the concave side. That's the standard. I've seen operators leaning on the headstock of big lathes, although I don't believe it did anything. On my little lathe I've been known to put a bit of pressure on a long shaft to get rid of a slight taper.

[John Stevenson 1Participant @johnstevenson1]

Posted by Michael Gilligan on 25/02/2014 12:43:22:

The preference for slight concavity is purely a matter of expediency … Flat is "perfection", which is unattainable.

WoodTurners and Potters have understood this for centuries; why do we find it so difficult?

MichaelG.

Why, because we work to thou's and they work to centimetres.

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Neil Wyatt on 25/02/2014 14:59:13:

The 'credible reason' was given by a poster above:

"I guess that is why many books on the subject suggest we err slightly away from perfect in the direction that is most likely to help us produce good work."

That is the principle behind the Schlesinger limits. One other sage poster observed changes in temperature throw everything out at these levels of precision. Setting up to chase tenths without a temperature controlled toolroom is running after rainbows.

Neil

.

Neil,

Whilst I accept that there may be the occasional instance where a convex face is helpful; I feel confident that in the vast majority of situations, concavity is "the direction that is most likely to help us produce good work." [and presumably Schlesinger was of that opinion, since he specifies flat-to-concave.]

As a "default setting" convexity would be troublesome in many situations … particularly where mating surfaces are produced on the same machine. These would wobble, or at least leave a gap around the edge: By contrast, a slight concavity gives the impression of perfect fit, and no wobble.

MichaelG.

Edited By Michael Gilligan on 25/02/2014 16:43:53

[I.M. OUTAHEREParticipant @i-m-outahere]

Sorry my previous post should have said concave not convex

I usually proof read before posting but for some unknown reason the website kept re directing me to the log on page and it took 3 attempts to get it to post .

Ian

[jason udallParticipant @jasonudall57142]

Mmmm …what impact on parting off…especially if using rear slide. …

[Another JohnSParticipant @anotherjohns]

Edward F connely In "Machine tool reconditioning" 530 pages

Recommends (For a tool room quality lathe) the end of a 12 inch test bar points .0000 to .0003" towards the operator

Quick post before a meeting:

I think you mean "towards, or away from" the operator; there is a little minus underneath the plus, at least in my copy of the book.

Ie, straight down the line, as straight as you can get it.

John.

[Michael GilliganParticipant @michaelgilligan61133]

Posted by jason udall on 25/02/2014 17:57:34:
Mmmm …what impact on parting off…especially if using rear slide. …

.

Jason,

You make an interesting point …

Normal parting off should be no problem [we're talking very slight concavity, and anyway Schlesinger's argument holds], but I don't think he considered the use of the rear parting tool.

[quote] The headstock should be so aligned that an arbor inserted in the spindle nose rises or inclines upwards only at its free end with respect to the bed ways, whilst inclining in the horizontal plane towards the tool post only. This will counteract the deformations resulting from the weight of the workpiece and the cutting force. [/quote]

MichaelG.

[Frank.N StormParticipant @frank-nstorm18349]

Allow me the small but important remark that the pic in the posting by John McN. (28.2; 08:18) shows measuring the direction of the spindle axis, which can point towards or away from the operator.

No one with his seven senses intact would correct a facing error by angling the headstock!!!

Facepalming,

Frank

[John McNamaraParticipant @johnmcnamara74883]

Hi Frank

I remember reading an article on Shaker furniture makers and how they made such finely turned and crafted chairs, Tables and settles to name a few. However as only their God was perfect each piece was made with a tiny flaw, to avoid showing disrespect. To set the record straight I am not religious.

The story came to mind when thinking about engineering perfection, we seek it but never achieve it. No matter how hard we try there is always a small error, it is just a matter of degree, Wayne R Moore In his book Foundations of mechanical accuracy, deals with the pursuit of it, in his case a crusade seeking it, Both Moore the theoretician and Edward Connelly in his hands on practical book came to the same conclusion. Nothing is perfect. All we can do is place limits on accuracy that are achievable.

When I Restored my lathe a process that took a year, I worked to the limits imposed by Mr Connolly, It has never let me down since
I should point out that When setting up a lathe the headstock is only one part of the total error budget, the saddle cross slide and topside all have to work together, I was only after doing this restoration that that the theory really sank in, Its the total error that counts. Oh and don't forget the Tailstock. And the bed must also be perfect: They never are….. I am very happy with the result. if….. only I could be as consistently accurate as the lathe is.

Regards
John

[Neil WyattModerator @neilwyatt]

An interesting observation. Having discovered that the cross slide ways on my lathe probably are correctly aligned, I faces some 2 3/4" steel yesterday. At first it was distinctly convex -as in very clearly if I put a rule across the face. bear in mind I had recently stripped down and this was the first machining since it was reassembled.

I noticed a little slop had come into the cross slide (I'd been taking 0.5mm cuts with a decent feedrate, enough to find any play!) Tightening the outermost gib screw completely got rid of the convexity, as far as I can tell without bluing the turned face (I haven't finished machining the blank yet).

This is mystifying, as this takes out slack when the cross slide is advanced and shouldn't affect it when it's wound out, but convexity suggests play when it is wound out. Any suggestions?

Neil

[Oompa LumpaParticipant @oompalumpa34302]

"when thinking about engineering perfection, we seek it but never achieve it. No matter how hard we try there is always a small error, it is just a matter of degree, Wayne R Moore In his book Foundations of mechanical accuracy, deals with the pursuit of it, in his case a crusade seeking it, Both Moore the theoretician and Edward Connelly in his hands on practical book came to the same conclusion. Nothing is perfect. All we can do is place limits on accuracy that are achievable."

I have lost a great deal of sleep over the persuit of perfection. Now I just do my best and if I am not satisfied I do it again. I no longer obsess over things as intently as I used to.

[Les Jones 1Participant @lesjones1]

Hi Neil,
I have a theory that might explain it. To make the explanation simpler I will assume there is an equal amount of clearance over the length of the gib. I also assume the cross slide will be overhanging the slides for all or part of its travel. As the cross slide moves in the length of gib to slide contact becomes smaller. As the cutting forces will try to push the tool towards the tailstock it is in effect trying to rotate the cross slide clockwise looking down. This means that the gib to slide clearance will be on the tailstock side towards the back of the lathe and the clearance towards the front of the lathe will on the headstock side. As the cross slide moves towards the back of the lathe the length of slide in contact will be shorter. The angle the cutting force manages to rotate the cross slide will increase because the adjacent side of the triangle formed by the clearance will decrease but the opposite sides length will remain the same. (The gib to slide clearance.) I hope this description makes sense without drawings.

Les.

[Phil WhitleyParticipant @philwhitley94135]

`Ok lets turn this question round(lol) How would all you concave dwellers out there "adjust" a lathe to make it turn slightly concave Without affecting the accuracy of the lathe in any other plane? I think you will find that a lathe should face completely flat, which is about as possible as a slide or a leadscrew with zero backlash. All that is being said in the ancient lore of magical machine tools is " in the time we have got to set up a lathe which retails at this price, get it facing as near to flat as possible, but it is better for the lathe to turn slightly concave, rather than convex. How near you get to flat depends on whether its a harbour freight or a Hardinge. I don't honestly know if my Colchester student roundhead faces slightly convex or concave because it is not important in 100% of the work I use it for. Although thinking about it, A T shirt emblazoned with the slogan "my lathe faces concave" would have a certain appeal to some sections of the home machining fraternity. (pulls on tin hat and heads for the Anderson shelter).

Phil

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Phil Whitley on 21/03/2014 22:42:03:

`Ok lets turn this question round(lol) How would all you concave dwellers out there "adjust" a lathe to make it turn slightly concave Without affecting the accuracy of the lathe in any other plane? <etc.>

.

Phil,

I presume that your question is capricious; but my answer is serious:

Find the fault [or combination of faults] that is causing the lathe to "face" convex, then rectify it accordingly.

MichaelG.

.

P.S. You are obviously aware [although perhaps others are not] that the "target condition" is that the lathe should face flat … The "acceptance" tolerances are simply a recognition of reality; and the permissible convexity is therefore specified as "Zero; plus nothing, minus a Gnat's Whisker".

[Les Jones 1Participant @lesjones1]

Hi Neil (And others),
Ignore my comments in my post last night. When I woke up this morning I realised the picture I had in my mind when I was thinking about the problem was of the topslide. (Set parallel to the cross slide.) Non of that reasoning will apply to the cross slide as the component of the cutting forces along the axis of the lathe would not cause a twisting effect on the cross slide as it is always between the ends of the slide way. (Not past the end of the slide way as it would be with the top slide.)

Les.

Edited By Les Jones 1 on 22/03/2014 08:16:47

Edited By Les Jones 1 on 22/03/2014 08:17:52

[MICHAEL WILLIAMSParticipant @michaelwilliams41215]

Hi Michael and Phil ,

Just a few thoughts :

(1) The target in facing off work in a lathe is simply to obtain a true flat surface as near as possible when specified for the workpiece – don’t see how there can be any argument . Same as target to obtain truly round and truly parallel cylindrical turning as near as possible and truly flat milled and ground surfaces as near as possible .

(2) The oft stated ‘fact’ that two surfaces have to be specifically concave to bed properly is wrong . Flat on flat as near as practically possible is what is required .

(3) At extreme levels of flatness the concepts of concave and convex have little meaning . When down to a level of flatness of the same order of magnitude as surface roughness and of the linearity errors in slides the definition of flatness becomes hard to define and statistical .

(4) At extreme levels of flatness also the workpiece and any mating workpiece are both very flexible . Either or both being microscopically concave or convex really doesn’t matter in most cases of practical assemblies .

(5) Even a nominally very flat surface actually looks like a variable height range of peaks and troughs at microscopic level . Geometry of peaks and troughs can be systematic or random . Bedding of one surface against another depends in part on how one surfaces set of peaks and troughs relates to the other surfaces set of peaks and troughs .

In practical engineering geometry of peaks and troughs can be controlled to a large extent by the machining process so as to give advantages for different purposes .

(6) So larger scale geometric accuracy , small scale geometric accuracy , roughness , lay and component flexibility all have effecst on how surfaces bed together .

(7) This is philosophical perhaps but it’s a question of who’s in charge of a particular design .

In quality engineering if flatness or some other surface geometry matters on a particular workpiece it is DEFINED in the design and on the manufacturing drawing .

To allow an important design features to be randomly modified on the shop floor by someones idea of needing ‘concavity’ et al in the machining process is just nonsensical .

Regards ,

Michael Williams .

[Phil WhitleyParticipant @philwhitley94135]

Posted by Michael Gilligan on 22/03/2014 07:26:50:

Posted by Phil Whitley on 21/03/2014 22:42:03:

`Ok lets turn this question round(lol) How would all you concave dwellers out there "adjust" a lathe to make it turn slightly concave Without affecting the accuracy of the lathe in any other plane? <etc.>

.

Phil,

I presume that your question is capricious; but my answer is serious:

Find the fault [or combination of faults] that is causing the lathe to "face" convex, then rectify it accordingly.

MichaelG.

.

P.S. You are obviously aware [although perhaps others are not] that the "target condition" is that the lathe should face flat … The "acceptance" tolerances are simply a recognition of reality; and the permissible convexity is therefore specified as "Zero; plus nothing, minus a Gnat's Whisker".

Thank you Michael, you have got my point exactly! Everyone on this thread should read and inwardly digest the very wise PS quoted above. I think in this case OCD stands for obsessive compulsive DTI use. Stop testing and start turning, then if a problem arises, find it and fix it.

Phil.

Capricious? Me? Nooooooo.

[Author]

Posts