lathe cross slide 'dishing'

[unc1esteve]

On another site many people insist that all lathes are purpose built to cut a concave surface when making a facing cut.  001 dish in a 3 inch diameter.  The reason for this is always stated that it allows two cylinders to be stacked upon each other.
They quote the English writer 'Tubal Cain' as one of their resources.  They use some lathe's tolerance specifications.  No documentation, just 'that is the way it is'.
I disagree.  If this is done on purpose then the lathe would be accurate only making this one cut for this one purpose.  "Tubal Cain ends his comments by saying 'or should be', so he is expressing an opinion.  He is not a manufacturer.  I can find no lathe manufacturer's advertisement that states this as a quality of the lathe.  I have been researching lathe design on line but so far have not found any information.  I have asked several sellers of lathes if the lathes they represent are made this way but have not received answers yet.
I understand the concave theory verses the convex facing.
Does anyone have or know of documentation on this subject?
Thank you.

[Fergus OMore]

Will  Georg Schlesinger do you?
Will Martin Cleeve( K C Hart) do you?

And what about Connolly?

the stock answer is 'Ringing parts together'

I wouldn't quite knock old Tom Walshaw, he made rings around a lot of people who thought that they were also 'Men of Iron'. Get his stuff from old copies of ME and see what he had to say.
Sorry, but he came from a Nest of Singing Birds- which we don't have today.

[philf]

On another site many people insist that all lathes are purpose built to cut a concave surface when making a facing cut.  001 dish in a 3 inch diameter.  The reason for this is always stated that it allows two cylinders to be stacked upon each other.
They quote the English writer 'Tubal Cain' as one of their resources.  They use some lathe's tolerance specifications.  No documentation, just 'that is the way it is'.
I disagree.  If this is done on purpose then the lathe would be accurate only making this one cut for this one purpose.  "Tubal Cain ends his comments by saying 'or should be', so he is expressing an opinion.  He is not a manufacturer.  I can find no lathe manufacturer's advertisement that states this as a quality of the lathe.  I have been researching lathe design on line but so far have not found any information.  I have asked several sellers of lathes if the lathes they represent are made this way but have not received answers yet.
I understand the concave theory verses the convex facing.
Does anyone have or know of documentation on this subject?
Thank you.

unc1esteve,

I have a copy of Testing Machine Tools by Dr. Georg Schlesinger. (Schlesinger limits are/were often quoted in machine tools specs). In it are various figures given for the lathe facing accuracy.

"For Finish Turning Lathes": "Lathe faces (hollow or concave only) within 0 to 0.02 per 300mm in dia."

"For Toolroom Lathes (Highest Degree of Accuracy)": "Lathe faces (concave only) within 0 to 0.015 per 300mm in dia."

"For Facing Lathes": "Lathe faces (concave; only for machines with headstock and carriage mounted on common baseplate) within 0 to 0.05 per 1,000mm in diam."

So, all of these are much better than 0.001" per 3" in diameter and, in all cases, they can be zero.

Hope this is of help.

Cheers.

Phil.

[Fergus OMore]

Let's be sure-- all this has absolutely nothing to do with crosslides but the concave facing is inbuilt into the headstock spindle- not the crosslide.

The crosslide doesn't need to be at 90 degrees or 3 thous in three inches or concave to a thickness of a tissue( Horner) or ability to withstand the efforts of a Churchill tank per Cleeve about Myfords to destroy the alignment.

You can chuck the whole crosslide out of the window as it is quite irrelevant for many lathe users. I have two lathe with one and one without! And so did Tom Walshaw and Cleeve and quite a few of us. I 've just got in the post from my good friend 'Miner' about epicyclic chucks mounted either fore or aft as my first reply was sent.

If we want to go off at a happy tangent, it was Cleeve that was instrumental in changing the narrow gib principle used at that time by Myford- as facing- by some of the greats on their machines was getting 'interesting'

It's ALL in Model Engineer- at some point.

[unc1esteve]

Thank you for the swift replies.
When I post a question I try to be as clear as possible but I seem not to succed.
I used the term cross slide because of the answers on the other site.
I am not knocking Mr. Tom D Walshaw.
I did not mention Dr. Georg Sclhesinger because I understand the difference between tolerance and error.
You say the concave is built into the head stock spindle.  This is what I am looking for.  An advertisement, article or comment from a lathe manufacture, supplier or seller that describes this.
Tolerance is accuracy.  A .001 dish is an error.  To tell a novice that this is normal to all lathes and it is for a specific reason is not good policy.

[Alan Haisley]

If the head-stock spindle is twisted so that a facing cut will be concave then a chuck or collet held piece will be taper cut with the collet end larger. It's easy enough to take a light facing cut from the center out so that the outer edge of a piece stands proud of the center if that's what you want.
Alan

[andyf]

I think I agree with you, Alan. If a lathe turns true cylinders rather than slight tapers, its spindle must be parallel with the bed. If the same lathe produces concavities or convexities when the cross-slide is used to make facing cuts, it seems to me that the likely cause is that the dovetail on top of the saddle is not perpendicular to the bed and the spindle axis.

Facing cuts on my Chinese lathe produce far more concavity than they should, and with a largish square held against the saddle dovetail, the Mk 1 human eyeball reveals the cause. If I ever find a 55 degree dovetail cutter at a reasonable price, I'll fix it. Too much metal needs to come off for a scraper to be used, other than for finishing off after the miller has done the hard work.

Andy

[Fergus OMore]

Gentlemen,
                  What you have forgotten( or I think that you have forgotten) is that a lathe is NOT a trignometrical figure. It is relatively a weak piece of equipment which will bend on cutting loads. If one creates a lathe to follow a pure rectangular shape, the cutting forces will create a convex shape. It then that this 'rocking about like a pea on a drum' of two convex faces occurs.

Clearly, I have little or no experience of what happens on Andy's lathe but I suspect that much of the problem lies in the lathe being flimsy.

Am I wrong? Well someone in a distant and nigh forgotten past wrote of what happened  when he stuck his elbow on the tail stock( bored rigid) during a long cut. Again, Thomas wrote at length about the subject in MEWM and earlier in ME. Tom Walshaw wrote extensively on the subject of lack of rigidity and his lathes constantly having to be adjusted - because of earth movement- in his retirement. Again, he wrote of problems of flatness in single point milling and filled Model Engineer with comments from people- who hadn't his maths skills

However- and it is worth thinking about- I 'did' the history of a chief draughtsman of a then very successful machine tool works. He left a few mediocre watercolours of a few lodges, a church and a couple of Scottish scenes and an engraved silver cigarette case.

In real money, it probably amounts to enough cash to fill a car with fuel-- and his wonderful business is - poof- gone- like him.

[philf]

Thank you for the swift replies.
When I post a question I try to be as clear as possible but I seem not to succed.
I used the term cross slide because of the answers on the other site.
I am not knocking Mr. Tom D Walshaw.
I did not mention Dr. Georg Sclhesinger because I understand the difference between tolerance and error.
You say the concave is built into the head stock spindle.  This is what I am looking for.  An advertisement, article or comment from a lathe manufacture, supplier or seller that describes this.
Tolerance is accuracy.  A .001 dish is an error.  To tell a novice that this is normal to all lathes and it is for a specific reason is not good policy.

unc1esteve,

Rather than specifying individual figures for a machine tool the manufacturers usually mention the standard to which they are made.

This is the test certificate which came with my lathe. There are two sets of figures for most parameters - one for the Industrial Quality (mine) and one for the Toolroom Quality version. Actual figures aren't quoted - just that they passed the individual test.





The particular test for Cross Slide perpendicularity to the Spindle Axis (11) has a typo in the metric version, showing 0.15mm instead of 0.015 (the Imperial version is correct at 0.0006").

You will see 4 standards mentioned, ISO, DIN, ASA & Schlesinger.

Is this the sort of information you're looking for?

Sorry Fergus - there are tolerances for the perpendicularity of the cross slide as well as the parallelism of the spindle to the carriage and, incidentally, in Machine Tool Reconditioning, Conelly says "Axis of the headstock spindle to be parallel to the outside ways of the bed in the vertical and horizontal planes". For a Toolroom Lathe a tolerance is given of 0 to +/- 0.0003" over 12" - not unidirectional (although Schlesinger does specify a unidirectional tolerance). Conelly specifies that the cross slide should have a unilateral tolerance of the squareness of it's travel to the axis of the headstock - for Toolroom Lathes 0 to 0.0005" over 12" diameter hollow or concave only.

Cheers.

Phil.

[bp]

To me as a simple ex draftsman used to applying tolerances, all of this indicates that the machine tool design fraternity considered that a flat face was very important, a convex face was very bad news and a (slightly) concave face was acceptable.  The degree of concavity determining whether the machine will be designated "Toolroom" or "Industrial".
cheers
Bill Pudney

[BillTodd]

On another site many people insist that all lathes are purpose built to cut a concave surface when making a facing cut.

They are misinterpreting the tolerance specification.

The target for a tolerance is always zero, normally this is in the middle of the tolerance band i.e. if something is spec'ed for 10mm  +/- 0.1mm then you'd aim for 10.000mm. However, some items must absolutely not be over or under size (like a bearing fit) , so specifications are +0/-X or +X/-0 .

The tolerance for a lathe for facing as previously mentioned is  +x/-0 (concave) . The ideal lathe has 0 concave and 0 convex, but is allowed to be slightly concave by to meet specification.

This means the lathe manufacturer , balancing the need for accuracy against the cost of production, will often play safe and leave the machine with a slight concave rather than do the extra work to get it right.

Bill

[edit] BP beat me to it  - The Bill's are right again 

[philf]

To me as a simple ex draftsman used to applying tolerances, all of this indicates that the machine tool design fraternity considered that a flat face was very important, a convex face was very bad news and a (slightly) concave face was acceptable.  The degree of concavity determining whether the machine will be designated "Toolroom" or "Industrial".
cheers
Bill Pudney

Bill, you may be a "Simple Ex Draftsman" but you have summed the facts up nicely!  (Not so simple after all!) 

I spent 38+ years in mechanical design - starting as a student apprentice when there were 40+ drawing boards and draftsmen/ design engineers in our office and retiring 3 years ago when there were no drawing boards left and only me doing any mechanical design on CAD. I never missed the drawing board. I still have several sets of drawing instruments which will never be used again - just too nice to throw away.

 

Phil.

[unc1esteve]

I am sorry I do not have the communication skills to convey what I was looking for.  Here are some edited quotes from the '7 x 12 mini lathe forum'.

 1.  'Does anyone have an explanation as to why I always get a "dishing" effect on a work surface mounted in the 4-jaw chuck or on the faceplate, quite noticeable on anything bigger than an inch across?The edges come out slightly higher than the middle, like a saucer.'

 2.  'I faced a 3" diam aluminum and got very good finished until I checked for
 flatness. I placed a straight edge on the face and the middle section is
 dished. There is light able to shine in the middle.'

 3.  'I then used the cross-slide to wind the DTI in along the straight edge for 2", until it reached the centre-line. The DTI reading increased by 0.008" over those two inches, which is rather more than a nominal amount. '
 
Answers

 1.  'The concave is made on purpose, common to all lathes, so that two cylinders will stack on to of each other and not rock.'

 2.  'Your lathe is OK. Go, replenish the earth, and subdue your worries.'

 3.  'But a slightly concave face is considered the norm, generally speaking. Just is.'

I do not have the equipment to measure .0005 in 12 inches to see how much light will shine under a straight edge.  I feel that being able to see light under a straight edge in one inch or .008 in two inches is not a tolerance common to all lathes.  These are errors.

To believe that these errors are purpose built into all lathes just so we can stack up two cylinders is believing in a fairy tale.  To tell a novice that their lathe is ok is poor practice.  How are we novices to become craftsmen if given this kind of incorrect information?

What I was looking for was something from a lathe manufacturer to confirm these statements.  Is not the target of a tolerance 'zero'?  In my mind these tolerances contradict these statements. 

Lathes are not designed to face a concave surface but to face as close to 'zero' as possible.  These gentlemen are just not correct.

[BillTodd]

Somewhere on this site is a long thread in which Bogstandard explains how he fixed up a friend's mini lathe. You could do worse than read through that

[edit] here it is, the most popular thread in this section:

http://madmodder.net/index.php?topic=627.0

[Alan Haisley]

It's obvious to me that the reported effect of the cutting followed by the measuring means that either the lathe spindle has enough play that under load it moves or the headstock itself has play that causes the same thing. i.e., the lathe in question is either not solid enough to avoid the headstock twisting under load or the spindle or headstock itself has some play that shows up under load. Since the DTI measurement won't place a significant load it shows the "twist".
The message originator might try orienting the cutter so that the right edge is parallel to the spindle and taking extremely light cuts. If they do that, I would predict that the finished facing cut will show less dishing. This would be consistent with what I think is happening.
Perhaps someone with more experience than I can come up with a method to isolate the movement and  remove as much of it as possible.
Alan

[Fergus OMore]

I think that following some very learned remarks on this, it would be prudent for Uncle Steve to tell us initially what lathe is giving rise to such concern, what tools and accuracy he has at his disposal  and what the results have been given on his machine- at this moment.

I am sure that I will not be the only correspondent here who will read the results with keen interest.

[Jonny]

Quite right Alan and the Bills, spot on.

Too much reading bull and not enough doing springs to mind. We had a saying at work a little knowledge is dangerous.

Heres a link to Harrison users site where the dimension of max permissable runout is stated. 0.01mm over 300mm! Machine came off at 0.002mm runout, hence the term runout not concave or convex, thats impossible unless play in slides or head.
Putting two and two together the term mislead 'concave' is actually a straight edge erring on the side of +90 degrees like a cone, big difference.

http://f1.grp.yahoofs.com/v1/wLSVT8xVkrxA6JOvDGXYf4_Zm-I3es3ErB6yneQAUmYEKbcNu69UZQvenQ_XfgnNLG-PuTltUxPgzz4GkvCgox4LFLEc30cYhVqJpWG08-Nqpro/Z1.%20VS%20330TR%20Accuracy%20Chart/C.%20Harrison%20Accuracy%20Chart%20-%20VS%20330TR

[philf]

Quite right Alan and the Bills, spot on.

Too much reading bull and not enough doing springs to mind. We had a saying at work a little knowledge is dangerous.

Heres a link to Harrison users site where the dimension of max permissable runout is stated. 0.01mm over 300mm! Machine came off at 0.002mm runout, hence the term runout not concave or convex, thats impossible unless play in slides or head.
Putting two and two together the term mislead 'concave' is actually a straight edge erring on the side of +90 degrees like a cone, big difference.

http://f1.grp.yahoofs.com/v1/wLSVT8xVkrxA6JOvDGXYf4_Zm-I3es3ErB6yneQAUmYEKbcNu69UZQvenQ_XfgnNLG-PuTltUxPgzz4GkvCgox4LFLEc30cYhVqJpWG08-Nqpro/Z1.%20VS%20330TR%20Accuracy%20Chart/C.%20Harrison%20Accuracy%20Chart%20-%20VS%20330TR

Jonny,

OK, concave may be slightly misleading in that it may infer a radius and you are correct when you say that the error will produce a cone ....... but, to me, describing it as concave is better than describing it as a cone which could be (for want of better words) "concave" or "convex". The term concave is used in more than one machine tool standard or reference.

My turn to be pedantic now - the Harrison test sheet describes the 0.002mm as deviation from squareness and not runout. Runout is used correctly at G8 - "Run-out of spindle nose centre" where the spindle is rotated and the dial gauge is fixed.

Take, for example, a faceplate faced on the same Harrison lathe without removing it - the runout (with the gauge fixed in the toolpost and the spindle rotated) would be zero (or very near to zero). If the gauge was traversed across the faceplate from the outside (operator) towards the centre it should also show a zero deviation - that is until it gets past the centre when the gauge will start to show the deviation from flat.

As Bill said - in an ideal world everything would be exactly parallel or perpendicular or concentric etc. etc. but manufacturing costs would be unreasonably high - hence working to tolerances and why there are various classes of tolerance for different classes of machine.

 

Phil.

[loply]

I have always been under the impression, and it seems logical, that the perpendicularity of a facing cut is controlled by the perpendicularity of the cross slide dove tails.

Misalignment of the head will ofcourse create the same effect (a convex or concave face), but we have to assume that's not the case.

A poor lathe may have a cross slide which is not perpendicular and will create a convex or concave face even if the head is properly aligned.

Clearly a convex face is going to be a major problem in most situations, whereas a convex one (if it's only slight) is likely to be more tolerable.

[Lew_Merrick_PE]

Back in the dark ages when I was attending University, the Mechanical Engineering department was peopled with Asian and Arabic students to a high degree.  One of those students ended up working for the Indian Ministry of Industry.  In the late-1980's and early1990's he got into the habit of sending me development machine tools to evaluate for him as he was unsure how often the "analysis" he got from his own people was more of the make the boss happy variety than clear, honest fact variety.  Among the things sent to me this way was an original pilot product unit of what is now the Chinese 7X blah lathe sold by many import/dealers.  It was quite nice right out of the box.  Yes, I spent no small amount of time making adjustments, but once they were made, I could turn an 8 inch long part and have it true and round within .0004 FIM.

Fast forward a few years when the 7X blah lathes from China started appearing.  One that sticks in my mind had the axis of the spindle out of true to the bed by more than .010 in 4 inches!  That one was the worst, but I have never had one brought to me where the spindle was true to the bed to a degree that I would accept.  Now, to be honest, I usually one see ones that people are having real problems with -- so this may be an unfair statement.  However, I would argue that the first thing one should do with these 7X blah Chinese lathes is to check the spindle alignment!

I happen to own a MT3 gage bar.  One can be made reasonably at fairly low cost using a MT3 tool holder of appropriate size (I typically use ø.500) with an appropriate length of mating drill rod (silver steel).  Access to another MT3 capable spindle to verify the concentricity of the set-up is in order, but that is usually not that hard to find in a local machine shop -- and shop owners are usually understanding about such needs (at least in my experience).

There is really not a lot one can do to fix problems with a lathe (7X blah or whatever) until you have the spindle aligned to the bed!

[Jonny]

My turn to be pedantic now - the Harrison test sheet describes the 0.002mm as deviation from squareness and not runout.  Phil.

Perfectly correct Phil, i stand corrected

[unc1esteve]

Conclusion

Thank you all for your responses.

 1.  I need to improve the way I communicate my ideas.  I repeat myself.
I was looking for documentation to prove or disprove the statement that 'all lathes were made a certain way on purpose, and the reason why they are made this way.  I received several answers from sellers of lathes.  One said yes, one said no, and two thought I was asking for help on how to machine a radius and suggested I purchase a radius tool.

 2.  The word concave as used here is misleading as it refers to a curve or radius.
Jonny's use of a straight line erring is much more accurate.  Also a facing cut is a straight cut at 90 degrees, not a curved surface.

 3.  I used the term 'cross slide' because of the answers on the other site.
Many suggested the problem of the 'dishing' is caused by the cross slide and commented how difficult it would be to correct the slide.  I disagree, it is not caused by the cross slide.  As Lew Merrick said the spindle must be aligned with the bed first.  Bogstandard showed it can be done with shims under the head.

 4.  Many do not understand the difference between tolerance and error.
A lathe is not designed to face a curve so we can stack cylinders.  It is designed to cut straight lines.  It is designed to face a 90 degree angle within a tolerance.  Were the lathe built with this error on purpose the this would be the only accurate cut.

 5.  The authors mentioned suggest to me that I am correct. 
Even though they use the word concave they are talking about a straight line of 90 degrees with a very small tolerance of less than 90 degrees over a long distance.  .001 in one inch, .008 in three inches are not tolerances but errors.  To suggest that all lathes are constructed this way is not correct.

 6.  If a concave or dish were the accepted norm the specifications would read +X/+X, or -X/-X.

 7.  The test certificate by Philf shows this if it is read correctly.

 8.  The mini lathe site talks about all of the errors built into the smaller machines but the only one they accept as not an error is the dishing problem.  The dishing problem and then the cross slide problems can be corrected by Bogstandard's methods.

 9.  I am not making tests on my equipment.  I was looking for documentation to prove or disprove other people's remarks.

10.  I feel that when a novice is astute enough to recognize his equipment is not functioning correctly it is poor practice to tell him a fairy tale.  We novices need to be taught correctly.

[Jonny]

I think this thread and links from it may be the definative answer regardless of what writers have written.

I think the answer you are looking for is in Phils and my link to Harrison owners site.
I feel its absolute bull what has been written else where, the error/runout on a facing cut is purely acceptable to makers tolerances nothing else.

Right who got a test sheet for a Smart and Brown or Dean Smith and Grace to prove the point.

[unc1esteve]

Jonny, I was not going to post again but I was not able to see the Harrison link.  I get a 'not found' message.

[bp]

A couple of minor points,
1  "Concave" does not infer/imply/suggest that the concavity follows a radius.  It simply means that the high points are on the outside and the low points are on the inside (think of a cave).  How the inside and outside points are joined up is relatively unimportant in the meaning of "concave".  In the case of a faced plane (hopefully!) the concavity will be conical.  It could be chewed out by a hungry woodpecker and still be concave, but unlikely to be generated on a machine tool.
2  I would bet that any test sheet from Hardinge, Smart & Brown, Schaublin (plus whoever else you care to mention)  would show an acceptable deviation in flatness, or concavity.  I would also bet that the deviation would be somewhat less that that shown by other manufacturers, such as that shown above.

Sorry if this can be categorised as fairytaleism
cheers
Bill Pudney