Lathe Spindle/Chuck Concentricity

[joegibParticipant @joegib]

Hi,

I regularly visit the American Home Shop Machinist website which may be known to some members here. An issue that quite often rears its head is which element of a screw-nose lathe's spindle determines concentricity when fitting a chuck? There seem to be two schools:

(1) The Screw School who maintain that the spindle screw is the paramount factor. They view it as a sort of spiral cone which, if closely fitted to the chuck backplate's female thread, performs a self-centering function. In their eyes, as the backplate spigot comes into contact with the spindle flange, the backplate thread slope is forced to rise up the the spindle thread slope until the two components lock. This tends to equalise any thread clearance around the diameter and thus self-centres the chuck. They hold that that the so-called spindle 'register' performs no useful function and claim that this term is a misnomer. So, when making a chuck backplate the key point is to machine its thread to be as close a fit on the spindle as possible short of causing a seizure.

(2) The Register School who maintain that the chuck register should be seen as acting something like a locating dowel. Here, the crucial factor is to fit the backplate spigot's recess as closely to the spindle register as possible short of, again, causing seizure. Under this view the backplate's female thread should be an easy fit on the spindle thread — close enough to ensure locking but not so close as to interfere with or constrain the register fit.

One 'authority' cited by the Screw School is Tony Griffiths of Lathes UK — see here (under 'Making a Backplate for a Screwed Spindle Thread', final paragraph):

**LINK**

Basically, he says that he used to take the register as the reference point until he encountered a Boxford spindle with no register and yet which showed high concentricity.

Yet only last night I was reading an article by Harold Hall covering this very issue (MEW155, Page 12) and he says of backplate fitting 'Most important is the parallel bore' [i.e. the backplate recess]. Further, 'As the location of the device is achieved by the parallel bore, the thread needs to be a little on the generous side'. (To be fair to Harold, I think his professional background is in electrical engineering so I don't suppose he'd claim to be an authority on mechanical engineering).

So, can any member with a professional background in mechanical engineering point to an authoritative pronouncement on this, please?

Joe

Edited By joegib on 17/10/2012 13:39:25

Edited By joegib on 17/10/2012 13:47:02

[joegibParticipant @joegib]

[Harold Hall 1Participant @haroldhall1]

You are correct Joe, I do not claim to be an authority on workshop activity, and am often concerned that people send me questions giving the impresion that they think I am. On the other hand, I often feel that my remoteness from the activity helps me to view situations, not blinkered by tradition.

However, to the question of fitting items to a lathe's screwed spindle nose I would like to ask one question and make one observation.

Question. Why do lathe manufacturers spend so much time and effort producing an accurate parallel portion to their lathes spindle if just a screw thread would work just as well?

Observation. It is extreemely difficult to keep the internal thread of a chuck or back plate perfectly clean.

Harold

[blowlampParticipant @blowlamp]

The two surfaces of the register are what aligns the chuck/faceplate to the spindle, not the screw thread.

Tony's final paragraph in the supplied link is incorrect in its assumption that, because he was lucky on one occasion, then it follows that his claim is proven. I'm surprised he says such a thing.

A simple proof can be had by making an accurate backplate minus the thread and checking for runout on the target machine whereby it's secured with a drawbar and thus of no influence to its alignment.

Many modern lathes have a flange fitting.

Martin.

[MICHAEL WILLIAMSParticipant @michaelwilliams41215]

There is no great mystery :

(a) Making a parallel register on a spindle fit another parallel register in a chuck backplate is relatively easy .

(b) Making a pair of screw threads match to the same level of accuracy is much more difficult .

Since in practice the plain register works perfectly well and the screw register has no real advantage most manufacturers go for plain .

Makers of very accurate lathes and other equipment have come up with a variety of alternative registering methods . The very best of all – though very dificult to do – is the Schaublin taper plus flat surface double touch system used in their very small quick change milling spindle system .

A somewhat simpler version of the same thing has been used on larger EMCO lathes .

Michael Williams .

[joegibParticipant @joegib]

Thanks for your quick response Harold.

Yes, I recall from the discussion threads I mentioned the 'screw' proponents were asked several times why manufacturers go to the trouble (and cost, of course) of machining this element. but they neglected to answer. There is also the question of why it's universally called the register. That term is used in a variety of industries to signify a locational device e.g printing, so how did it come to be used here if it has no function?

Joe

[colin hawesParticipant @colinhawes85982]

My 84 year old 7 " drummond lathe has a free thread fit and a good register and is very repeatable for concentricity. I believe that the register is the only important area but the thread must be perfectly square to its axis to seat the chuck to the back face of the register. Colin

[LambtonParticipant @lambton]

Page 36 of the manual supplied with my Myford Super 7 some years ago gives guidance as follows.

"CHUCK BACK PLATES & THREADED BODY CHUCKS

Register bores are held to very close limits. When back plates or threaded body chucks are supplied as separate units after the machine has left these works the register bore may need very light scraping of polishing with fine emery cloth. Do not screw equipment on to the spindle nose without ensuring the spindle register diameter is lightly smeared with fine oil"

Clearly Myford designed their lathe to the "register school".

I have found that many part-machined Myford back plates offered by the normal accessory suppliers neither have a close fitting thread or an accurate register diameter. "Rattle fit "comes to mind.

I have very carefully made a dummy register on a short piece of steel to the exact diameter of my lathe spindle register. I take with me when buying a commercial back plate and you would be surprised how difficult it is to find one that is correct. One supplier seeing me do this said that all his stock was made in England on CNC machinery so they were all identical and could not be wrong. Such faith!

When changing chucks I ensure both the spindle nose and the chuck bore are clean and free from foreign matter and then smear the spindle threads and the register with a little Rocol Anti-seize compound. This make the chuck easy to remove even after heavier work or being left in place for a long time.

[Harold Hall 1Participant @haroldhall1]

Adding to my answer above.

If a well made internal thread is made to fit a threaded spindle, then one is just in with a chance, providing the internal thread is kept clean.

If though a well fitted register is machined to fit the lathes register, then it cannot fail to work.

Harold

Edited By Harold Hall 1 on 17/10/2012 18:18:14

[NJHParticipant @njh]

Hi

I recall, many years ago at a ME exhibition, Myford selling, quite cheaply, short lengths of bar with the Myford screw thread and register machined on them. They also sell ( sold?) a screw thread and register on a No 2MT arbour for mounting chucks in the tailstock etc.

I have found the No.2 MT version useful from time to time but am still seeking a use for the other version – although it might serve for " Lambton's" ( Please give a First name for us to use!) purpose. Anyone got any other bright ideas?

Regards

Norman

Edited By NJH on 17/10/2012 18:33:11

[AndyfParticipant @andyf]

A dummy spindle nose comes in useful if you want to turn something which screws on to your real one. For example, if you want to make a screw-on ER collet chuck, you can test the thread, and then the register diameter, without removing the job from the chuck. Once right, you can fit the embryo collet chuck on your real spindle and machine the taper etc, thus keeping everything concentric..

Andy

[Jeff DaymanParticipant @jeffdayman43397]

In support of Harold's and others' comments I can share what I was taught and what has proven correct in my 28 years in industry, much of it having to do with tooling and fixturing. The basic concept of fixturing is to restrict the parts' relative movements to each other, normally six "degrees of freedom". Below are the basic rules of fixturing.

1. You can not repeatably and accurately locate anything on a thread. They are handy for fastening and driving things, but not for location. If there is ANY clearance in the thread, the parts will have six degrees of freedom.

2. You can locate a cylindrical part to another cylindrical part with a close fitting register (male round part fits in female round bore with minimum possible clearance) PLUS a planar surface engagement. The parts will have 1 degree of freedom with a close fitting register system, only rotation.

3. You can locate two flat plates to each other using two dowels PLUS a planar surface engagement. One dowel is not enough and three is too many. Fits of dowels in holes need to be close for good location. This first plate to plate method is how 99% of locating fits between plates are achieved in industry. With close fits, nearly zero degrees of freedom is possible.

4. The second method for extreme locating accuracy between two plates is the full kinematic system consisting of a sphere fitting a cone, a sphere fitting a V groove, and a third sphere resting on a planar surface, or, 3 spheres fitting in 3 V grooves arranged radially 120 degrees apart. This joint is done in measuring lab equipment but due to high cost and difficulty in making the features with adequate relative accuracy is not commonly used much outside the metrology business. A good link for this type of joint system with zero degrees of freedom is at **LINK**

JD

[MICHAEL WILLIAMSParticipant @michaelwilliams41215]

Hi Jeff ,

Nice explanation .

Full kinematic joints are pretty rare in general engineering but they do sometimes turn up in surprising places – jet engines and Renishaw probes come to mind .

Regards ,

Michael Williams .

[Alan HopwoodParticipant @alanhopwood63369]

I could certainly suggest a use for the illustrated Myford nose on a plain shank. Whilst it might not please the purist engineer, I have occation to use a Myford size chuck on the Harrison L15 I use at work. This has a 10" threejaw chuck, so when I want a more delicate part held I often chuck the small chuck in the larger one. I also have an Albrecht 1/2" chuck on a spigot which gets put in to hold little screws etc. for modification. As the saying goes in my industry "It's near enough for farm work".

Alan.

[NJHParticipant @njh]

Thanks Alan

Clocking it up in a 4-jaw independant might offer more precision though!

My equipment is a rather the opposite of yours – the Myford is my "big" lathe and my other lathe is a collet only model with collets 0.4mm to 4.8mm ( x 0.2 mm steps) !

Andy – yes good idea but I already have a screw-on ER25 chuck ( I didn't think of making my own at the time!) but maybe some other similar project may occur and it would certainly be useful for that.

Regards

Norman

[Jeff DaymanParticipant @jeffdayman43397]

Michael Williams – thanks for your comment – you have a private message

JD

[MICHAEL WILLIAMSParticipant @michaelwilliams41215]

Jeff ,

Thanks for the notes – looks very interesting .

MikeW

[MadMikeParticipant @madmike]

You are mostly correct guys about chuck mounting. Threaded mounting is simply not accurate and definitely does not have repeatability.

A "register" is the best compromise if you read the explanation by Jeff Dayman (very well put Jeff incidentally).

many manufacturers have used parallel registers with again some possible limitations in accuracy end repeatability. Jeffs final solution whilst the most accurate etc has some serious cost and manufacturing implications for OEM suppliers.

IMHO a great compromise is a tapered register, concentricity and repeatability become very good in practise. How to secure the thing? Well many simply fit bolts and tighten onto the register. What an absolute nightmare when changing chuck/faceplate etc etc. Our dear friends at Myford did offer an alternative solution on the 254 range and that of course was a Camlock chuck. These variously known as D1, D2, D3 etc, dependent on size, have a tapered register and the chuck can be simply released or tightened using a "chuck key" type tool which gives accurate and rapid chuck changes in seconds. I have at various times checked mine by turning sample bars, removing the chuck, spinning the spindle and then refitting the chuck, and finally clocking the test piece. Invariable I get repeatable concentricity of 0.0001 to 0.0002 inches.

Despite this manufacturers, evn today, tend to go for the cheapest possible solution leaving we, their customers, to suffer as a result.

Thank Heaven for my Camlock D3 mounting.

Rant over.

[Ady1Participant @ady1]

The two surfaces of the register are what aligns the chuck/faceplate to the spindle, not the screw thread

That's my understanding of it

The back register part, the L shape gives you axial and radial concentricity

[AndyfParticipant @andyf]

Posted by NJH on 17/10/2012 19:31:22:

…….. Andy – yes good idea but I already have a screw-on ER25 chuck ( I didn't think of making my own at the time!) but maybe some other similar project may occur and it would certainly be useful for that…..

Norman, as it was made by Myford themselves, I bet it would be worth something to someone. Maybe you could rent it out to those who want to make properly fitting backplates and other attachments.

Andy

[Alan HopwoodParticipant @alanhopwood63369]

The various methods of fitting a chuck all have some merits, but we shouldn't forget that if you want concentricity the only two ways are either;a) turn the whole thing at one go, or b) turn between centres. All the griptrues, collets, or special fixtures in the world do not give total repeatablity. The most important thing is that the job does not move during a cut. Please don't ask how I know this, unfortunately many examples can be furnished from the scrap bin.

Norman's comment re a four jaw is good, but even this has limitations on a ten inch four jaw as the jaws struggle to grip small diameter stuff.

Alan.

[NJHParticipant @njh]

Hi Alan

I was suggesting that you could put the Myford adapter in the 4-jaw and clock it up to achieve greater accuracy. I guess however that changing a 10 inch chuck is something you don't want to do very often!

Norman

[MICHAEL WILLIAMSParticipant @michaelwilliams41215]

Not quite on the original topic but NJH mentioned changing chucks . One man can just about change a 10 inch chuck but above that size life gets very interesting ! The biggest chuck that I have ever actually seen changed was 48 inches diameter and fitted to a DSG . It took turner , foreman and crane operator over an hour .

Equally interesting are the methods used to mount and true very big workpieces when workpiece is dangling loose from a crane . Between centres jobs are not too bad but putting things in big chucks and even worse on big faceplates is an art !

Don't know why but the vertical axis lathe which is much better suited to many of these chuck/faceplate jobs was quite a rarity . In a working lifetime of seeing thousands of ordinary lathes I haven't come across more than a very few vertical axis ones in regular use . The only place that had several of them was a company making large diameter relatively thin machined casings for the aerospace industry .

Michael Williams .

 

Edited By MICHAEL WILLIAMS on 18/10/2012 11:46:26

[jason udallParticipant @jasonudall57142]

…. was a company making large diameter relatively thin machined casings for the aerospace industry .

Michael Williams .

…And bell foundries….

[joegibParticipant @joegib]

Many thanks, Gents, for taking the time to comment on this.

It's a comfort to know I won't have to put new backplates on my 3-jaw chucks! It's also nice to see from Jeff's informative remarks that there's no Great Atlantic Divide on this but that there are genuine differences across the board. Anyway, I'm happy from the pretty uniform views expressed here that Harold's 'classic' practice for fitting a backplate can readily be defended.

As it happens I have 2 spindles for my Myford ML10 plus a Myford-produced dummy spindle nose of the type mentioned by Norman. As a complete amateur I was hesitant to enter the debate on that forum but I did make a post reporting my findings in measuring the 3 Myford spindle noses. As regards the registers, each measured 1.250" dead with not a smidgeon of difference. However, a simple measurement across the thread crests showed more variance. One spindle measured 1.125" (nominal), another 1.124" and finally the dummy measured something 1ike 1.121 (I can't be more precise as the post is long gone and my dummy nose has gone walkabout). Anyway, I suggested this showed that one fairly well-respected manufacturer, at least, was punctilious about meeting their nominal register spec but less so as regards the thread dimension. This argued that Myford regarded the register as the ruling factor, not the thread (Lambton's quote from the Super 7 manual confirms this). In the event this data was ignored by the Screw School as they'd done with other challenges.

The point of this tale is to raise a note of caution about using the Myford dummy spindle nose as a gauge for preparing a backplate. The register will probably be fine for this purpose but any variance from nominal in the thread might produce unpredictable results when fitting to the target machine. It might be wise, Norman, to check whether there is any difference in your case.

Joe

[Author]

Posts