View Full Version : Chuck Register Clearance?

07-14-2007, 11:58 AM
I'm not sure of the correct terminology so please bear with me. My lathe has a 2.25" spindle and there's a smooth unthreaded area past the threads leading to the shoulder that the chuck tightens against. Is that area called the "register"?

If so, how much clearance should the backplate have in that smooth unthreaded area?

The backplate on my old Skinner 3 jaw happens to be 2.28" over my spindles 2.25" in that area. That's .030 of clearance which I'm thinking could also be called .030 of slop. That old 3 jaw chucks up all over the map and I've seen runout at .015 on occasion. I looked at it real close one day with various pieces of doweling and that time it was pretty consistently .006 out towards the middle of the #2 and #3 jaw.

I first noticed the 2.28" when I modeled the backplate I made for a 4 jaw chuck on the dimensions I found on the Skinner. At the time I assumed threads are to some extent self centering, but it could be I'm fooling myself once again.

Whaddayu guys think?

New backplate for the 3-jaw?


J Tiers
07-14-2007, 01:14 PM
The "register" (so-called) is a "religious" issue, which brings up "religious wars" whenever it is discussed. I am prepared to be flamed as an idiot here.... won't change my opinion a bit.

I am NOT a true believer, noting that it forms an extra restraint that is kinematically unnecessary. I have chucks with similar clearance, and they set up consistently.

The parallel fit SHOULD be a tapered fit if the "register" were TRULY intended to be a precision alinement means. Otherwise it must be a "wringing" fit if errors in the small numbers of thousandths are to be avoided. But a "wringing" fit would not last 50 years, or probably 50 DAYS if the chuck were changed regularly.

Making it a tapered fit would turn it into the same sort of thing as a cam-lock etc nose. Those have a precision taper to set alinement.

The "register" DOES have a function, which is to get your chuck started onto the final tightening CLOSE to centered, so that the threads and shoulder can center it. The threads form a 'cone" alinement means if they are well formed.

In fact, however, the threaded mount is what is stupid. It just isn't going to be consistent over the long term, and has other problems as well. That is why the long taper and cam-loc mounts came into being.

if yours is sloppy and sets up any old way, probably the backplate threads are shot, sloppy and worn, or even oversize, and the chuck can tighten up in a false position out-of-center. If so, new backplate could be a good idea.

I will predict at least 7 people will call me an incredibly stupid idiot who should stay away from machines for saying such a thing as the above, but too bad...... I said it.

07-14-2007, 06:42 PM
You won't get any argument from me. Sounds perfectly reasonable.

07-14-2007, 06:50 PM
IMHO, ideally any accessory held on the spindle would be a close fit to the register and would seat against the shoulder all around. This way you wouldn't be relying on the threads so much to center the chuck or backplate. Any accessory I make to screw onto the spindle, I take care to machine for a close fit to the register. I have a little book that I keep nearby where I detail all the sizes of important parts (like the register area ) so that as I make a new backplate, for instance, I can be confident that it will fit first time without having to remove it to test, then re-mount etc. Part of this information I keep is to compensate for the inside measurement inaccuracy of the dial indicator(s) I use. If I measure an inside diameter with a particular indicator, I know that the actual size is two and a half thou larger. This figure is with me manipulating the indicator, and might be different if someone else took the measurements. With care then, I can machine a backplate to keep the play over the register surface within a half thou or less. To me, a thou is a lot of slop on a critical component like that where you'd like it to be repeatable.

I suppose if you were to want to make a new backplate to possibly enable your chuck to repeat more closely, you'd have to first characterize your chuck to find out if the jaws can repeat, and at what diameters- without removing the chuck. If they can't be re-tightened onto a workpiece without remaining concentric, then it would be a waste to make a new backplate. If the workings of the jaws prove to be good, however, you could then arrange to turn a stub (on someone elses lathe perhaps) then mount your chuck to that stub and true up the back of the chuck and the register surface. Then you'd fit your backplate to your spindle, then turn that to fit the chuck. Theoretically, that should get your chuck running much more concentric.

Personally, I would think that a chuck should be able to pass this test- mount the chuck, turn a stub, remove and remount chuck, and measure runout of the stub. Without loosening the jaws, re-mount the chuck a few times. If it can't consistently be within a thou or so, I wouldn't be happy. Another test- turn a stub, loosen then re-tighten the jaws without changing the position of the stub, then check runout. Again, if it couldn't repeat within a thou or so, I wouldn't be happy. With this last test, if you change the position of the stub or rotate it at all as you loosen then re-tighten the jaws, the test is meaningless. Also, you must use the same sequence or procedure to tighten the jaws each time, or the test is meaningless.

Mark Hockett
07-14-2007, 07:44 PM
J Tiers,
I also agree with you. While working on the spindles Rose Engine lathe project that I posted about previously, we had many 1" X 8 TPI chucks and faceplates. I measured every one of them and they were all over the place. The tightest one had about .005" clearance and the most clearance was over .030". We also measured 5 different lathe spindles and found that almost none of them measured 1" on the register most were less than 1".

07-14-2007, 08:05 PM
Both of my chucks have FAR more than 30 thou clearance there.

I've always thought it is that perpendicular (to axis) interface between the back of backplate and (whatever that other surface is called...) that results in accurately seating the chuck.

07-14-2007, 10:04 PM
The shoulder is there to keep the chuck's axis of rotation parallel to the spindle axis. It won't help to get the chuck's axis to be concentric with the spindle axis. That's the job that the register could do, if the mating surface inside the back plate was a close fit to it.

Rich Carlstedt
07-14-2007, 11:33 PM
J Tiers is right on.
There is no "register' on a threaded spindle.
Those who make a close fit of the shoulder will deal with chucks that do not
want to come off..
Here are three facts which should be read.

First, there is a simple physical test .
Take a 6-32 screw and a 8 -32 nut and spin them together.
Now exert a pulling force between the two, and you will see that the two will
immediately center on each other . it is impossible to engage them on "one' side. the reason, is that a taper thread is a self centering helix, and the pitch center of the thread is control of both x and Y axis. All that you need is a perfect "face" on the back of the spindle to give total and repeatable control
of X,Y, and Z.

Second. A six jaw Buck chuck is considered to be one of the most accurate and repeatable chucks made commercially.
They can easily repeat to tenths, even when removed and later replaced.
My buck does this.
My buck has .030 "clearance" where the spindle shoulder is (erroniously called a register)
If it really used the shoulder for registation, why is it accurate ?
Is Buck so dumb to not know this ?
Of course not see #3 below.

There are "No Standards" for such a shoulder in ANY manufacturing standards
book or list!
There ARE standards for helix configuration, and perpendicularity of the face.
The reason is simple. it is not a concern for machine tool makers.
As a Manufacturing Engineer and member of SME, I have tried to find one.

The myth of a registration dies hard, because there have been some famous
writers in our magazines who passed on this dimension as a required diameter.It is unfortunite that they did not do their homework.

The reason for doing both the (1) Thread and the (2)Face at the same time and setup, when making a chuck back or faceplate, is fully understood when the above is considered. to do so any other way presents problems and variations.


07-14-2007, 11:41 PM
Thanx for the responses guys. As Darryl suggests I think I'd better look at the jaws more closely before investing time and material money in what could be an exercise in futility.

Again not knowing right from wrong, I can see that the removables don't line up with the masters. If I put a piece in deep enough to engage the masters there's a crack of light on the removable ones.

Besides that the #1 jaw seems to lead the others. If I close the chuck all the way up on just the masters #1 and #2 meet up first. Typically I'll find an .005 gap between #2 and #3 and the gap between #3 and #1 looks like it's big enough to drive a truck thru at .020 or so.

Makes no sense to touch up the jaws if the body can't repeat tho. I'm gonna walk out there and mount/dismount while checking runout each time. More later ....


07-14-2007, 11:49 PM
Wow. Rich checked in while I was typing! Good info. I'm still gonna have a look at body repeatability just for fun.

I know it's just a 3-jaw but I would like it to be a little more accurate. There have already been times when the stock was only .005 over and it's been a problem to get concentric without going under.

Thanx again.


07-15-2007, 01:03 AM
Got some results and apparently a new backplate isn't gonna do a bit of good. I chucked a piece of 3/4 PG shafting up and without unchucking dismounted the the chuck and re-mounted several times. Got the same .004 of runout. Same high spot between #2 and #3 too.

There's probably some room for improvement in the jaws but that means grinding and I'm not there ... yet. :D

So thanx again guys. At least I won't be chasing my tail over a new backplate.


Rich Carlstedt
07-15-2007, 01:25 AM
Before you buy new jaws, see if the chuck is one piece or if it
is bolted to a back plate.
Sometimes you can lossen the back plate bolts and shift the chuck,
to center it. use a indicator on your high spot and try to move it
one half the error..
If the back plate had a 'precise" shoulder to chuck fit, you may not budge it.
but if it was made with clearance, it will adjust ! and may be the reason the jaws are now out. it happens in a crash !

Paul Alciatore
07-15-2007, 02:45 AM
The thing about 3 jaw inaccuracies is that an older chuck frequently has wear on the scroll. This will give different results when chucking different diameters of stock as you are using different positions on the scroll. Worse yet, the wear is not going to be the same at the three positions, 120 degrees apart where the jaws are. So you tighten one diameter stock and it is high at jaw one. Another diameter may put that high spot, actually a worn spot on the scroll, at jaw 2 or 3. Results are inconsistent. Grinding the jaws will NOT correct this kind of error.

I have an old 3 jaw that has this kind of error. I am thinking of making a collet chuck out of it. The scroll would be completely removed for this conversion and be replaced with a threaded plate to tighten the collet. Of course, I have made many parts with it using the single setup method.

07-15-2007, 11:08 AM

J Tiers
07-15-2007, 02:39 PM
From reading this thread thus far, it seems to be regarded as "Gospel" and a universally held belief that 3-jaw chucks are meant to be repeatable and accurate.

Maybe where your at.

But not here.

or here. But they should be just as bad and in the same place when removed and replaced.

I notice that there are few if any NEW lathes sold with screwed spindle nozzles and that most if not all have "bolt-on" or "cam-lock" flanges.

I wonder why?

For the cheap asian lathes, the simple bolt-on is cheaper and puts the onus of centering in a blurred/shared responsibility between lathe maker, chuck maker and user. Convenient.

For industrial machines, the cam loc or other taper nose is simply light-years better. See my earlier post.

On the other hand.....

The bolt-on may severely limit chuck choices.

The screw-on type may be the dead cheapest to USE, and easiest to obtain or make new adapters for.

The cam-loc etc will be the most expensive to get more backplates etc for. Often double or triple the price of a threaded nose backplate, can be many hundreds of dollars more. One adapter and chuck may cost more than the entire lathe did used.

The thing about 3 jaw inaccuracies is that an older chuck frequently has wear on the scroll. This will give different results when chucking different diameters of stock as you are using different positions on the scroll. Worse yet, the wear is not going to be the same at the three positions, 120 degrees apart where the jaws are. So you tighten one diameter stock and it is high at jaw one. Another diameter may put that high spot, actually a worn spot on the scroll, at jaw 2 or 3. Results are inconsistent. Grinding the jaws will NOT correct this kind of error.

But shimming the scroll may very well fix it, or significantly improve it. Often there is radial slop in the scroll vs its "bearing". A small shim there may make an unusable drifting, moving chuck OK.

John Stevenson
07-15-2007, 02:55 PM
I can appreciate having a worn 3 jaw chuck, lets face it many of the machines in the home shop are showing some age but the good news is chucks have rapidly come down in price to the point where it's cheaper to buy a new chuck than a set of jaws.

So now if you have a chuck with worn scroll there is no reason why you have to put up with it.
Forget the mentality that it's only a bit worn, use it on the mill or drill for workholding where repeatability doesn't come into it and get a new chuck.

Even the better Chinese chucks are made to a higher standard than that worn clunker you are having to bash and dial in every time.
Lifes too short to piss about.

07-15-2007, 03:00 PM
I know you're right about 3 jaws Tiff and I've almost got my 4 jaw up on the lathe. :D While I did have to foresight to order the 5/8x11 tap I needed to finish the backplate the last time I ordered from ENCO, I did NOT have the foresight to look in the Handbook and see what size drill I needed! :mad:

Went looking for some body shift as Rich suggested and found the body is in fact .004 high between #2 and #3, or low at #1. Take your pick. Pulling the backplate revealed a slight shouldered area that I thought had a click of movement, but it all came to naught. I can't get it to budge offa .004 out.

It's got the scroll errors Paul alluded too also.

I'm sure I'll be happier when that drill shows up and I can get my 4 jaw mounted. I've used a 4 jaw exactly once on my buddies Atlas and got a 5/8 shaft dialed in under a quarter. That made me very happy.


07-15-2007, 03:03 PM
I need to type faster. I can't keep up with you guys!


J Tiers
07-15-2007, 03:13 PM
Went looking for some body shift as Rich suggested and found the body is in fact .004 high between #2 and #3, or low at #1. Take your pick. Pulling the backplate revealed a slight shouldered area that I thought had a click of movement, but it all came to naught. I can't get it to budge offa .004 out.

It's got the scroll errors Paul alluded too also.

So take off the chuck, and skim a few thou off the OD of the shoulder, then put back and "tap your chuck in" then tighten securely . "poor-man's adjust-tru".

Check to see if your scroll has radial clearance. if so, a few thou of shim may bring it into a "normal range of error".

07-16-2007, 02:27 AM
This old Skinner 3 jaw of mine has some life left in it after all. I did as JTiers suggested and skinned the shoulder on the backplate for some more clearance. After some quality "adjustment" time I'm running every piece of precision shaft or dowel I've got from 3/16" to 3/4" at under .002! Waaay better! No more high spot.

The other thing I found is of the 3 locations to tighten the chuck there's one that works the best. I don't know why that is, but it works everytime.

At any rate, I'm not gonna have to be afraid of going under on .005 oversized stock anymore. As long as I don't crash it and knock it off along the way ....


07-16-2007, 02:53 AM

07-16-2007, 08:41 AM

07-16-2007, 08:48 AM
Me too Tiffie!

Now that the old Skinner is operating somewhere near optimum do you suppose I should dowel that backplate on? Still learning here and it's quite likely I will be exerting "excessive force" on it at some point. If I knock it off I could always re-align, but who needs the hassle?

Never doweled anything in my life, but I'm assuming reamed vs drilled holes would be best? I own exactly one reamer and it's a 3/8". That would make for a couple of pins fat enough to take some newbie abuse.

Whaddayu guys think?


07-16-2007, 09:33 AM

Rich Carlstedt
07-17-2007, 12:14 AM
Doweling it is not bad in my opinion.
.375 is a bit oversize however.
It might crack the cast Iron near the outside edge with any press fit load.
try to stay in the 3/16 diameter area if possible.
To shear 3 of these would also rip out your jaws.
Using standard hardened dowels will (with 3) require
a force of 4,850 pounds to shear them..
the Cast Iron would fail before that

Without reamers, drill a number drill, like #13 (.185)
to depth (.625 ?) and follow with a sharp 3/16 run very slowly.
you should be close. You can always make your pins to fit,
but they will be softer.. One more thing when making them
if you can, drill a 1/16 hole through the center of the pin.
"If" you ever break it, it can be removed with a EZout.


J Tiers
07-17-2007, 12:32 AM
Don't forget, if you do NOT dowel it, you will be able to adjust -in a piece of STOCK to the same amount as you can the chuck.

That means you potentially can get your WORK running 0.000 by tapping-in the chuck and measuring on the work.

That was the meaning of the "poor man's adjust-tru" comment.

An adjust-tru type 3 jaw chuck has 4 screws to move it around and center the work.

You can get a similar effect with the slight movability of your chuck.

07-17-2007, 08:11 AM