Chuck out of true

I’ve checked out the “area of interest” and there may be a minor “groove” but no raised area.

I used a lever DTI on the register, photo shows maximum variation...

I'll have a go at polishing register with some very fine wet & dry & cutting fluid.

I just happened to be given some round bar the right size for a back plate, so may have a go at making a new one. Not overly confident, but it'll be good practice.

Concentricity and dimensional precision (of the registers) are the two important aspects. All done without removal from the lathe chuck should ensure that. Threading precision is not so important as long as it is in line with the registers and is of the correct thread form. Again, if done without removal from the chuck should not be anything but on line.

Initially an imposing challenge, but when you have done it once, you will wonder why you were worried - even if the first attempt is not too “pretty”!

Have a look at the three part series of u-toob vids by doubleboost on making one.

**LINK**

Thanks,

Watched doubleboost videos. Makes me feel a lot better knowing it’s not just me who sometimes takes too much off, and good to see the solution when you do.

Grotto, do take care when"polishing" the machine spindle register -- you are polishing to remove burrs only not removing any material that alters the OD !

The register I was suggesting you needed the DTI for id the backplate chuck register seen here adjacent to the indicator probe - this is normally a very close fit in the chuck.

When you come to finnish the diamiters on a back plate, take care that it is not warm as when it cools the size changes, a bore will shrink and an outside diamiter nice fit will be loose.

David

Spot on. I found that on the first back plate I machined about 25 years ago. I very carefully cut the spigot, for the chuck, to a nice close fit - but it was much looser the next time I tried it. There was plenty of metal, so I cut it again, as I recall. An undercut at the inside corner of the spigot is also important if the chuck has a square corner. This is because they are unlikely to fit together perfectly without some clearance at that point.

The three places on the backplate that need to be accurate, if concentricity of the chuck is to be optimised, are:

1 ) Register for mandrel. Once this is satisfactory, the "backplate to be" can be screwed onto the Mandrel, and the other faces machined without disturbing the embryo backplate.

2 ) Squareness of front face of backplate

3 ) Outer Diameter of backplate where it fits in chuck body.

An adequate chamfer between the front face and O D is necessary, to prevent any foul in the corner of the recess in the back of the chuck.

Even with all the above at maximum precision, work held in a 3 jaw chuck is unlikely to runs absolutely true, 0.003" to 0.005" is fairly good. (I have seen one with less than 0.001" run out, but that is exceptional. )

For absolute concentricity, you have to mount the work in a 4 jaw chuck, and clock it and adjust until it is acceptably concentric.

Howard

You can also get very good concentricity in a 3-jaw by using a set of soft jaws which you can bore to a set dimension whilst gripping a scrap piece of known round bar. These work best for components you've machined them to fit, but are often also useful for parts within a small range around that diameter.

My Warco's 3-jaw used with bar in the 6 - 20 mm diameter range will also often run within 0.001" TIR. I'm thinking that general build tolerances for such equipment may have improved since the 1970s.

Yes, soft jaws, or a home made collet should produce absolute concentricity.

Very useful equipment, but not what the OP has, by the looks of it.

Chucks taking soft jaws are probably in the minority in the hobby world, much as they are useful.

Howard

Well I made a new back plate. Took an awfully long time as I hadn’t planned the operation as well as I should and had lots of set backs. Did it in the 4 jaw, made the register hole first, then cut the thread (leaving it in the chuck the whole time. Took the chuck off so I could test fit on register about 50 times as I was paranoid about making it too loose.

Result has exceeded all my expectations (which were pretty low). I was a bit worried about getting the holes in the back plate to line up with the chuck, but realised the holes in the original back plate were drilled over size to allow a little movement.

I've run a dti around the new back plate and get no no real movement which is good.

I've fitted it to the chuck, and although it's better than it was, it's still not great...

3 thou is pretty fair for a used 3-jaw.

As Pete, 3 thous is adequate for most work - all turning when you don’t need to remove, and then replace, the job in the chuck.

I haven't been here for a long time and have just come across this debate. As usual lots of theories and lots of great advice. However despite the original statement that some work did not turn out well, we have not, as far as I can see, actually established exactly what was wrong, and the scale of the problem.

The fact that the chuck is "running out" may or may not be the result of a scuff mark in the back plate or some swarf in the register.

The only true test in this situation is to machine some metal in the offending chuck. How about turning the classic Myford Dumb Bell test piece and then measuring that for run out before it is removed from the chuck. Then we will all know what we are looking for. Machining back plates, polishing registers may or may not be the answer. Let's look at the problem from an engineering perspective, after all if this occurred in a manufacturing environment you would not try maching bits off of the machine or its tooling until you had established the problem. We have so far, IMHO, simply been focussing on the symptom rather than the cause.

Mike,

I agree with what you say ... but I should just mention that my interest in the marks on the backplate's cylindrical register was prompted by Grotto's statement in his opening post:

"when I removed it I noticed that it was very tight on the head stock thread."

... and I was keen to discourage him from tapping or screwcutting, until all other options had been explored.

MichaelG.

I'm interested in the focus on the cylindrical register. It may not be as important as some may think.

This is my take on it.

In order to fit a chuck there has to be some clearance between the cylindrical register and the mating part of the chuck nose. Ideally a close running fit. The metal in tight contact is the screw thread (actually the rear flanks of the thread) and the flat register on the back of the chuck and correspondingly on the face of the nose.

Dings in any of these will affect the fit of the chuck. As will eccentricty of the screw thread. The cylindrical register can only impose a limit to run out as it is not a mating face (not in intimate contact).

The most likely cause of a bad fit is swarf or burrs ion the tops of the threads so that should be addressed first followed by high spots on the back mating surface and then the cylindrical register.

As I said this is my take on things and I'm happy to hear other arguments.

In addition as has be alluded to 3 jaw scroll chucks are not ever going to be that precise so one can only go so far with them.

MadMike suggests doing a turning test which will tell you nothing about the chuck fitting as it only tells you how good your headstock allignment is and to an extent the state of your spindle bearings.

regards Martin

Hi Martin on my lathe the thread is clearance at least 5 to 10 thou clear and the location is the register diamiter and location face there is only a few tenths of clearance on the diamiter. I get a consistent location repeatability of a few tenths taking it off and on again.

David

Mad mike,

Go back to the earlier posts (even the first) by Grotto and you will see there is an apparent step change when the chuck suddenly disliked being threaded into position. That is why most attention has been directed towards the threads and registers.

The thread has deviated somewhat since its beginnings.

Edited By not done it yet on 08/08/2019 10:18:24

That's essentially what I said that the cylindrical register 'limits' the run out but even though you say your thread has a large clearance, it doesn't when it's tightened. Sounds like everything is nice and clean and concentric with you.

regards Martin

NDIY, the original post started by referring to components that had not turned out well/accuately. The subsequent discussions followed as the poster remarked upon the tight thread fitting of the chuck. The chuck may not have been removed for months, who knows?

My views about turning a test piece are suggested as the condition of any of the component parts of a chuck are not necessarily the issue. The test piece will tell whether the machine is producing round, concentric, tapered or inaccurate product. The key in all of this is to determine what the machine actually does. I have seen people complain about chucks running out of true, after putting a DTI on the outer of a chuck body. Three jaw chucks are simply a convenient holding device and are not by definition a super accurate piece of tooling. Testing the output from the lathe will help to firstly identify if there is an accuracy issue, and then help to establish the cause and any subsequent corrective action. Everything else is just guesswork.

If any of my "engineers" had put this problem to me without a full test of machine capability I would have sacked them. Diagnosis remotely without detailed analysis is not a good idea. If we purport to be engineers then we should act like one. If the original poster is a newbie then we must help and guide him in the correct process in situations like this.

.

MadMike ^^^

That's the opening post ... I have emboldened the key words

NDIY and myself are of like mind on this.

MichaelG.

Edited By Michael Gilligan on 12/08/2019 15:28:51