Improve 3-jaw chuck repeatability

[old martParticipant @oldmart]

First you need a range of precision ground cylindrical steel bars to test your jaws. The humble drill shank, if not suffering chuck damage would be a start and a Morse taper socket would also be useful. Tightening each of the usual three sockets in sucession from loose and recording the runout close to the jaws will tell you which socket is the best to use.

The adherance to a tight register does not allow any better runout than the best socket of any chuck. All of the three jaw chucks at the museum except the serrated jaw one that always uses soft jaws has a loose register of about 0.010", 0.25mm, allowing adjustment to zero if required. In practice we don't bother with anything under 0.002", 0.05mm tir for a three jaw scroll as 4 jaw independants are used for fully adjustable and better workholding than a 3 jaw.

Three jaw scrolls are normally held onto backplates by three screws, studs or bolts, and it is easy to add another three for security as ours are.

Edited By old mart on 08/08/2023 16:17:53

[Ian PParticipant @ianp]

Posted by Sonic Escape on 08/08/2023 12:21:24:

Then I think it is time to try to remove the chuck to see where is the problem.

One more thing, if I push the chuck I can make the indicator to deviate by up to 0.025mm at 2cm from the chuck. Is this normal or maybe the bearings are worn out?

Have you calibrated or checked your 'push' for repeatability

Not all pushes are equal

Ian P

[Chris MateParticipant @chrismate31303]

Even theoretically a grinder is questionable as the stone wears off in one/every sweep, and in the end one is faced with temperature and your own accuracy of any measurements..

[Bill PhinnParticipant @billphinn90025]

Posted by Chris Mate on 07/08/2023 19:42:18:

So how do one readjust a collet unless its mounted in an adjustable holder-?

Edited By Chris Mate on 07/08/2023 19:47:05

You use an EasyZero collet nut.

[Howard LewisParticipant @howardlewis46836]

If the spindle is moving measureably, that needs to be cured first of all.

You cannot measure accurately anything that is moving about, either longitudinally or radially.

Minimising the longitudinal play should improve the radial, at least a little, (More if the bearings are conical, ball or taper roller ).

There will be play in the chuck, because of the clearances between parts (Scroll to body; Scroll to each jaw; each jaw to the slot in which it slides. ) And is the distance from jaw face to scroll engagement point equal on each jaw?

Is the jaw face perpendicular to the grooves on which the jaw slides?

Given the build up of clearances, it is lucky to find a 3 jaw chuck that holds work concentric within 0.075 mm, on a new unworn chuck. Especially given the prices for hobby machine chucks.

For high precision Industrial chucks you need to be sitting down!

Solve each problem , in turn, in a logical sequence.

Then look at overall accuracy.

Howard

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Bill Phinn on 08/08/2023 17:48:11:

Posted by Chris Mate on 07/08/2023 19:42:18:

So how do one readjust a collet unless its mounted in an adjustable holder-?

Edited By Chris Mate on 07/08/2023 19:47:05

You use an EasyZero collet nut.

.

That looks very slick, Bill

Although it’s basically moving the adjustment to the other end of the collet.

MichaelG.

[old martParticipant @oldmart]

We could easily get 0.001", 0.025mm by pushing the Smart & Brown model A spindle, but as it is plain bearing, the hydrostatic film would take up most of that when the spindle is turning. Everything has some flexability, you would get less with a DSG than with a Myford obviously.

[Sonic EscapeParticipant @sonicescape38234]

I'll try to summarize the main ideas so far:

1. There is no longitudinal play. Only 0.003mm if I push the chuck with both hands as hard as I can.
2. Radial play is up to 0.01mm measured at 10mm from the chuck. If I push it with one hand. Always return to original position. So it must be just metal bending not play? See the video bellow.
3. The main problem is the 0.120-0.130mm runout. This was confirmed with a test bar and two carbide mils. There is no improvement if I use a different socket hole of the chuck.
4. EasyZero collet nut is interesting but it can dial out only 0.05mm maximum.

I'm looking for a C-spanner to remove the chuck. Now I'm curious to see who is causing this runout. If the spindle is fine I'll replace the chuck. Anyway this one feels very gritty and usually it requires both hands to turn the key.

[Chris MateParticipant @chrismate31303]

Something I found about the chuck that will throw me around is if its not clean from chips in the scroll.

[Howard LewisParticipant @howardlewis46836]

If everything else has minimal play, and the run out is only for work in the chuck, you have two choices, either live with it , or buy a new one. But don't be disappointed if a new is no better!

(Run out of that magnitude is not uncommon with a used 3 jaw, or in some cases a new one! )

I have only seen one 3 jaw with minimal runn ot, all the otghers have had some, depending on how worn they are.

Run out of the magnitude that you are measuring is not that unusual, and might well vary according tghe diameter ofc the piece claped in the chuck(As I said scroll to body, scroll to each jaw,each jaw to body, dimensions / squareness of faces within each jaw )

If you think of all the clearances present in a 3 jaw chuck, just be surprised that run out is a small as it is.

If absolute concentricity is essential, you either turn the two or more diameters at one setting, without disturbing the chuck, or you use a four jazw and clock the work until it is concentric to within what you are prepared to accept.

If you reduce a 25 mmbar to 20 mm and elsewhere to 15 mm; the 20 mmand 25 mm diameters will be concentric, so the initial run out won't matter when you part off trhe job.

Howard         Why do the typos only become obnvious after pressing "send"

Edited By Howard Lewis on 09/08/2023 08:44:23

[old martParticipant @oldmart]

As I get older, my typos get more common and even after proofreading, they hide until after pressing "add posting".

Strange, I got this reply correct first time, a first for me.

[SillyOldDufferModerator @sillyoldduffer]

Posted by Chris Mate on 08/08/2023 23:13:47:

Something I found about the chuck that will throw me around is if its not clean from chips in the scroll.

+1 Swarf in the chuck may be the problem.

Rather than remove and strip the chuck, last time I cleaned mine on the lathe. I removed the jaws and then rotated the scroll by hand with a large matchstick held in the slot. As the scroll turns, swarf catches on the matchstick and can be seen and flipped out.

Not the best way to clean it, but the chuck doesn't have to be taken off the lathe. A big advantage when you don't have the necessary C-spanner yet!

Oil the scroll after cleaning.

Dave

[Howard LewisParticipant @howardlewis46836]

When I clean the scroll on my 6" 3 jaw, in situ, the tool used is a toothbrush, narrowed to fit into the slot for the jaws.

Rotating the scroll eventually works any swarf into centre so that it falls out and can be brushed away.

(A job for a bottle brush pushed through the spindle, from back towards the bed . )

But having cleaned the chuck, and got everything else as good as possible, remember that ikt a 3 jaw chuck.

So 0.012 – 0.013 mm is not too wildly out, especially if the chuck is not new.

A 3 jaw chuck is not, and by virtue?? of all the clearnances within it, cannot be a precision device capable of matching the concentricity produced by good collets, or a carefully adjusted independent 4 jaw chuck

If you can get down to 0.013mm eccentricity, I'd stop worrying and just get on with using it.

Howard

Edited By Howard Lewis on 11/08/2023 16:49:40

[Bill PhinnParticipant @billphinn90025]

Posted by Howard Lewis on 11/08/2023 16:42:13:

A 3 jaw chuck is not, and by virtue?? of all the clearnances within it, cannot be a precision device capable of matching the concentricity produced by good collets.

It depends what you're paying for your 3 jaw chuck. Bison and Pratt Burnerd have made adjustable 3 jaw chucks for a long time. To be specific, both claim "adjusting accuracy of 0.005mm" and "repeatability 0.015mm" for their adjustable 100mm 3 jaw chucks. I don't know any ER collets on the market that claim greater accuracy than 0.005mm runout [except maybe Regofix?].

[old martParticipant @oldmart]

Pratt Burnerd and Bison adjustable chucks cost as much as many people spend on their lathe.

[David George 1Participant @davidgeorge1]

I have a Pratt Burnard chuck which if i take care repeats to less than a thou. I use the key hole marked with the O all the time and check it with a few diferant sized hardened and ground dowels from 8mm dia to 13mm and 25mm to ensure it runns true especialy if i have had a problem of tool jam etc. the chuck is screwed to the back plate with 6 cap screws three from the front into the back plate and three through the back plate into the chuck. there is clearance on the location between the chuck and back plate so with the screws slightly loose I can put a dowel into the chuck and tighten the chuck with the O key hole, then tap the chuck till i get a zero reading. Nip up the screws and check runnout again and if ok fully tighten all screws. this set up works for me and i get consistant good holding and location on all jobs as long as they are round etc.

David

[Author]

Posts