Truing a 3 jaw chuck

I have a Taiwanese lathe and fitted a new Chinese brand 3 jaw scroll chuck to the existing back plate and had to machine a small rebate so the chuck fitted snugly on the back plate. on checking the run out on a mandrel chucked into the jaws it was .003" run out which is what the tolerance was on the inspection certificate.
My question is do I machine the rebate on the backing plate down more and set the chuck up using the bolts or should i leave it as it is and dress the jaws of the chuck to get rid of the run out.
Naturally I am reluctant to try dressing the jaws as I am afraid of doing permanent damage.
Thanks in Advance
Paul

The problem with your chuck is flex in the face, which allows the jaws to flex open under pressure. The amount will be different at different opening diameters.

If you enlarge the 'rebate' by half a mm or so, you can loosen the chuck mounting bolts slightly, tap the chuck around to clock your work, then tighten the mounting bolts. Best you can do with that quality chuck. Since you are only modifying the chuck there is no permanent damage to the lathe.

I suspect you don't really need better precision in most cases, so leave it as is would be my advice.
You can make your own adjustable 3-jaw as Red suggests, but -
Why not buy a 4-jaw? That would solve all problems!
Just my 2 cents,
fusker

Agree to a jaw wrong taper or wear and flex will cause much run out.
I agree with Red.

One fellow here trued the chuck back face on a new china chuck and so made it run very good (actually better not perfect).
As if all the chuck front features had been machined at manufacturing close from the front, and then the manufacturer finished the back with a little error. No way of knowing if that is true with your chuck. if you had a dead true test bar held between centers you might be able to check that With chucking the chuck on the test bar and running an indicator to the back. Still, the bar going full through will be different than just chucking a part way through the jaws.

Trued face plate should be marked so going back on at the same line up to the lathe.

Certainly, all the features of the headstock should run near zero.(.0003 or better IMHO)

Older scroll chucks trued-up may run good on the diameter where trued and often still out at other diameters.

A good quality 4 jaw is worth the purchase because that makes a lathe able to run close jobs...A poor quality 4 jaw may have wobble so even checking good at one place may not check that good on down the part.

Being quick and accustomed to finish skim a close part between centers is a handy habit.

Red James said:

The problem with your chuck is flex in the face, which allows the jaws to flex open under pressure. The amount will be different at different opening diameters.

If you enlarge the 'rebate' by half a mm or so, you can loosen the chuck mounting bolts slightly, tap the chuck around to clock your work, then tighten the mounting bolts. Best you can do with that quality chuck. Since you are only modifying the chuck there is no permanent damage to the lathe.

Click to expand...

I do exactly what Red James said, and have done for 45 odd years ....and get yourself some soft jaws.

Thanks for the various suggestions, Do you mean a 4 jaw scroll chuck? I have a four jaw chuck independent which i use when I need zero run out and dial the work in, the reason I would like the 3 jaw on zero run out is so I can do precision work without having to dial in the material first and be able to take work out and refit it into the chuck. Would it be feasible to set the chuck up on a mandrel between centres and check the mounting face with a dial indicator and true it up if needed?

That is what I mentioned in post 4 but I have never done so. If considering that, I would chuck up on the test bar a number of places and mark the chuck a number of times to see that it repeats a number of times at a number of places.
Still at other diameters, it may not the same.

It is .003" out at pretty much all diameters which leads me to believe it is the back face of the chuck which is out by that amount. I
will mount it on a mandrel and dial it and see how it comes up. I will try it in a few different positions.I have a big chunk of 39.5 mm 1045 chromed steel bar I use as a mandrel.

david6855 said:

I would like the 3 jaw on zero run out is so I can do precision work without having to dial in the material first and be able to take work out and refit it into the chuck.

Click to expand...

May as well try pissing UP a rope unless your idea of 'zero run out' is measured with a grade-school plastic ruler.

Regardless of 'jaw count', nothing you can do to the jaws can eliminate scroll wear, damage, distortion, imperfect manufacture, or imperfect positioning. All of which get worse, not better, as the scroll operated chuck is USED for anything meaningful.

Either find a collet system that can hold the size stock you need 'rather well centred' (never even then 'perfectly') . ELSE .. learn that mucking about with a 3-J, even a Set-True/Adjust-True isn't necessarily as fast as ...

-- using a DECENT 3-J within its inherent limits when you CAN do [1].

Then swapping to a collet system ELSE 4-J independent when you SHOULD do.

IOW - settle for making tea or coffee reliably.

Not trying to boil the entire ocean with a not-even-adjust-true Chinese 3-Jaw chuck as your primary cookware.

And sorry . .your budget or lack thereof dasn't impress the Laws of Physics by the mass of a single charmed Quark. Wishing and chiseling won't change that.

Yah wants better repeating precision, yah gots to invest in better grades of tooling and more options among it [2].

Meanwhile soft-jaws tuned-up for a specific task, same scroll position, worn, damaged, or otherwise, will be faster and cheaper that f*****g with the chuck under the soft-jaws.

PM no doubt has electroplating experts who could in fact gold-plate a dried turd. Don't expect them to try and make it into a reg'lar exercise of the sort folks just won't quit trying to apply to cheap, worn, damaged, or 'all of the above' 3-J scroll-operated chucks, year after year.


Bill

[1] Anything under ten thou is eminently usable for what a scroll-operated chuck SHOULD be doing. THREE thou is damned good. Trying to 'hold' not just hit better numbers in real-world USE is chasing diminishing returns. Even rather good collets sweat to break a half-thou, net-net on repeatability. Specs only call for 2 to 6 'tenths' on most collets manufactured in volume, brand-new and undamaged.

[2] Those uber-precise over multi-thousand piece-count 3-J nose-art found on CNC critters?
Power operated. AND NOT by scroll-back. And not even remotely close to 'cheap', even if Chinese.

david6855 said:

It is .003" out at pretty much all diameters which leads me to believe it is the back face of the chuck which is out by that amount. I
will mount it on a mandrel and dial it and see how it comes up. I will try it in a few different positions.I have a big chunk of 39.5 mm 1045 chromed steel bar I use as a mandrel.

Click to expand...

QT: [pretty much all diameters] but is the error all the same way each time? perhaps + to jaw #3? Is it wobble or run-out?

Put a test mandrel in the .003 error chuck and map wobble and throw (even run out mostly with not wobble) a number of times and at different diameters..

*.003 is reasonaable for a scroll 3 jaw so be very sure you are going to improve it with the repair.

Post your test findings before you true it up so others may agree with your intentions.

I tried to search that thread but could not find it. Perhaps someone marked it for reference and will post a link.
It was really interesting. Made it seem that if he had ground the jaws first that act would have made the jaws true to the back and then off to the jaw slots

Easy and practical solution is to machine in some slop in the rebate of the chuck (as previously recommended a few times) so it can be loosened tapped true in a minute or less on the infrequent occasions when good trueness is needed. I did that "fix" long ago and would never go back.

Intellectually atractive and very theoretical solution is on is to make the Chuck scroll true at all diameters. Good luck with that. Even if the scroll as new is by some magic machined true +/- .0005 throughout its range, it won't wear true and you'll be back trying to "fix" it.

Denis

dgfoster said:

very theoretical solution is on is to make the Chuck scroll true at all diameters. Good luck with that.

Click to expand...

Yep. ...even if the scroll never shifted in the chuck body.

Useful-enough workholder for do-all-ops-in-one-go working. So long as one has some stock to waste on diameters, wouldn't much matter if it were out twenty-thou, and in a different direction, each slug of raw stock.

Playin' wit yerself to expect to take work out, re-clamp, same end OR reversed, and expect 'zero' of anything but the warm glow of satisfaction.



Bill

Just to ask the obvious, but is the backplate running true on the face and the register? If thats off, anything bolted to it will be off as well.

Agree that for .003 (to .015 or so) Reds solution is best.
If unusable perhaps .008 to .015 or more off, with wobble and you were stuck with it because the return shipping costs would cost as much as the chuck then the repair justified.

Still, it won't hurt to examine it and think of what manufacturing errors caused the defect and what might be the solution to fix it.
I think all the front machining is likely done to very close tolerances because all is done in one fixturing and on a fair machine. then with flipping the part (chuck) the tram, table or fixture set is not dead exact so that is where the error comes from. Then likely they true the OD and face off the back.
With that, the inners are off a tad to the front, OD, and the back.

It would be interesting to know the manufacturing process of a quality name brand chuck and compare that to a low price chuck.

michiganbuck said:

Agree that for .003 (to .015 or so) Reds solution is best.

Click to expand...

Yah, but no more than that. You wanna cut yer losses early-on. OP HAS a 4-J arredy. Might even have collets.

'In general' for a lathe, it is best to not even START with any 3-J that doesn't have two-piece jaws AND a set-tru back, no matter who makes it. Indexers, Dividing-Heads, grinders - different loading, different choices.

Otherwise, save yer money. Get by with a cheap 4-J, even a second one parked with jaws already reversed if you must. I may quit at six. Or not.

ER and 5C collet nose-clamping goods can be had on plate or solid-body that can be HELD in a 4-J 'til one gets around to permanent backplates.

I have both on D1-3 backs, but keep the 'loose' ones as well.

That way, I can not only centre to as close as I have patience for. I can intentionally mount a collet system eccentrically for making BFBI camlocks & c. ER nose-mount can hold slugs of stock a smaller lathe could not pass through its spindle as bar even without the space consumed by a drawbar. 5C has all manner of collets cheap, hex, square, step, pot, expanding internal included.

Save yer TIME, net-net.

F***with, scrap and re-work time is time just as much as closure and clamping time is.

Bill