Chuck run-out problem on a Rotary Table

[SillyOldDufferModerator @sillyoldduffer]

This week's educational project was my first attempt to cut a gear. I made a fly-cutter formed to an approximation of the Meccano gear and attempted to reproduce Meccano Pinion 26. (19 teeth on a 13.7mm diameter blank, 38 or 40DP.)

I eventually produced this flawed item.

Most of what's wrong with the gear is due to mistakes and/or the inadequacies of my highly amateur attempt at making a form-tool.

But the picture shows that the gear is cut far more deeply on side than the other, indicating a serious run-out problem.

The error is about 0.35mm. Tests with a dial indicator show that It's not due to to misalignment of the chuck on its adaptor, or to misalignment of the adaptor to the rotary table. The error is only serious on rod held by the chuck's jaws.

Graphing the error relative to jaw position shows that jaw 2 is slack. I was able to drastically reduce the run-out by shimming the gear blank holder at jaw2 with some aluminium cut from a fizzy drinks can.

My problem is, what's causing the run-out? I think it's a new problem. The chuck in question is only used on my rotary table and it's about 2 years old. I've used it many times in the vertical position without noticing any errors. It's not much used and it's had an easy life.

Recently I removed swarf from the scroll and jaws and may have done something to it then. Cutting the gear is the first time I've ever used the chuck in the horizontal position. There's no obvious damage or dirt on the scroll or jaws. The error is present when the chuck is vertical as well as horizontal, so it's not caused by the orientation.

I'm baffled and will welcome any suggestions.

Thanks,

Dave

[SillyOldDufferModerator @sillyoldduffer]

[Paul LousickParticipant @paullousick59116]

Hi Dave,

The first thing to check is if you have replaced the jaws in the correct position. They should be numbered and have a corresponding number on the chuck body. Replace the jaws in order, starting with No 1.

Sometimes on an old worn chuck the jaws can be assembled in a different order. This is a trial and error process and does not always work. Check by inserting a round bar in the chuck and a dial indicator.

Paul

Edited By Paul Lousick on 06/10/2016 21:23:46

[Michael GilliganParticipant @michaelgilligan61133]

Sorry, Dave

Despite your meticulous documentation; the problem itself doesn't really surprise me.

A four jaw, or a collet would be much more appropriate.

[Yes, I know we are having a bit of a run on such issues]

MichaelG.

[SillyOldDufferModerator @sillyoldduffer]

Thanks for the suggestion Paul. Not that unfortunately. I've just double checked.

Ta,

Dave

[SillyOldDufferModerator @sillyoldduffer]

Posted by Michael Gilligan on 06/10/2016 21:26:17:

Sorry, Dave

Despite your meticulous documentation; the problem itself doesn't really surprise me.

A four jaw, or a collet would be much more appropriate.

[Yes, I know we are having a bit of a run on such issues]

MichaelG.

That's a really good hint Michael. At the moment a 4-jaw feels like the right way to go, miserly though I am.

Is a 0.35mm run-out not excessive though? None of my other chucks are anywhere near that bad.

Thanks,

Dave

[Clive FosterParticipant @clivefoster55965]

If gear cutting its best to use a 4 jaw chuck that can be adjusted to run dead nuts true.

I presume the starter rod was straight, true and properly round. I've spent time chasing my tail on similar before realising that the material was not quite straight. Can be surprisingly hard to see half a mm / 20 thou bend on short piece.

What breed is your 3 jaw and how good is it supposed be? Comparing with the test sheets and book with my new in 2004 Bison 200 mm 3 jaw the run out on yours is about 10 times worse than new spec and about double acceptable values on a used chuck. Going by earlier threads even a relatively inexpensive chuck should be 0.10 mm or less TIR when in good order.

Clive

[Michael GilliganParticipant @michaelgilligan61133]

Posted by SillyOldDuffer on 06/10/2016 21:34:17:

Is a 0.35mm run-out not excessive though? None of my other chucks are anywhere near that bad.

.

It does seem rather high for a little-used chuck … especially if it was better before you cleaned it.

I would have a very good look for bruises on, or tiny bits of swarf embedded in, the scroll.

MichaelG.

[SillyOldDufferModerator @sillyoldduffer]

Posted by Clive Foster on 06/10/2016 21:41:23:

If gear cutting its best to use a 4 jaw chuck that can be adjusted to run dead nuts true.

I presume the starter rod was straight, true and properly round. I've spent time chasing my tail on similar before realising that the material was not quite straight. Can be surprisingly hard to see half a mm / 20 thou bend on short piece.

What breed is your 3 jaw and how good is it supposed be? Comparing with the test sheets and book with my new in 2004 Bison 200 mm 3 jaw the run out on yours is about 10 times worse than new spec and about double acceptable values on a used chuck. Going by earlier threads even a relatively inexpensive chuck should be 0.10 mm or less TIR when in good order.

Clive

Hi Clive,

I'm confident that the chuck is at fault because once I realised there was a problem I got out my test rod.

The chuck is far eastern and inexpensive. What's odd is that it seems to have got out of order suddenly. I think I would have noticed it had poor run-out before now, though it's true that cutting this smallish gear must have been it's most severe test.

All my other cheapo chucks are better than 0.1mm, but perhaps I've just been lucky.

Another vote for a 4-jaw from you too. I'm warming to the idea!

THanks,

Dave

[MWParticipant @mw27036]

Bad grinding on the jaws? bellmouthing? a close inspection of the jaws themselves may offer some clues. Especially looking at the circle they form.

Michael W

[Ian PParticipant @ianp]

Dave

The problem is not really due to chuck run-out but more one of technique.

Assuming you want to create a blank and then transfer it to a different machine to cut the teeth than a four jaw on the RT would allow you to precisely centre the blank, however….

I think I would start with an oversize blank with no centre hole stuck in your inaccurate 3 jaw and use the RT to rotate the blank continuously while you mill the OD to the final diameter. For a truly concentric bore using a small slot drill (well, smaller than the final ID) create the bore again by rotating the table whilst gradually lowering the cutter, eventually using the cutter to mill the bore to the correct ID size.

If you already had a stock of blanks you want to use then stick a bit of bar in the RT chuck and mill it to the correct diameter to create a stub mandrel that the year can be pressed/locked on to.

Ian P

[SillyOldDufferModerator @sillyoldduffer]

Quick lack of progress report with special thanks to the two Michaels. I'm going to have a good look at the scrolls later, but the chief suspect now is the jaw grinding. If this is the fault, it means that Inspector Clouseau here has been using a squiffy chuck for over 2 years without noticing!

This picture is of the jaws in my lathe. The chuck is satisfactory and you can see that there's a reasonable triangle where the jaws meet. (Jaw 2 is on the left.)

Next picture is the offending Rotary Table Chuck, where the jaws meet in a noticeably coarse triangle.

Ian P's suggestion about using a different technique has considerable appeal and reminds me again that there's usually more than one way of skinning a cat. And now I wonder how much of the quality work I admire so much is produced by people who know how to overcome the shortcomings of their tools, not by using high-end kit? If that's the name of the game I still have lots to learn!

Thanks,

Dave

[Tim StevensParticipant @timstevens64731]

It does sound like a seriously bruised scroll in the exact position you need.

One solution could be to set the table horizontal (for now), loosen the chuck on the backplate/table, and allow it to move about. This might mean overcoming any register to allow the floating to happen. Then fit the bare spindle/mandrel in the chuck (ie the bit that the gear is attached to, if separate, or the gear and spindle if integral) and nip up the chuck holding bolts just enough to hold it. Clock the surface of the blank as the table is rotated, and note the high point. Then tap the chuck sideways at this point with a plastic or brass hammer, to centralise it. It will take some time and cursing, but eventually it will be right. Tighten the bolts further, and check that nothing has moved. Then put the table back upright and cut your gear. Once the blank is centralised on the backplate, it will remain concentric, and so will the gear cut round it, however cockeyed the chuck etc.

Worth a try?

Tim

[SillyOldDufferModerator @sillyoldduffer]

Posted by Tim Stevens on 07/10/2016 17:48:12:

It does sound like a seriously bruised scroll in the exact position you need.

…

Worth a try?

Tim

Yes I think so Tim.

Today's experiments revealed that the run-out varies from as bad as 0.4 to as 'good' as 0.15mm, this effect being achieved by removing and reinserting the jaws. I would have thought that a faulty jaw would be consistently wrong.

Then I measured the outer-jaw set and found that it also has exactly the same run-out problem. Finding that two jaw sets have the same error got me suspecting the scroll.

I had a good look at the scroll through the slots and it seems OK, but then I wouldn't expect to see a small error.

Thinking critically about things I've done that could damage a scroll I came up with an 'own-goal' scenario.

Is it possible to damage a scroll by winding the rotary table round and catching the end of a 'left in' chuck key on the body of the table? I've done that a few times and thinking about it, the torque resulting from a 90:1 reduction drive on the far end of a chuck handle might be super-human.

I'm careful never to leave a key in the chuck of my lathe, but have certainly been careless with doing just that on my rotary table.

Thanks,

Dave

[MWParticipant @mw27036]

I think you could be forgiven with an RT, I tend to remove my tools from what ever they were fiddling with and have a generic "Swarf-come tool pit/ tray" set up infront of me; whose contents could vary in anything from a mic, to a spanner, to nylon swarf to a spring steel pocket ruler and black sharpie. I find it very useful and comforting. I also have a trolley behind me containing my hammers…i always like to have them close to hand.

 

Getting back to your Chuck, The RT, is it a front mount chuck on a tee slotted base? or a threaded nose? The latter we can pretty much rule out being the cause in all but the most battered of circumstances.

The front mount is an interesting one because it's difficult to centre, I puzzled myself with how i could set it up til i realised i could turn a dowel to fit the back of the chuck to the RTs MT2 spindle bore, jobs' a goodun.

But i suspect you've already ruled that out, and are pretty sure it's the chuck, I did notice the chinese one has a sharper angle ground on the tooth and they do form a much rougher triangle.

This is strange, because as you say, it was working fine before, I would look at the chucks own scroll and the corresponding jaws scroll, it could be that jaw 2 has worn thin and it might have some noticable slack?

PS. I really doubt it's this also, but it just occured to me that when fitting the jaws, i always look for the bite point when going round the numbers. I slip it in and out as i carefully turn the key and wait for it to just take it. i never turn it beyond the bite point or it could go in a odd order?

normally this is very very obvious, but maybe theres a chuck out there where its very subtle? 

Michael W

Edited By Michael Walters on 07/10/2016 18:40:09

[Michael GilliganParticipant @michaelgilligan61133]

Dave,

It's a bit of a faff, but; could you 'clock' the travel of one jaw as you wind it in? … I'm now wondering if the scroll is not concentric with its bearing; which would give a 'periodic error'.

[A long travel DTI would be convenient for this, but I'm sure you could improvise]

A plot of displacement against key rotation should, ideally, be a straight line.

Do two runs; one with the jaw biased to each flank of the scroll.

[Rubber band might suffice]

MichaelG.

[Anonymous]

Posted by SillyOldDuffer on 07/10/2016 11:46:10:

And now I wonder how much of the quality work I admire so much is produced by people who know how to overcome the shortcomings of their tools, not by using high-end kit?

Not me, I need all the help I can get. I try to use quality tooling where I can, albeit all old and secondhand.

Neither of the pictures of the chucks look inspiring. On the 3-jaw chuck on my lathe, which I rarely use as the jaws are badly worn, the jaws meet with a neat triangle in the middle, and the sides of the jaws mate without gaps. Interestingly the 3-jaw chuck on my dividing head mates neatly but doesn't have a triangle in the middle, so I wonder if it was a 'special' for the dividing head, intended more for precision than gripping power?

Andrew

[SillyOldDufferModerator @sillyoldduffer]

Posted by Michael Walters on 07/10/2016 18:33:36:

… Getting back to your Chuck, The RT, is it a front mount chuck on a tee slotted base? or a threaded nose?

… could be that jaw 2 has worn thin and it might have some noticable slack?

… maybe theres a chuck out there where its very subtle?

Michael W

Edited By Michael Walters on 07/10/2016 18:40:09

Hi Michael,

It's a front mount chuck on a tee-slotted base, and I've confirmed that the alignment at that point is OK.

I've looked at the chuck threads on all three jaws and they're unworn.

I wondered if I was doing something daft loading the jaws but it's standard stuff. I also tried loading the jaws in the right order into the wrong slots and got the expected result, that lowest run-out was obtained by putting the jaws in the correct slots, 1,2,3. What I forgot to do was to note down the measurements so see if the error follows a jaw round the chuck, doh!

Ta,

Dave

[SillyOldDufferModerator @sillyoldduffer]

Posted by Michael Gilligan on 07/10/2016 18:42:58:

Dave,

It's a bit of a faff, but; could you 'clock' the travel of one jaw as you wind it in? … I'm now wondering if the scroll is not concentric with its bearing; which would give a 'periodic error'.

[A long travel DTI would be convenient for this, but I'm sure you could improvise]

A plot of displacement against key rotation should, ideally, be a straight line.

Do two runs; one with the jaw biased to each flank of the scroll.

[Rubber band might suffice]

MichaelG.

Yup, I'll have a go at that. Have to be Sunday though as I'm busy tomorrow.

Cheers,

Dave

[SillyOldDufferModerator @sillyoldduffer]

Fixed it, hurrah!

I did the check suggested by MichaelG and didn't find anything abnormal.

Then I discovered that the chuck mount could be rocked very slightly on the rotary table, and that the degree of rocking corresponded with the chuck run-out. ( Not obviously loose, it was necessary to give the chuck a good hard pull to move it. )

Removing the mounting plate from the rotary table revealed a fettling problem. Two of the chuck mounting bolts were long enough to bottom out on the rotary table, one being longer than the other. Making it worse, some swarf had got into the holes such that the chuck was jacked off the rotary table and no longer parallel with it.

Cleaning the swarf off and reducing the length of the chuck securing bolts by a millimetre has done the trick. Reassembled, the chuck run-out at 50mm is a much more reasonable 0.03mm.

I would have noticed the problem before if I'd ever used the the rotary table without the chuck. All became clear as soon as the chuck mounting plate came off .

Lesson learned – check everything!

Thanks again for all the advice and pointers.

Dave

[MWParticipant @mw27036]

Posted by SillyOldDuffer on 09/10/2016 17:43:31:

Fixed it, hurrah!

I did the check suggested by MichaelG and didn't find anything abnormal.

Then I discovered that the chuck mount could be rocked very slightly on the rotary table, and that the degree of rocking corresponded with the chuck run-out. ( Not obviously loose, it was necessary to give the chuck a good hard pull to move it. )

Lesson learned – check everything!

Thanks again for all the advice and pointers.

Dave

Ah, thats the sort of problem you're always relieved to have, in the sense that it could have been more terminal with the chuck or RT.

I get a similar problem on my lathe when the gibs come too loose and i end up with tapered work. So i have a habit of checking the dia. on both ends of the work to make sure it's turning parallel as the gibs tend to wriggle loose after a hefty job.

Michael W

[Michael GilliganParticipant @michaelgilligan61133]

Good result, Dave

MichaelG.

[Author]

Posts