3jaw chuck alignment problem

[goldstar31]

It would be interesting if we had a report of the outcome from the use of his independent 4 jaw chuck or perhaps his face plate. I would reasonably expect that a drill chuck would not hold a test bar of sorts.

Just a thought or two looking again after discussing the problem with a 92 year old mate of mine.

Norm

 

[Anatol]

I told you I would return!
Our story so far : similar substantial runout on two 3jaw chucks and a collet chuck, with the same 9" piece of 1" cold rolled bar in.
The plot thickens, but tjwal, petertha, cogsy and norm are hot on the scent! (exciting isn't it?)

The 'origin' of the observed swept cone of the bar in rotation is back about the point of the fwd spindle bearing. So I thought I should check for worn spindle bearings or a bent spindle. So today I pulled the back cover off the headstock and put not one but two indicators on the aft end of the spindle, about 120 deg from each other - one 'thous', one 'tenths'. Both ran dead true! No needle movement except over a spot of varnish or something on the shaft. I was amazed. I also put a 3' bar in the collet chuck and tried to wobble it, to check for bearing slop. Rock solid. So this is heartening.

Suspicion is now directed at the spindle nose/back plate.

One of the cam locks (?) on the spindle nose would tighten too far. So I adjusted the relevant camlock pin, and tried again. This time I put an indicator on the bar in the chuck as I tightened the cam locks. I could see runout diminishing as I tightened. So I really torqued the camlocks and it did reduce the runout, but not into acceptable range.

I guess my next move is to try to shim the backplate to get the bar straight (I also need to buy/get a precision ground bar.) Since I get much the same error on all chucks, I suppose I may need to shave off the spindle nose plate. Not something I'm going to rush into!

In the meantime, I want make sure I understand the camlock system. The locks simply pull the chuck back plate hard up against the spindle surface - right?
Can the locks and pins themselves effect alignment if they're adjusted right?
The spindle nose has three 'locks' and three cap screws. I took one out this revealed a spring down a well. But that 'well' or hole does not see to contact with the lock part. What is the function of the stuff under the capscrews? Can the locks be adjusted?

thanks all!

 

[goldstar31]

Is the mounting face for the chuck FLAT? It would help if you removed all the bolts and whatever and lightly blue the back of the chuck and use it to find the possible high spots.

 

[Hopper]

...

Our story so far : similar substantial runout on two 3jaw chucks and a collet chuck, with the same 9" piece of 1" cold rolled bar in.

How do you know that piece of bar is not bent or deformed?
Have you tried other pieces? Or some ground bar out of a computer printer etc.?

 

[petertha]

Essentially yes, the camlock pulls together the tapered recess of the back plate or chuck or whatever to the corresponding spindle nose. And this could entirely be your problem if either/or are not matching.
To clarify, when you replaced from chuck A to chuck B & observed pretty much same runout issue, did you mean you bolted 2 different chuck BODIES to the same backplate? Or did you mean each chuck assembly had their own backplate. I assumed the latter. it makes a big difference. If you bolted chuck B to the same backplate then you havent really eliminated the backplate as the problem source. Thats why I suggested the easiest thing for you to do is borrow a friends complete D1-4 chuck/backplate assembly & bolt mount it. If he had 0.002" runout, yours should be comparable.

I bought a new D1-4 face plate backplate & turns out it had a partially finished spindle nose recess. So despite D1-4 tightening, it was not being clamped mated co-axially or co-radially to the lathe spindle, it had a wobble. A quick check is to blue one of the surfaces, clamp up, remove & look for partial transfer pattern. Its not bulletproof but might give you quick insight.

 

[Anatol]

Is the mounting face for the chuck FLAT? It would help if you removed all the bolts and whatever and lightly blue the back of the chuck and use it to find the possible high spots.

"Is the mounting face for the chuck FLAT?"

this is at the question isn't it

> lightly blue the back of the chuck and use it to find the possible high spots.

and if I did it with three chucks, I'd approximate Whitworths method !

If I establish that the spindle nose plate (if that's what you call it) is not 'flat' - or - not perpendicular to the spindle axis, the remedy is the same right - face off the plate?
What is correct procedure? If the line of the spindle axis is not parallel to the plane of the ways, will faced surface still be perpendicular to the axis of the spindle? I think it will but I'm having trouble visualizing this.

 

[Anatol]

How do you know that piece of bar is not bent or deformed?

It really doesn't matter at present - correct me if I'm wrong - since I'm just getting relative, indicative values - deflection through a 360 deg swing. As my wise friend Tom noted - you can draw a perfect circle with a bent stick.

" Or some ground bar out of a computer printer etc.?"
nice idea. where else in junked machinery can one find precision ground bar?

 

[Anatol]

Essentially yes, the camlock pulls together the tapered recess of the back plate or chuck or whatever to the corresponding spindle nose. And this could entirely be your problem if either/or are not matching.
To clarify, when you replaced from chuck A to chuck B & observed pretty much same runout issue, did you mean you bolted 2 different chuck BODIES to the same backplate? Or did you mean each chuck assembly had their own backplate. I assumed the latter. it makes a big difference. If you bolted chuck B to the same backplate then you havent really eliminated the backplate as the problem source. Thats why I suggested the easiest thing for you to do is borrow a friends complete D1-4 chuck/backplate assembly & bolt mount it. If he had 0.002" runout, yours should be comparable.

I bought a new D1-4 face plate backplate & turns out it had a partially finished spindle nose recess. So despite D1-4 tightening, it was not being clamped mated co-axially or co-radially to the lathe spindle, it had a wobble. A quick check is to blue one of the surfaces, clamp up, remove & look for partial transfer pattern. Its not bulletproof but might give you quick insight.

"did you mean you bolted 2 different chuck BODIES to the same backplate?"

3 chucks, each with its own backplate.
(In the process I discovered that camlock pins come with both imperial snd metric threads. In USA, metric threaded camplock pins seem rarer and more expensive. Anyone know of a reasonable supplier? )

"Thats why I suggested the easiest thing for you to do is borrow a friends ..."

Who has friends ? That's why I come here (joke). No machinist friends, at least not with D1-4

"complete D1-4 chuck/backplate assembly & bolt mount it. If he had 0.002" runout, yours should be comparable."

I may also be able to fool around swapping backplates among chucks. I also have a e1-4 4 jaw, but its such a beast - over 12" dia" I didn't want to heave it around when I had others to test with
and per Norm, blue is the next test, but I'll also try shims to straighten the bar up

 

[Cogsy]

It really doesn't matter at present - correct me if I'm wrong - since I'm just getting relative, indicative values - deflection through a 360 deg swing. As my wise friend Tom noted - you can draw a perfect circle with a bent stick.

In this case it does matter as you're holding one end of the potentially bent stick. Imagine fixing one end of a banana at a point in space (the chuck) and rotating it around the centreline of that fixed end. The other end is going to describe a huge arc, reducing towards the chuck jaws, which is comparable to what you are experiencing. The end of the banana will be describing a circle but you will be getting a lot of deflection in the 360 degree swing.

This is relatively easy to check though. Put the bar in place and mark the extreme point of deflection. Then loosen the chuck slightly, rotate the bar 180 degrees and retighten the chuck. Now find the maximum deflection again. If it matches with the previously marked point on the bar then it's a bent bar, but if it stays relative to the chuck/spindle then the bar is not your issue.

 

[Hopper]

It really doesn't matter at present - correct me if I'm wrong - since I'm just getting relative, indicative values - deflection through a 360 deg swing. As my wise friend Tom noted - you can draw a perfect circle with a bent stick.

If the bar is bent, the end will move around in a circle, causing your dial indicator to show runout. even if the chuck is running dead true. An out-of-round piece of bar will show a runout reading right next to the jaws, even in a perfect chuck.

Other sources of ground bar includes gudgeon pins from car engines.

 

[petertha]

If I establish that the spindle nose plate (if that's what you call it) is not 'flat' - or - not perpendicular to the spindle axis, the remedy is the same right - face off the plate? What is correct procedure? If the line of the spindle axis is not parallel to the plane of the ways, will faced surface still be perpendicular to the axis of the spindle? I think it will but I'm having trouble visualizing this.

Normally that is exactly how you prepare a semi-machined chuck back plate blank for mounting a chuck. It has a pre-machined internal tapered center hole profile that matches the spindle nose and the D1-4 pins. Tightening the D1-4 mechanism pulls the backplate in flush to the spindle, that fit has to be perfect to begin with. Turn the OD + face + lip/boss to match the recess in the chuck back plate. Drill mount holes & mount the chuck. If the chuck was accurate meaning jaws are true to the machined recess on the back & your backplate machined tolerance are good, it assures the best fit.

Similarly it is common to take a skim pass across a face plate to true it.

But in your case I would not recommend this specifically because you haven't come up with a solution yet. If for example your spindle nose is egged or distorted, then when you mount a perfectly good backplate / chuck, it has the potential of mounting differently every time. So machining it 'true' probably wont be repeatable next time you mount it & you are chasing your tail. I feel like if you tried 3 independent backplate / chuck assemblies with the same runout problem its statistically looking very slim that your issue is chucks & back plates. It must be aft of that - the spindle nose or the D1-4 draw mechanism or the spindle itself.

 

[Anatol]

In this case it does matter .... The end of the banana will be describing a circle but you will be getting a lot of deflection in the 360 degree swing..

You are right of course, its obvious. Thankyou, and apologies to Hopper. (The only thing I can say in my defense is that I was on a heavy dose of painkillers at the time.) I have made the measurements with several bars - of different diameters - and got similar results, but will do the test your recommend. I've also ordered a 1" precision ground bar.

 

[goldstar31]

Perhaps a couple of dial mikes reading in tenths???????

 

[Anatol]

I feel like if you tried 3 independent backplate / chuck assemblies with the same runout problem its statistically looking very slim that your issue is chucks & back plates. It must be aft of that - the spindle nose or the D1-4 draw mechanism or the spindle itself.

Thanks, this is my reasoning too, but I want to be very sure before I start hacking up my spindle nose. If I can straighten one or more chucks with shims between nose and back plate, this will determine that the issue is at that mating surface - right? Then I should measure the nose axially and radially, with as precise a device as I have - tenths.

A more general question - If I use the fwd end of the headstock as datum, this measurement will be wrt spindle axis.
If I measure off the ways, cross slide or tool post, I'll introduce a potential error IF the spindle axis is not parallel with
the plane of the ways.

Thankyou also for all the explanatory pics and diagrams.

 

[Anatol]

Perhaps a couple of dial mikes reading in tenths???????

Not quite sure what this is in response to. I have a couple mikes. How do you suggest I use them in this context?
thx!

 

[goldstar31]

Not quite sure what this is in response to. I have a couple mikes. How do you suggest I use them in this context?
thx!

You have or should have accessed the lathe test charts in Schlesinger which shows the way 2 dial gauges are used in several tests. Whether they are 'fine' enough is a answer that only you can supply.
You test the lathe according to -say- Schlesinger and if corrections are required you are the only person to literally get on with revising the alignments discovered.

We can only do so much. The rest is really up to you.

Norman

 

[Cogsy]

A more general question - If I use the fwd end of the headstock as datum, this measurement will be wrt spindle axis.
If I measure off the ways, cross slide or tool post, I'll introduce a potential error IF the spindle axis is not parallel with
the plane of the ways.

I'm not sure why you would be measuring from any datum point for runout, all you need is a fixed point to take measurements from (i.e. your indicator should be secure when you take a reading). You're not going to be able to measure relative distance anywhere near as precisely as runout with an indicator where you're measuring differences rather than absolutes.

You mentioned you have ordered a ground bar, hopefully it has an MT taper you can fit directly to your spindle. This bar will simply amplify any runout issue with the spindle and allow for easier measuring with indicators. If it runs true to your spindle bore then the problem narrows down to the spindle register, chuck backplates or chucks themselves. It will certainly get you closer to a solution.

To be honest, if this is a new error I can't see it being a bent spindle without some major event occurring. My money is still on the chuck/backplate/spindle interfaces as the problem.

 

[DJP]

I am curious if this lathe was purchased used. I recently bought a quality 4 inch mechanic's vice at a deep discount only to find that it was bent. The jaws wouldn't close square and the threaded shaft would bind. I returned it for a refund and vowed to be more careful before taking home a bargain.

Is there a possibility that your lathe was dropped during transport? Can you use the tail stock with a live centre to see alignment problems deep in the casting?

Your only solution may be to shim the head stock to get it close enough.

I'm sure that we would all like to see this issue resolved.

 

[petertha]

If I use the fwd end of the headstock as datum, this measurement will be wrt spindle axis. If I measure off the ways, cross slide or tool post, I'll introduce a potential error IF the spindle axis is not parallel with the plane of the ways.

The 2 important surfaces that the camlock system registers to a corresponding mounted accessory on are the tapered cone (end of red arrow) and the rear face (end of green arrow). The nose front (orange arrow) doesn't contribute anything. Notice this back plate adapter is just a hole, it is not in contact with the lathe spindle. So when back plate & spindle are mounted together, cones coincide with one another and the faces are maybe 0.001" apart constant offset. Tightening the cams then bring the faces together. Ideally in a nice fitting arrangement the back plate has just a teeny bit of 'stick' &then pops off. Its a delicate fit.

So as pointed out ^above^ it matters not where you mount the DTI when checking the lathe as long as it is fixed. You are rotating the spindle & looking for deviation. For example if the spindle cone is not circular the DTI will wander in & out showing this deviation. Same thing with the nose flange face.

 

[DJP]

With this cam locking system of retention it may be prudent to tighten the cams in stages. Tighten each cam to 1/2 of the final torque, then tighten them again in sequence coming closer to the final torque applied which should be equal for all cams. You may need a small torque wrench to do this properly. The process is consistent with tightening wheel nuts of a car to ensure that the rotors are not stressed adding runout.

Just one more idea for discussion to solve this issue.