Atlas 618 workpiece wobbles

[WCraig]

TLDR...

What could be causing stock in my newish 4-jaw to wobble 25-30 thous when projecting just 2.5" beyond the jaws? Bad chuck? Bad spindle bearings? Something else?

Long, possibly meandering story...

So, my little Atlas 618 lathe has had a hard life these past 75 years. There are deep scars to the front of the compound where the chuck has crashed. Hard. Multiple times. I believe the spindle is bent. A drive centre or an end mill holder in the morse taper socket of the headstock spindle wobbles by about 3 thous near the spindle nose. The wobble gets more pronounced when I measure further out from the nose.



BUT, I don't understand why the 4-inch, 4-jaw won't hold work true. The backing plate for the 4-jaw was machined to fit the chuck on the lathe. It ought to be true to the real rotation of the lathe, no? I actually took a fresh cut off the face of the backing plate today. I used marker to ensure that I cut the entire surface that the chuck body mounts to. Despite this, a dial indicator showed that fresh face was out of true by about 1/2 thou or a little more!?!



I'm guessing that the spindle is running differently when there is tool pressure or not. I would like to add more preload but I can't get the collar (part M6-32, per the diagram) to turn any tighter. The end of the set screw that is supposed to be in that collar is all buggered up and the lead shot is long gone. The whole collar has been chewed up with locking pliers in the past. Nonetheless, it is as tight as I can get it.

Back to the 4-jaw. After taking the fresh facing cut, I reassembled the chuck body to the backing plate and put it back on the lathe. All mating parts were cleaned. The body of the chuck runs out less than 2 thous whether measured close to the backing plate or out at the jaw end. That makes sense to me since there is a tiny amount of play between the chuck body and the registration boss on the backing plate.



I then dialed in a piece of 1 inch drill rod in the 4 jaw and got it centred within a thou near the jaws. This is when I found that I still have relatively large runout just 2.5 inches out from the jaws. I measured it at about 30 thous on the first test. In case the stock was bent, I rechucked it at 180 degrees to the first attempt. The runout then measured about 25 thous.



BTW, if I alternately press and pull on the projecting workpiece, I can move it about 3 thous in each direction--6 thous total movement across the bed. That seems like a lot but no where near the 25-ish thou wobble.

I guess it is possible that the chuck jaws do not hold the workpiece in line with the body of the chuck. The chuck was from Shars and was pretty inexpensive. Is it really possible that it was that badly made? Is there any test I can do to determine if the chuck is at fault?

I do plan to upgrade to a better lathe but I'd still like to understand this problem better. Thanks for any suggestions or pointers.

Craig

 

[cross thread]

Have you put an indicator on the inside of the spindle ? (In the MT bore) If you do that , and lets say its good .001-.003 then the problem is the chuck . It sounds like a jaw has been ground incorrectly , or the body it's self . A possible test , put a micrometer on the chuck face ( the surface that goes to the face plate) now measure the jaw slots for parallel in relationship the to the face that mates with the FP . In your first picture , the part of the spindle right in back of the threads , the part that the chuck will register on , put an indicator on that .

 

[WCraig]

I'm confident that the spindle is bent. I've measured multiple MT tools in the spindle and they all show an increasing wobble away from the headstock. Measuring inside the MT socket isn't easy as there is a little damage right at the end. I've made sure that nothing is protruding and MT tools do seat OK.

My question is more about understanding why stuff in the chuck is so far out of true.

Craig

 

[francist]

Maybe there’s a wonky jaw or two? Not sure how that would happen but just reaching for ideas. If it’s a four jaw independent you could swap the jaws from location to location and see if the aberration follows any particular one?

-f

 

[Manual Mac]

If your Spindle is slightly bent, I believe you could still zero out the 4 jaw chuck in one place, but if the DI is moved either towards the headstock or the tail stock it will now have more runout the further you move it?
or maybe I’m not understanding this correctly ?

 

[WCraig]

Maybe there’s a wonky jaw or two? Not sure how that would happen but just reaching for ideas. If it’s a four jaw independent you could swap the jaws from location to location and see if the aberration follows any particular one?

-f

I took your idea a little different direction. At the moment, the chuck is separated from the back plate. I set it on a flat surface (which happened to be a rotary table but I was just using it as a flat surface) with the jaws holding the workpiece straight up. In theory, the workpiece should be at right angles to the surface. I then rigged a dial indicator to measure between the side of the chuck and the workpiece. I labelled the 4 jaws (N-W-S-E) and measured the relative distance from the chuck side to the workpiece two times--once near the jaws and again near the end of the protruding workpiece. The workpiece was leaning to the S jaw by 8-9 thous. I should redo this with the workpiece rotated 180 degrees and see if it now leans to N or not. Nonetheless, I don't think the chuck is out by the 25-30 thous I saw earlier.

Craig

 

[WCraig]

If your Spindle is slightly bent, I believe you could still zero out the 4 jaw chuck in one place, but if the DI is moved either towards the headstock or the tail stock it will now have more runout the further you move it?
or maybe I’m not understanding this correctly ?

Maybe I'm wrong, but I'm thinking of it like a crankshaft. The front and rear bearings of the headstock determine the true center of rotation. Intermediate bends don't matter as long as the chuck's backing plate is machined true to that center of rotation. (And the chuck is mounted squarely to the backing plate.)

This is making me think that maybe the bearings are bad? I pulled the spindle when I first got the lathe and then life got in the way and I haven't done much with it until recently. When I had it apart, I thought the bearings and races looked OK. I do remember having a _lot_ of trouble getting the back bearing into place during reassembly. The inner race was a lot tighter on the spindle than what other people seemed to have described. I got it together by freezing the spindle and warming the bearing over a light bulb. I don't feel any play in the spindle but maybe I never really got the bearing seated right? It was a real race to get all the parts together before they warmed/cooled.

Craig