Taper

[140mower]

Perfect. Thanks Mike.

As an aside, I really enjoy your use of the English language to make clear consise statements about very complicated things using relatively few words. It certainly works for me. My answer to that same question would have been 4 pages long and added nothing else. So, I'm jealous of your talent.

Edit - and then I see you added to it and essentially doubled it plus added a photo...... LMAO!

I think you might actually have one of my many faults too. I write something and then realize that not all readers are the same and Somebody who thinks differently might miss my meaning. Then the words get added. Then I worry I might have insulted the original reader. Then I realize I need to fix more. Next thing you know, I have 4 pages of mostly unnecessary verbia.

Not to worry, we all have moments of verbal diarrhea. I am just happy to see someone going through the effort of helping those of us in need.

 

[Susquatch]

but then when he cut stock - even short stock - he does. well, one sure way for that to happen is for the stock to be held by the chuck in a way that isn't concentric to the axis of rotation.

For the most part I think we loudly agree too. But not on the particular point I quoted above. Which is where I stumbled before.

I'm gunna, try to take a page out of @Mcgyver's book and be short and suscinct.

It doesn't matter how out of alignment a part is held in the chuck jaws. When it is subsequently turned, the turned part will be concentric to the spindle axis.

This assumes good bearings and a spindle with no perceptible run-out.

To create a taper on a short part, the spindle axis MUST be out of alignment with the ways (bent, worn, damaged, twisted, etc).

This assumes that the cross-slide is not slipping and the cutter is on center.

You have a nice lathe and I bet it is reasonably well aligned. Try it!

Use your 4-jaw or a big shim to mount a 5 inch part say 2" in diameter with a 3" stick out that is deliberately mounted as you described.

in a way that isn't concentric to the axis of rotation.

Use unequal shims to make it obcenely out of whack. Then cut it until the cutter cuts clean all the way around for all 3 inches of stickout. The final part will be concentric with the axis of the lathe. It has to be.

 

[mbond]

I think we are talking across purposes again. Ignoring all the bad stuff that could happen with a severely compromised machine, any cut will be concentric with the spindle axis. Exactly as you say, it has to be. But when you take that part out of the lathe, and measure it you won't get 'conformity' between the sections that were chucked and turned

 

[Susquatch]

But when you take that part out of the lathe, and measure it you won't get 'conformity' between the sections that were chucked and turned

Of course not! LMFAO. I did actually address that very point earlier by saying that the turned portion would be concentric. I deliberately left the stub out of that statement.

The OP never said he did that. But who knows, maybe that's what he did. If so, it's a clear no-no. Besides, that's not a taper - more like two non-concentric cylinders.

Virtually no hobby level 3jaw chuck is perfect that way. Even those that are really good are usually only good at one diameter and don't do as well for other diameters. My own 5C collet chuck is actually very good, and my 3jaw is ok, but I'd never trust it completely that way. Jaw tension, material compression, and host of other factors make it very difficult to do. That's just one reason why order of ops is so important. Heck, I don't usually even trust indexing the part to the chuck! Once it is done I'll almost always part the stub off. If I have to keep the stub, I'll dial the good end in on my 4jaw so I can turn the stub.

 

[mbond]

Sometimes I have a hard time keeping my mouth shut. And this seems like one of those times.

 

[Susquatch]

Sometimes I have a hard time keeping my mouth shut. And this seems like one of those times.

Oh what the hell! Go for it. Why should I be the only one with verbal diarrhea? LOL!

If not, that's ok. I "think" we are all on the same page now.

 

[mbond]

Haven't I put my foot into it enough for one day?

 

[Susquatch]

Haven't I put my foot into it enough for one day?

Not really. Wander around the forum a bit. Pretty sure I step in it at least a half dozen times the average day and then..... I leave tracks on the brides clean floor afterwards too.....

 

[mbond]

Let me try again. At least to clarify my own mind. However badly I expressed this before, something keeps bugging me about this one.

I am probably talking about effects that are too small to cause the observed problems, and here are a few things that might better belong in the other thread, but since I am encouraged to spew ...

First, there really isn't such a thing as a single spindle axis. No matter how good the bearings, there will always be _some_ level of eccentricity. And since we assume ball or roller bearings, that eccentricity will be of a special kind - it will cycle around the notional center with a period proportional to the size of those balls / rollers to the other parts of the bearing.

Second between mounting plate, the chuck itself and the way it grips the work, some degree of eccentricity is assured. Unless the machine is really bad, the largest source is probably user error (chucking the work at an angle).

These two factors alone will produce a series of spiral shaped ellipses in the work. If enough cuts are taken at speed x feed rates that aren't natural multiples of those other rates, the work should become round along an _effective_ axis of rotation.

But there is another factor that will produce the taper of my imagination (tm). And that depends more on the direction of feed. Feeding away from the chuck will be much more resilient (2/3 more theoretically) than feeding towards the chuck. And that relates to to tool pressure versus the slenderness ratio of the work. No matter how stiff the material, or how light the cut, some lateral pressure will be exerted on the work by the tool. If the material does not remain in the elastic range, then it will simply break. but if it does, then it will reliably make a taper that grows larger the further away from the chuck that the tool goes. In the ranges quoted in this thread it may by microns

With a perfect chuck, the reaction forces will be 4 times greater with an unsupported work piece versus one supported with a tail stock. An imperfect chuck will have higher forces on some jaws and some parts of each jaw - which will make everything worse. I thought I had an idea of a way that a poor chuck could cause microns turn into tenths or maybe even thou, but I totally doubt what I was saying before


probably I am way over thinking this, but I don't think I have ever been happy with the Hogan's Hero's line - I know nothing. I want to know. And mistakes are a great way to learn

 

[Susquatch]

Let me try again.

Sure! Why not! We all learn from healthy discussions like this.

I am probably talking about effects that are too small to cause the observed problems, and here are a few things that might better belong in the other thread, but since I am encouraged to spew ...

I believe in my heart of hearts that this is true. I'm no big time lathe repair shop, but I've aligned a few lathes quite successfully using my own approach. What I can tell you with confidence is that you are right - there is no such thing as a perfect spindle. Nor is there a perfect measuring tool. We humans have never figured out how to measure zero.

The OP's problem is big enough to measure reliably within the limits of what can be measured reliably.

That said, I think it helps to be able to recognize what can be improved upon and what is too little to worry about. It also helps to have an idea of the kind of precision you can expect from the various kinds of lathes available. A top notch precision lathe will usually run circles around a clapped out machine no matter how good it once was, and it will usually out run a hobby lathe and tool room lathe. Our expectations must be reasonable for the machine we are considering.

First, there really isn't such a thing as a single spindle axis. No matter how good the bearings, there will always be _some_ level of eccentricity.

Yup. You are correct. See also above about reasonable expectations.

I think the OP's measured taper is worse than I'd expect his machine to produce. Also, because it is taper (not runout), I think it can probably be improved. But there will be a limit beyond which further attempts to refine it are futile.

And since we assume ball or roller bearings, that eccentricity will be of a special kind - it will cycle around the notional center with a period proportional to the size of those balls / rollers to the other parts of the bearing.

Yes, I'd expect some level of eccentricity too. I don't know if it's period would be proportional to the size of the balls/rollers to the rest of the bearing. I'd have to see an analysis of that to get any sense of comfort about that.

Second between mounting plate, the chuck itself and the way it grips the work, some degree of eccentricity is assured.

Yes, but it isn't normally at a level that needs to be worried about. If it is, then it's time to accept it for what it is or do something about it. On my lathe, it's less than a tenth. I'm quite happy with it and have no desire to try and fix it.

Unless the machine is really bad, the largest source is probably user error (chucking the work at an angle).

This is where we fell off before. It doesn't matter if the work is chucked at an angle. After turning, the axis of the final product will be concentric with the axis of the spindle - assuming it is solidly held, is short enough to not droop measureably, and the part is never removed from the chuck till the work is complete. Yes, the stub held in the chuck will not be aligned with the rest of the work, but we are not going to reverse the part or use the stub in any way. And yes, it will have some eccentricity, but not enough to worry about.

These two factors alone will produce a series of spiral shaped ellipses in the work. If enough cuts are taken at speed x feed rates that aren't natural multiples of those other rates, the work should become round along an _effective_ axis of rotation.

I don't know about a spiral being the result. I've never seen that happen. If it does happen, it's at a level that I can't see or measure. I've done this enough times to know that it's not something I'd worry about.

But there is another factor that will produce the taper of my imagination (tm). And that depends more on the direction of feed. Feeding away from the chuck will be much more resilient (2/3 more theoretically) than feeding towards the chuck. And that relates to to tool pressure versus the slenderness ratio of the work. No matter how stiff the material, or how light the cut, some lateral pressure will be exerted on the work by the tool. If the material does not remain in the elastic range, then it will simply break. but if it does, then it will reliably make a taper that grows larger the further away from the chuck that the tool goes. In the ranges quoted in this thread it may by microns

You should re-read your own thread about cribbage pegs.

All of what you say can happen but there are ways to reduce the chances and the level of impact. In my experience the impact is neglidgeable. It might not be on a worn out machine or one with bad bearings.

About all I can say for sure is that every lathe I have aligned worked out well within reasonable limits.

With a perfect chuck, the reaction forces will be 4 times greater with an unsupported work piece versus one supported with a tail stock. An imperfect chuck will have higher forces on some jaws and some parts of each jaw - which will make everything worse. I thought I had an idea of a way that a poor chuck could cause microns turn into tenths or maybe even thou, but I totally doubt what I was saying before


probably I am way over thinking this, but I don't think I have ever been happy with the Hogan's Hero's line - I know nothing. I want to know. And mistakes are a great way to learn

Ya, I agree that you are over thinking this as it relates to the OPs concerns.

Some might say who cares beyond that. Personally, I don't think for one minute that you are overthinking it with respect to understanding what could happen under other circumstances as well as exploring the theoretical limits of what is practical and achievable.

Sometimes it's worth it to set up a series of experiments to explore such things. If you do, I'd be keenly interested in following along.

Most of my experiments of this nature have been done while aligning a lathe. Especially comparing various methods others have used. When the dust settled, I have ended up settling on the dumbbell bar as the best method (for me) for minimizing unintentional tapers.

I am also slowly working away at improving my methods and I have an unfinished project on an improved dumb bell.

 

[Dabbler]

Feeding away from the chuck will be much more resilient (2/3 more theoretically) than feeding towards the chuck.

it would seem so, but it is not true. Reference Joe Piezynski's videos making plastic needles of any arbitrary length from 1" nylon or delrin.

With a perfect chuck, the reaction forces will be 4 times greater with an unsupported work piece versus one supported with a tail stock.

If by 'reaction forces' you mean the cutting forces on the work piece, they are the same no matter where you are along the bar. If, you mean is the stiffness of the work piece, it is in the order of the inverse of the length cubed. This is a standard beam calculation you learn in engineering school.

 

[mbond]

It has been about 20 years since I was in engineering at university. I don't don't work in the field, but I hope that I haven't forgotten mechanics 101. I'm constantly reminded of my other failings, so maybe my hopes are too optimistic

When I say reaction forces, I mean those forces applied by the chuck on the work to keep it in place. Because the tool applies its force (however little or much that is) at some distance away, and the chuck is the only support, the chuck itself must apply a moment (torque) to resists that force. If the work is supported at both ends, then the force can be resisted directly (ignoring the real world problem of deflection in the work) and so no moment is needed - thus theoretically reducing the required reaction by a factor of 4. The real world value will be less by some amount

I still think that I am way over thinking this, and the effects that I'm taking about are orders of magnitude smaller than what the OP is reporting. Which makes my posts on this thread unhelpful at best. but I was encouraged to spew, so ... i don't feel so bad about doing it

re the direction of feed, I think it depends a lot on the material properties of the work. Probably that could be a whole other long conversation, but probably the key factor is the 'thrust angle' generated by the tool. That's another long conversation formenting too it seems

 

[Susquatch]

When I say reaction forces, I mean those forces applied by the chuck on the work to keep it in place. Because the tool applies its force (however little or much that is) at some distance away, and the chuck is the only support, the chuck itself must apply a moment (torque) to resists that force. If the work is supported at both ends, then the force can be resisted directly (ignoring the real world problem of deflection in the work) and so no moment is needed - thus theoretically reducing the required reaction by a factor of 4. The real world value will be less by some amount.

These normal torque forces are not really minor and neither are the tool pressure forces - both of which must be reacted by the chuck. However, when cutting a test bar or taper to evaluate alignment we always take really light skim cuts which actually are relatively minor.

The test the OP did does not use the tail stock so no relief there either.

Re the direction of feed, I think it depends a lot on the material properties of the work. Probably that could be a whole other long conversation, but probably the key factor is the 'thrust angle' generated by the tool. That's another long conversation formenting too it seems

I can't remember what we were debating here anymore. I think it was a bit of a red herring as far as how it might affect a taper. It matters a lot when making needles which are prolly impossible to do cutting away from the chuck. But cutting a big taper can be done in either direction without changing the outcome.

You feeling more comfy about this yet?

 

[mbond]

How long ago did i say i should stop talking?

 

[Brent H]

@patro :

4 jaw chuck in your lathe and put in a solid (say 2” diameter) piece and a 6” stick out. Set the chuck to have zero run out very close to the chuck. Take light cuts until you are shaving the whole stock. Measure. If you are not cutting a taper _ nice. If you are, you may need to adjust your headstock. Smaller lathes, shim. Larger lathes may have bolt/nut arrangements to move the headstock. I needed to shim my Utilathe to get the taper out. About 0.003”. Once that taper cut is out you can then set up your tailstock.