Concentricity Issues

[[Ni...]]

Hey all,
Need some help figuring out why my concentricity results on our Zeiss Prismo always greatly differ from those answers achieved manually on a lathe using indicators at various locations. I'm clearly doing something incorrectly so if I could get some assistance that would be appreciated! The situation is as follows:

Alignment -
End face picked up as a plane for Spatial Rotation
OD picked up in 3 locations as a cylinder X & Z origin
End face also used as a Y origin.
(everything is done scanning)

The Datum A Ø is measured as a cylinder in 2 different locations and the Ø7.088" is measured the same as well, only as a circle instead. (Cylindricity of the Datum A is always intolerance @ .0004" max)
Note: The parts OAL is 26.500".

Tube.jpg

[[Ja...]]

Are you rotating your alignment (Y axis) to the O.D.?

[[Ch...]]

I measure concentricity across a distance is a huge rabbit hole. I measure stuff like this all of the time and it can get tricky. Pay attention to the bottom about the concentricity

Profile_Concentricity_CALYPSO.pdf

[[Ni...]]

Please sign in to view this quote.

The part is laying down on V-blocks down the Y axis of the machine. So Y origin is the end face, X & Z is the part OD

[[Ow...]]

Yeah, definitely going to want to review the document Chad shared, specifically the section on Concentricity , and review your options to evaluate it differently.
Furthermore, they're most likely checking runout on the lathe, not concentricity.

[[Aa...]]

Chad posted a valuable resource, and Owen is spot on about lathes typically checking runout, not concentricity.

The specification is poorly designed, and qualitatively speaking, the use of a RMB datum and the distance from the datum together mean it will have a very high sensitivity coefficient. Simultaneous position callouts or MMB/MMC, etc. would probably be a much more appropriate application.

[[Br...]]

Yes they are checking runout on a lathe if they are looking at the FIM while the spindle rotates. BUT Circular Runout is a combination of Concentricity and Circularity. Therefore, there should only be more error in the runout check than the Concentricity check as Concentricity takes the form error out of the equation. In other words, If the Runout is good, then so is the Concentricity.

[[An...]]

Here is a proposal.
See attached!

Coaxiality__1.pdf

[[Aa...]]

Please sign in to view this quote.

That would be true if they were truly spinning the part about the RMB/RFS axis of the datum feature. However, in many cases, they have to clamp on something other than the datum and "indicate in" the datum feature. What they are essentially doing is positioning and orienting the workpiece such that the runout is good on both the datum and the controlled feature. And that's not the same thing as verifying the controlled feature to the datum feature at RMB/RFS. It's much more similar to a simultaneous requirement.

[[Br...]]

Please sign in to view this quote.

Ok well that's a very specific anecdote. The argument was that on a lathe runout is being evaluated-not concentricity. My response was that there is an implied concentricity within runout. That's all I was addressing. In your case the problem isn't that concentricity isn't being evaluated, its that it isn't being evaluated to the correct datum feature. There's all kinds of problems that could come from how someone inspects something on a lathe. The part could be chucked too tight for instance so it isn't being checked in a free state. The part springs as its released from the chuck and the operator and inspector get two very different results. You have to take these situations on a case by case basis.

[[Aa...]]

Please sign in to view this quote.

I guess my point was that it is a rare exception that on a lathe, concentricity is being evaluated to the correct datum feature (RFS/RMB). It's not a specific anecdote. It's the nature of measuring on a lathe.

[[Br...]]

It is a specific anecdote. You absolutely could chuck on the datum feature and check the runout. And I don't think that's rare. You could do it a multitude of ways. This is beside the point and a little subjective anyway. If that was really your point in the first place then why would you say

Please sign in to view this quote.

Am I really to assume by that you really meant that they simply don't measure back to the correct datum features? That's quite a leap.

[[Aa...]]

Please sign in to view this quote.

I guess that depends on the accuracy of your chucking equipment relative to the tolerance you're trying to hold and the sensitivity of the specification. I deal in the realm of tight gauge tolerances, where across the board the chucking methods available on lathes and OD grinders are unacceptable as datum feature simulators. In the case of the OP's part, I would say based on the concentricity tolerance, the length/diameter ratio, and the relative distance between the controlled feature and the datum feature that the chucking methods are probably not sufficient, especially considering the loose form requirement on the datum feature. It would take much less than an accumulative .0005" in taper and lobing in the datum feature to make the workpiece "point" when re-chucked so that it doesn't repeat within the .001" tolerance.

[[Br...]]

You still didn't answer my question. What does all this have to with the topic of Runout vs. Concentricity? You said :

Please sign in to view this quote.

This was what my response about Runout and Concentricity was regarding. Where are we going with this? Why is all this information necessary? I'm confused about what you are trying to argue here. If the problem is the lathe setup, and all the problems with the processing equipment, fine. But that's not what was said, and that's not what I was addressing. We're really getting into the weeds here on what was a pretty simple topic.

[[Aa...]]

I'm not here to be argumentative, or to make enemies, although it feels like that's about all I'm accomplishing. Before saying anything else, I want to make clear that I harbor no hard feelings over this discussion, and hope you don't either.

Owen's post, and my agreement with it, came before your response about Runout and Concentricity. So I was neither reiterating what you said nor arguing with it there. I was merely trying to highlight what I perceive to be an under-appreciated point. You apparently agreed, because you went on to further develop it.

The crux of the OP's struggle (which for all I know may be over already) is the difference between concentricity measured by CMM and runout measured by a lathe with an indicator and a chuck for a datum feature simulator that I don't think should be trusted to seat with enough repeatability to give reliable results. I could very well be wrong. There are too many details I don't know about the hardware and the form of the datum feature. The best way to know is to have the operator reload the parts multiple times to make sure the runout repeats. As far as we know, that didn't happen, and the operator merely indicated the controlled feature after the turning operation was complete.

All this is relevant because it could explain why the lathe operator was getting much better runout than the CMM gave for a concentricity result.

[[Br...]]

All good Aaron, If I thought you were just trying to be argumentative for the sake of arguing i wouldn't have continued this discussion. I am just trying understand what you are getting at. I think the problem is semantics. When you say "Runout" you are referring to the complete operation of Lathe inspection whatever that entails. I assumed everyone meant Runout to be the Geometric tolerance called Runout, which is a combination of the geometric tolerance Concentricity and Circularity.

[[Ni...]]

Thank you all for the help. I will comb through all of your info and put it to use. As for not trusting the results on the lathe or stating it's for runout....i'm not convinced. But I see the points of argument.

[[Ba...]]

I think the datum you actually used at CMM and lathe are different, at CMM, the datum you select is actually the axis of cylinder A.
on lathe, i don't know how you fixture fix your part, but i guess the datum the lathe used is not the axis of cylinder A, but one line which perpendicular to the end face and intersect the centerpoint of cylinder A. i.e. what the lathe checked is not concentricity refer to cylinder A, but the runout|Ø.001|M|A . the first datum M is the end face. At lathe, cylinder A provide only one point to the datum, not an axis.
I think the datum used differently is the main cause of your issue. and this part is very long, i think the result read at CMM is move bigger then on lathe.