Runout Tolerance Applied on an Assembly 1

[greenimi]

Why runout is used in this figure? (instead of profile)

What would be the difference between using profile (maybe even with dynamic profile modifier) instead of currently shown total runout?
Are there any mathematical difference between shown total runout and profile (again with dynamic profile maybe) ?

 

[3DDave]

But at least the AED standard is consistent with the fact the datum axis is part of the FCF whereas Y14.5 lacks, hence creating great confusion.
...
That's exactly what I don't like about the new "addition", being a "different story" !
Different story, but using the same "title" and the same "author"

The AED-1 is example is for a different purpose and doesn't seem to deal with a case that Y14.5-2018 does and has a different author; they are not even in the same series of documents. .

This is why a white paper discussion of the reason for a change or adoption of a requirement into a standard needs to be published ahead of the issuance of the public review draft and kept for public access after release of a standard that includes the changed or new requirement.

What great confusion exists between the two?

 

[jassco]

Figure 12-16 is no good.

 

[greenimi]

The AED-1 is example is for a different purpose and doesn't seem to deal with a case that Y14.5-2018 does and has a different author; they are not even in the same series of documents.

3DDave,
When I said "Different story, but using the same "title" and the same "author"" I did not compare AED-1 with Y14.5.
I compared runout examples from Y14.5 with other runout from Y14.5 (which Burunduk identified it as "There is an exception added, for assembly drawings." ). And those examples are from the same Section (Tolerances of Runout)

So, nice try, but not convincing.....

This is why a white paper discussion of the reason for a change or adoption of a requirement into a standard needs to be published ahead of the issuance of the public review draft and kept for public access after release of a standard that includes the changed or new requirement.

I agree with this.

 

[greenimi]

Figure 12-16 is no good.

Well, but it is part of the standard, therefore MUST be good ! Must be usable.

 

[Burunduk]

So giving up definition robustness and consistency for the sake of simplicity is your (and committee's) chosen option, right?
Is this the way a STANDARD supposed to work?
Instead of education (continuous improvement) I will just surrender........

I don't think adding runout for assemblies gave up "definition robustness and consistency".

Robustness is not compromised because there is only one interpretation to the requirements in 12-16, and the tolerance zones, their behavior and relationship to the specified datum reference frame is crystal clear. From there, you could verify it in many ways, and at least one method is obvious - a dial indicator set up at the right orientation to the datum feature simulators and swept along each of the surfaces when they rotate.

Consistency is something else. You can't expect an exception to be consistent with a general rule, or it wouldn't be an exception. If you give that exception a different title and symbol it isn't an exception either.

 

[greenimi]

I don't think adding runout for assemblies gave up "definition robustness and consistency".

Robustness is not compromised because there is only one interpretation to the requirements in 12-16, and the tolerance zones, their behavior and relationship to the specified datum reference frame is crystal clear. From there, you could verify it in many ways, and at least one method is obvious - a dial indicator set up at the right orientation to the datum feature simulators and swept along each of the surfaces when they rotate.

Consistency is something else. You can't expect an exception to be consistent with a general rule, or it wouldn't be an exception. If you give that exception a different title and symbol it isn't an exception either.

This is not how I read from the runout definition
12.2 RUNOUT
Circular runout and total runout are geometric tolerances used to control applicable characteristics of surfaces of revolution relative to the datum axis RMB.

In fig 12-16 total runout is used, but I don't see the DATUM AXIS.
I see just an axis, not the datum axis since is not part of the FCF.
Surface of revolution I see.
Relative to AN AXIS I see.
But I don't see the DATUM AXIS RMB

So it is in a direct violation of their own definition.
Well, you will argue....ohhh that's the exception, right?
It is not an exception since it is in a direct contradiction with their own definition.

 

[Burunduk]

Greenimi,
I suppose any exception to a rule could be seen as a violation. After all, what makes an exception an exception if it doesn’t contradict the rule it’s carved out from?

 

[mfgenggear]

Burunduk
You are very patient person. Kudos buddy.
You are much appreciated for to trying to help.

 

[greenimi]

Burunduk,

I know you want to defend the committee for their screw up. They created an unnecessary distraction. Nothing prompted them to artificially create this issue. They created a solution which is now looking for a problem.

Just maybe certain individual with a clear agenda to put in the standard what “his company is using for years with (allegedly) successful acceptance and implementation within that particular industry”. We all have those individuals……who want to leave their marks (like the dog on your lawn) before the history will kick them out into the anonymity.

 

[greenimi]

By the way, didn’t you hear about politics within 14.5? Like some “promised” modifiers –dynamic profile or unequally disposed profile definition, --to certain members?
Well, what’s happened in Vegas stays in Vegas.

 

[Burunduk]

greenimi,
I can't possibly defend a screw-up that I don't even consider a real screw-up. To ensure that everything is perfect and prevent even the smallest chance for confusion, the committee should have - after adding assembly runout - rewritten the rest of the runout section to clarify that whenever the axis of revolution is indicated to be mandatory as a datum, it applies only to detail drawings. But realistically, do you really expect something like that to happen?
I don't, I'm not naive.

Honestly, I think there are bigger issues to worry about - like the poor definition of profile of a line, as one example.

 

[greenimi]

To ensure that everything is perfect and prevent even the smallest chance for confusion, the committee should have - after adding assembly runout - rewritten the rest of the runout section to clarify that whenever the axis of revolution is indicated to be mandatory as a datum, it applies only to detail drawings

Again, my opinion is that this "subject" should have been placed in AED-1 (2023) not Y14.5.

AED-1: 2023 has its own chapter called 5.10 Assembly Tolerancing
"Assembly dimensioning and tolerancing pertain to the geometric control of component features that are ONLY applicable on assembly product definition. This section outlines the conditions necessary for such practice to be valid."
The following two primary types of features need to be controlled with tolerance at an assembly level:
- Existing Features with Adjustment at Assembly
- Features Manufactured at Assembly.


They also have " Assembly Level Datum" paragraph.


Just that "Runout Tolerance Application on an Assembly" does not fit well here within Y14.5. feels like have been added in a hurry, just to get it done, without thinking of the unintended consequences of such approach.

 

[3DDave]

The only problem I see with it is that in conventional runout there is an identified axis and the measurements are made in a plane that passes through that axis. If the measurement is not made in that plane the measurement will not be correct, hence my reference to the restoration of the implied datum concept. Instead of providing the inspector the ability to create a fixture and pre-position a dial indicator prior to receiving the part this requires an inspector to hunt around to find the axis of rotation so that the alignment can be made.

The other part that is a bit odd is that most such installations don't offer any adjustment to be made at the assembly level, so Fig. 12-16 seems designed to reject items beyond the ability of the drawing level to do so.

What AED did was to identify an axis, but one that is typically not accessible to be used as a datum feature. They differentiate between machining or other adjustment at the assembly level and rejecting parts beyond the ability of the drawing level to alter, which is good but doesn't fix the original problem.

It mainly reads like a company standard with the results of long discussions between manufacturing and engineering about specific problems and not about more fundamental concerns, like the development of the [IN] symbol. Looks like engineering got tired of inspectors not reading the drawing and making some weird assumption when INDIVIDUAL was used so AED is there to use nails, glue, rivets, and some weld, to make sure the INDIVIDUAL nature of the requirement stays fixed to the Feature Control Symbol.

Sure, a proven communication method seems like a good idea - it should not be part of some independent standard that is outside the series on engineering documentation.

 

[Burunduk]

greenimi, you’ll probably be pleased to hear that, after thinking about this enough, I realized that the tolerancing idea in 12-16 is indeed flawed - at least when considering the intent of the 2018 edition to define runout (both circular and total) based on tolerance zones rather than a dial indicator and a spinning workpiece.

Consider total runout on the cylindrical surface of the wheel in the example. You can think of its tolerance zone as similar to that of cylindricity, except that it has an orientation constraint to the specified datum reference frame taken from the bottom base of that assembly. If the cylinder is offset from the center of the shaft-bore interface (which defines the axis of rotation), then checking it with a dial indicator on the spinning wheel (as in the old definition) might show more runout than the tolerance allows.

However, if you check it without spinning using a CMM or whatever other method that can virtualize the datum reference frame and the theoretical tolerance zone, which is constrained in rotation but free to translate - you can allow the tolerance zone to float in X, Y, and Z to best fit the actual cylinder regardless the functional axis of rotation. This would inevitably create a conflict between the traditional method and the newly supported ones.

So yep, when you're right, you're right!

 

[greenimi]

Consider total runout on the cylindrical surface of the wheel in the example. You can think of its tolerance zone as similar to that of cylindricity, except that it has an orientation constraint to the specified datum reference frame taken from the bottom base of that assembly. If the cylinder is offset from the center of the shaft-bore interface (which defines the axis of rotation), then checking it with a dial indicator on the spinning wheel (as in the old definition) might show more runout than the tolerance allows.

However, if you check it without spinning using a CMM or whatever other method that can virtualize the datum reference frame and the theoretical tolerance zone, which is constrained in rotation but free to translate - you can allow the tolerance zone to float in X, Y, and Z to best fit the actual cylinder regardless the functional axis of rotation. This would inevitably create a conflict between the traditional method and the newly supported ones.

Burunduk,

Being “right” is surely comforting and appealing for most people. The only problem, I do not like to be “right” because I won’t learn anything. When I am “right” I am only reiterating and repeating what I already know without getting more information, without getting smarter and wiser myself. I came here not to contradict you, but to “force” you (or anyone else for that matter) to give me your knowledge (or at least part of your knowledge which I do not have). Ok! Enough with honesty.

Let me think on what you just said above…..very good insight.

 

[greenimi]

All surface elements shall be within a tolerance zone consisting of two coaxial cylinders with a radial separation equal to the tolerance value specified. The tolerance zone is constrained in translation (coaxial) to the datum axis. (2018)

Total runout provides control of all surface elements. The tolerance is applied simultaneously to all circular and profile measuring positions as the part is rotated
360 º about the datum axis. When verifying total runout, the indicator is fixed in orientation normal to and translates along the toleranced surface. (2009)
Burunduk, 3DDave and all,
Despite of this outcome I will continue to argue—mostly with myself
Just wrapping my head around the change is the definition from 2018 to 2009 (both copied here), would you say that if the 2009 definition was kept intact, then fig 12-16 had been impossible to be legally added
OR
would you say that even in 2009 the concept of 12-16 is still possible.
In other words, I am asking if the change in the definition of runout made it possible addition of the exception or the exception could have been added in 2009 too?

I understand that after this lengthy discussion this could be a moot point, but as I said, digging deeper and deeper in the reasoning and the concept depicted in 12-16 is my intent of opening up this thread.

 

[Burunduk]

Actually, I think assembly runout could work as an exception under the 2009 definition of runout. It cannot work even as an exception under the 2018 tolerance zone based definition.

Under the 2009 dial indicator based definition of runout tolerances, the exception would be only that the assembly interface that provides the rotation axis is not a truly a datum feature, in the sense that this feature is not given a datum feature symbol and it is not referenced in the FCF. But effectively, any eccentricity of the toleranced feature relative to it would be detected by the the dial indicator, which would still be part of the definition, even for assemblies (the 2009-based version of 12-16 would show a dial indicator fixed relative to datums A and B in orientation). And it does make sense not to reference the bore/shaft that gives the axis of rotation in the assembly as it is already "fixtured", and functionally you may be interested in the more "global" control as done 12-16, with the orientation of the dial indicator set by the base of the assembly and not the axis of rotation.

With the runout as a control related only to a tolerance zone fixed in orientation to the datums as defined for assemblies in the 2018 version, it falls apart. Because, you are not required to detect any TIR/FIM. So even a very eccentric (relative to the rotation axis) cylinder could pass inspection considering only the tolerance zone of two coaxial cylinders unfixed in location to anything but only oriented to the datum reference frame.

 

[greenimi]

Burunduk,
I will digest your latest reply, but just reading it at the cursory level, your replay straightened my "idea" of politics within Y14.5. I am saying this because, i think (I don't have any proof however) that this is another case of "promised" figure/concept to be added in the next revision for some individuals.
But, in the same time, they decided to also improve the runout definition (which is a good thing IMHO), but now the "promised" figure is not kosher with the verbiage/text.

I am not sure how people from the math standard will comment those changes , allegedly done with the improvement scope and not to change the runout definition as clearly indicated in the 2018 text.
I am not envy math folks who are required to deal with .5 inconsistent and reckless behavior.

 

[Burunduk]

greenimi,
I wouldn't say that 12-16 is not "kosher" with the text, it is just that the control will not work the way it would had the concept been added to the 2009 edition and involved rotation of the part and a dial indicator reading, per definition.

If your suspected scenario is true (and I agree it can be), the individuals who were waiting to have the concept added to the standard got their figure and sub-paragraph as promised, but in combination with the now purely tolerance zone based definition, the resulting control is impaired in a way that probably no one expected. The said individuals probably still wanted to reject features based on their eccentricity relative to the rotation axis and not just based on form and orientation deviations relative to the specified datum features.

See the image below. On top of figure 12-16, I drew in red a possible condition of eccentricity deviation of the cylinder relative to the rotation axis. Note how the orientation is accurate, but the feature is off-center. The cylindrical feature would be rejected by recording the total movement distance of the dial indicator tip during rotation, however it would be approved by the floating coaxial-cylinders tolerance zone shown in the "Means This" portion and as described in the text (the tolerance zone is not required to be coaxial with the rotation axis).

 

[greenimi]

Burunduk and all,

I think I understand your statements and I agree with them. And that’s the sad thing.
If two men agree on everything, you may be sure that one of them is doing the thinking. (Lyndon B. Johnson)
Is this case it is YOU. I am learning from you.

I guess we need someone else to offer us different / opposite perspective.


Now, just a very quick follow up inquiry about the same subject (of runout):
Since now ( per 2018 ASME standard) the tolerance zone has been changed from a dial indicator (part rotated 360 degrees about the datum axis) to a “two boundaries” tolerance zone (two coaxial cylinders with a specified radial separation, for example) do you think that is opening up the usage of the runout (circular or total) on the threads (outside threads, male threads, for example)?

I am thinking that runout COULD be used for the outside diameter of gears and also can be used for the splines (interrupted, non-continuous surfaces), but what about extended to threads? Do you think runout could be legally used?