Can indirect control override direct one? 2

[CheckerHater]

There was very heated exchange recently on this forum caused by ambiguous understanding of some standard terminology. Unfortunately, most participants concentrated on emotional part of the discussion, and overlooked serious underlying problems.
Here I have no intention to start another fight, but rather try to figure out if it is possible to find some common understanding in the realm of “default”, “specified otherwise”, “direct”, “indirect”, etc.

Imagine that we have drawing of a part, that has cylindrical feature. Let’s call that feature “X”.

The drawing title block has a requirement: “UNLESS OTHERWISE SPECIFIED, MAX ROUNDNESS ERROR IS 0.05”

There is also direct 0.10 Runout control applied to feature “X” via FCF.

As we all know, Runout indirectly controls Roundness. Armed with that knowledge, please answer the following question:

What is the maximum allowed Roundness error of feature “X”?
a) 0.05
b) 0.10
c) None of the above (please explain)

Thank you in advance for your opinions.


"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[CheckerHater]

@TheTick:

O, I am pefectly fine "off the hook"

Simple fact that RUNOUT = CYLINDRICITY + SOMETHING_ELSE, and SOMETHING_ELSE is not zero, is mathematical proof that runout measurement of 0.10 always satisfies cylindricity requirement of 0.10

The same way, PARALLELISM measurement of 0.10 guarantees that flatness requirement of 0.10 is satisfied.

 

[CheckerHater]

Sorry greenemi,

It wasn't me who called the question "a trap". Just trying to protect myself

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[pmarc]

Thread title: Can indirect control override direct one?

I think the answer to that one is pretty simple now. Based on the responses you, powerhound and greenimi provided to my question, I think you have to agree that indirect control can override direct one.

Imagine that we have drawing of a part, that has cylindrical feature. Let’s call that feature “X”.

The drawing title block has a requirement: “UNLESS OTHERWISE SPECIFIED, MAX ROUNDNESS ERROR IS 0.05”

There is also direct 0.10 Runout control applied to feature “X” via FCF.

As we all know, Runout indirectly controls Roundness. Armed with that knowledge, please answer the following question:

What is the maximum allowed Roundness error of feature “X”?
a) 0.05
b) 0.10
c) None of the above (please explain)

My answer to this one is - impossible to say because we do not know what the total size tolerance for the diameter of feature "X" is.
- Assuming you added the missing info to the question, and the total size tolerance is less than 0.05, then my answer would be "c) None of the above".
- Assuming the total size tolerance is equal to 0.05, then we actually have nothing to discuss and the answer has to be "a) 0.05".
- Assuming the total size tolerance is greater than 0.05 but less than 0.10, say 0.08, then we are back in the previous "heated" thread where we tried (at least I hope we tried) to solve the dillemma whether size tolerance, as an indirect control of parallelism, is or is not able to override the UOS parallelism requirement given in the title block. In this case there will be a question whether size tolerance, as an indirect control of roundness, is or is not able to override the UOS roundness requirement given in the title block.
- Assuming the total size tolerance is greater than or equal to 0.10, say 0.15, then we are again back in the previous thread, but this time we have to solve the dillemma whether runout tolerance, as an indirect control of roundness, is or is not able to override the UOS roundness requirement given in the title block.

And regarding the 3rd and 4th bullets, if I did not make myself clear enough in the previous thread, I will try now.
I do understand why most people think the correct answer would be "a) 0.05" in both cases. But I also see why someone may say: "I disagree, it is 0.08 in the 3rd and 0.10 in the 4th. If in certain cases direct roundness tolerance from the title block can be overriden by indirect roundness control, then why couldn't it be overriden in all other cases?" If the answer to that is: "It can't, because roundness tolerance from the title block (0.05) is lesser than the indirect tolerance value", then another question is: "Can you prove that my thinking is incorrect?"

This is what I have been missing in this entire debate (both threads) so far, CH. You have not proven that the thinking other than yours is incorrect. Or in other words, you have not proven that your and only your interpretation is the correct one.

If I could summarize this discussion really shortly, in my opinion one of the most important takeaways from it is that the UOS notes must be used wisely. It is not that they should not be used at all. In some cases they can do really great job, but sometimes they may get us into serious unexpected troubles.

 

[3DDave]

Not certain where I saw the prohibition, but the Y14.5 standard gives only one way to specify geometric tolerances and that is via a feature control frame. If there is some other method used, then it is up to the creator of the drawing to supply their own interpretation of what it means, preferably in writing. In this case it seems like the usage is outside the Y14.5 standard.

Were this a real situation I'd suggest the drawing creator not wimp out and they go back and apply the controls to the features the way the standard says to. And then I'd probably select 0.05, unless the size tolerance was smaller or some other control was also applied.

The problem with the UOS version is that there isn't a good way to determine what factors override UOS application - is it when an explicit control is looser or tighter and does that include the size tolerance? Perhaps the note could be expanded to include those conditions, but then it's back to -why?- when a perfectly good standard is sitting nearby with examples.

 

[Belanger]

Wow -- I think many of these comments would have been different if you all used real names rather than hiding behind handles.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[pylfrm]

pmarc,

I'll answer "a) 0.02" for your modified question. However, I would not consider the 0.05 roundness tolerance to be overridden. I'd say all three tolerances apply.

I agree with the main takeaway that UOS should be used carefully. Perhaps alternate wordings should be considered. While not a drop-in replacement, "UNLESS SPECIFICALLY EXEMPTED" might be useful in some cases.


3DDave,

Adoption of symbols indicating dimensional requirements, as shown in Fig. C-2 of Nonmandatory Appendix C, does not preclude the use of equivalent terms or abbreviations where symbology is considered inappropriate.

pylfrm

 

[CheckerHater]

I said it before an I will say it again - nothing is overridden so far. All requirements apply together perfectly well.

Just follow me here:

UOS MAX ROUNDNESS ERROR IS 0.05 = Default requirement
MAX DIMENSIONAL ERROR IS 0.02 = Default requirement fulfilled (or you can call it “satisfied”, or whatever). Roundness cannot exceed 0.02 falls under "Roundness cannot exceed 0.05"
MAX RUNOUT ERROR IS 0.10 = Default requirement is fulfilled (There is NO conflict between default and local requirement. If we agree that runout is compound control, and form and location are two major contributing factors, we simply clarify that remaining 0.05 error may be attributed to locational error)

@Belanger:
Many of these comments would have been different if “experts” like yourself responded with opinions on the problem on hand, rather than personalities of the posters.

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[CheckerHater]

Now continue. I am about to attend the meeting and have no idea when I'll be back.

Let say, MAX DIMENSIONAL ERROR IS 0.15 and Rule 1 is in action.

Rule 1 is compound control in some way similar to runout as I described it before. The total tolerance value is distributed between local size and form errors. How exactly is it distributed? THERE IS NO RULE. Nowhere in the standard it says how to split the value between the two. Pmarc believes that it's my burden to proof that something doesn't exist. Sorry, this is not how it works.

Roundness (cylindricity) is a refinement of Size tolerance - that's a fact. 0.05 is a refinement of 0.10, 0.15, 1.00, whatever.

Now I would like pmarc to show that somethig DOES exist: would you please show us the rule that states, that increase in dimensional tolerance must result in increase in associated roundness?

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[CheckerHater]

Let's continue.

Let's throw runout out of the window.

Case 1:

UOS MAX ROUNDNESS ERROR STILL IS 0.05 = Default requirement
MAX DIMENSIONAL ERROR IS 0.15

According to pmarc's theory (and I am all for carefully considering it) IF size tolerance (indirectly) controls form, THEN local roundness requirement is widened and default roundness is, therefore, overridden. (Am I getting it right?)

My argument is, that in absence of any rule stating how exactly compound tolerance should be distributed between its components, both requirements must be satisfied together, just like they appear on the face of the drawing: Size tolerance 0.15, roundness (cylindricity) tolerance 0.05, roundness is a refinement of size.

Case 2:

UOS MAX ROUNDNESS ERROR STILL IS 0.05 = Default requirement
MAX DIMENSIONAL ERROR IS 0.02

My argument is, that in absence of direct local roundness control, local roundness is limited by the size tolerance: Size tolerance 0.02, roundness (cylindricity) tolerance 0.02, which is still less than 0.05, so both requirements still are satisfied together. It is questionable if this condition may be called "overriding" as the global requirement is technically satisfied.

Case 3:

UOS MAX ROUNDNESS ERROR STILL IS 0.05 = Default requirement
MAX DIMENSIONAL ERROR IS 0.10
LOCAL ROUNDNESS SPECIFIED BY FCF IS 0.02

My argument is, that in presence of direct local roundness control, local roundness is limited by it: Size tolerance 0.10, roundness (cylindricity) tolerance 0.02, NEITHER SIZE TOLERANCE OR UOS RULE HAVE NO CONTROL OVER LOCAL ROUNDNESS ERROR. This is the only condition that truly may be called "overriding"

Hence my opinion is: only direct local requirement may override direct UOS requirement.



"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[Belanger]

CH -- I haven't responded because it's a legitimate problem, with good arguments on both sides, and I'm not sure what the true answer should be.
That's something that you could have acknowledged in the previous QC Inspector thread, instead of saying that you were tired about arguing over something so "obvious."

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[CheckerHater]

@Belanger:

Sorry if I was too harsh.

You don't have a solution, but I am sure you have an opinion. If your real name being exposed prevents you from sharing your opinion with us, is it really a good thing?

So, maybe having "pen names" actually allows us to explore the problem further without looking at "ranks"?

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[Belanger]

Well, you do make a good point -- I try to think about my replies a little more carefully than if I had a screen name, because I have my real name out there.
I suppose my opinion here would be that we shouldn't even create a situation where there is a contradiction between the general notes/title block and the callout on a particular item (kind of like 3DDave's opinion). But I'll think about it some more later today.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[pmarc]

I really don't want to repeat myself again. J-P's and 3DDave's comments sum up my point really nicely.

I would just like to add that per para. 3.5(e) in Y14.5 it is acceptable to place feature control frames in general notes.

 

[CheckerHater]

It was acceptable per 1966 standard to use verbal notes instead of FCFs, and I guess the standards should be backwards-compatible (unless specified otherwise?)

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[pmarc]

One edit in my latest comment:
I meant 2009 version of the Y14.5.
In 1994's para. 3.5 there was nothing about FCF placement in general notes. Additionally, para. 3.3.5 said that it was prohibited to place material condition modifiers in general notes. This statement is gone in 2009.

 

[Belanger]

CH -- I promised to think more about your original question, since you asked about my opinion. Assuming the size tolerance is larger than 0.1, I would have leaned toward the circularity being allowed 0.1 (from the runout). This is because the roundness was "unless otherwise specified," and runout (in a way) specifies roundness.
However, that is my answer only if forced to choose. The best answer is that it's a bad way to call things out, since it's confusing, as we all now presumably agree.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[CheckerHater]

@JP: I think we could also agree that ASME could do better job with definitions of “default”, “specified otherwise”, “direct”, “indirect”, "refine", etc. Or is it all our fault?

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future