Tolerance Zone Extent 2

[axym]

The recent thread on Surface Profile and Size brought up an interesting effect regarding tolerance zone extent.

Here is a drawing for a fictitious part, that illustrates the effect. I have intentionally avoided the use of directly toleranced dimensions.

It has been proposed that an as-produced part like the blue one would conform to the 2 mm profile tolerance to A. This doesn't seem right to me, but so far I can't point to anything in Y14.5 that would disallow this. The standard does not give much detail on the extend of tolerance zones (or datum feature simulators).

Opinions?

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[CheckerHater]

The tolerance zone extension issue still exists, even when the basic part geometry is fully defined

The tolerance zone extension issue doesn't exist, plain and simple.

By the definition:

Profile tolerance zone DOESN'T EXIST outside of nominal surface.

You either define your nominal surface with enough dimensions or not. (That's including CAD data)

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

 

[Burunduk]

CH, if the profile tolerance zone does not exist outside the nominal surface there should be a way to set its dimensions. How do you suggest doing that?
Suppose that your nominal surface is the top surface of a part 100+/-2 wide and 70+/-2 long, plus there are form and orientation variations possible depending on how the peripheral surfaces are controlled. What are the dimensions of the "nominal surface"?

 

[CheckerHater]

Add 2 dimensions... or 2 other dimensions...

And your nominal surface is defined:

Also Burunduk, could you sketch your example with "undefinable" surface please. I am having hard time visualizing it.

Thank you

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

 

[Burunduk]

CH,
I can re-form my question to address the issue using your last graphic.
The "nominal surface" is the R147.28 top arc. Would you say that the extent of the tolerance zone (in the direction of the width) for a profile control applied on it is 51.70 exactly?

 

[CheckerHater]

Not me, but the Standard says so:

"Along", not "outside"

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

 

[Burunduk]

CH, according to your interpretation of the Y14.5.1M definition which you posted earlier - "Profile tolerance zone DOESN'T EXIST outside of nominal surface." I assume that there is an agreement that "nominal surface" as referred in that definition is what is defined in Y14.5 as "true profile".

According to the Y14.5-2009 quote you provided now, the tolerance zone is "extending along the length and width (or circumference) of the considered feature or features.", and as you noted - "not outside".

Can I conclude from these two posts that your interpretation is that "considered feature"="true profile"="nominal surface", and that the width-extent of the tolerance zone in your last figure is precisely 51.70 as the basic dimension defines?

I am not saying that you are wrong or right, just trying to understand your view on this topic.

 

[CheckerHater]

The width of nominal surface (not tolerance zone) is exactly 51.70, unless something else is specified on the drawing.

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

 

[Burunduk]

The width of nominal surface (not tolerance zone)

Could you clarify regarding the part which you put in brackets - "not tolerance zone" what case you envision for the extent of the tolerance zone to differ from the extent of the nominal surface and not be 51.70?

 

[chez311]

CH,

The tolerance zone must be extended some amount even in simple cases. From Y14.5-2009 para 8.3.1:

"Where a profile tolerance encompasses a sharp corner, the tolerance zone extends to the intersection of the boundary lines. See Fig. 8-12. Since the intersecting surfaces may lie anywhere within the converging zone, the actual part contour could be rounded."

For a sharp corner (non-tangent) if we limit the extent of the tolerance zone to only those points which are normal to the "nominal geometry" which based on your response (1 Jul 20 15:48) you take to mean the specified basic geometry then there are points which will not have a tolerance zone boundary - it would be undefined what would happen in these corners.

 

[CheckerHater]

Also: I don't see simultaneous requirement in the OP drawing.

Tolerance zones are completely independent. There is no sharp corners between the tolerance zones. There is no need to blend them.

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

 

[axym]

The standard states that the tolerance zone "extends along the length and with (or circumference) of the considered feature or features" and also that "profile tolerances apply normal (perpendicular) to the true profile at all points along the profile". This would seem to correspond to the extent of the as-modeled surface in my example, with the 51.70 basic width.

If this were strictly true, however, then the zone would not cover the entire feature if it was produced wider than the 51.70 basic width. The profile figures typically show the tolerance zone as extended past the as-modeled surface by an unspecified amount. The zone is also shown as extended even when the profile is specified to apply between points that are defined by basic dimensions. So there is the implication that the zone must be able to extend past the as-modeled surface, but no details are given.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[3DDave]

This is exactly what removing sharp edges takes care of. No one seems to complain about that.

 

[CheckerHater]

axym, Burunduk, and chez311,

Do you insist that according to Y14.5-2009 para 8.3.1 tolerance zone can be extended this much:

Also, on the matter of " it would be undefined what would happen in these corners" (chez):

Standard provides means to deal with corners that have to be explicitly blended (or not).
Interestingly, you still have to provide dimensions:

So it is still the designers responsibility to provide explanation "what would happen". Lack of this explanation is not a Standards fault.

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

 

[chez311]

CH,

I didn't suggest to blend the tolerance zones. The verbiage of the standard specifically states "the tolerance zone extends to the intersection of the boundary lines". I would say this applies to the intersection of two profile tolerance zones, even if not grouped as a pattern.

In regards to your figure, yes you have found a case where what happens is undefined even if we extend the tolerance zones. This requires that the peak of that point be either at a point of common tangency of the two arc, or a similar result could be had if the peak is close enough to this point of tangency and the width of the tolerance zone wide enough that they don't intersect. I don't see how your solution provides any more clarity in this case - either way it is undefined what happens near the peak.

Do you insist that according to Y14.5-2009 para 8.3.1 tolerance zone can be extended this much:

No, as you'll see from my previous reply (30 Jun 20 20:15) I would say that clearly violates design intent, I don't think any rational person would accept a part like the one on the right.

So it is still the designers responsibility to provide explanation "what would happen". Lack of this explanation is not a Standards fault.

I wholeheartedly agree, I agreed with you on this several times. The issue at hand arises from insufficient constraint. It would be up to the designer to ensure that such strange behavior (from which can arise multiple interpretations) is not allowed by providing sufficient constraint and properly specifying the applied tolerances.

 

[CheckerHater]

chez,

I do not provide any particular solution.

One could be ISO definition with the "ball rolling"

Real world sharp edge could require "dull sharp edges" note or, again, ISO symbology for corners.

It all winds down again to creator of the drawing - if you think something could be misinterpreted - explain it, don't blame the standard.

I think we could agree on that.

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

 

[chez311]

CH,

Sorry, I should have said your interpretation instead of solution. There are likely more than a few potential solutions.

Yes we agree on that point and others.

 

[Burunduk]

CH, thank you for clarifying your interpretation on 1 Jul 20 16:33.

Here is a statement by you that I completely agree with:

Tolerance zones are completely independent. There is no sharp corners between the tolerance zones. There is no need to blend them.

Refer to the following sketch, a modification of the part in the OP. It is also incomplete and I am aware that there are some specifications missing such as the perpendicularity of datum feature B relative to A. I believe that there are enough definitions to properly address the topic. I must also clarify that I think that datum feature B is missing as a secondary datum reference for the arc (or a simultaneous requirement of all features to A instead), but based on your statements and the figure you posted in the other thread I think you will disagree.

If I understand your statements so far correctly, the tolerance zone for the top arc is limited by the true profile in the shown way: it spans only at normal offset along the length of the arc, which you find to be limited by the basic width dimension between the two vertical faces.
Based on your comments here and the figure you posted at the related thread, you also shouldn't see a problem with the A being the only datum reference, as long as all the required basic dimensions are specified.

My question is, suppose that a part utilizes the entire "n" tolerance on the vertical sides and comes out at width W+n. And suppose that the radius is very accurately coincident with the true profile other than spanning somewhat longer than the true profile to accommodate the W+n width. What should be decided regarding the portions of the arc that are marked up in yellow? They are outside of the "m" tolerance zone that controls the arc. Should the arc be found non-conforming to the requirement because it doesn't end at the theoretical corner dimensioned by the basic Q, and spanning wider and lower than it's tolerance zone?
Thank you.

axym, Burunduk, and chez311,

Do you insist that according to Y14.5-2009 para 8.3.1 tolerance zone can be extended this much:

My opinion is: according to the letter of the standard, it can. Will it be approved under any realistic circumstances at any factory? Very very unlikely. Perhaps in some futuristic fully automated inspection process where parts can end the process without human judgment involved.

 

[Burunduk]

CH,
I would also appreciate it if you address the following case.
This one has the arc fully constrained and doesn't have the floating tolerance zone issue as in your recommended schemes.
Still, the same problem that results from your fixed-extent tolerance zone interpretation occurs here as well.

Should the part as produced be rejected for the arc falling outside of the tolerance zone?
Or maybe the part should be approved because the portion at the extent of the tolerance zone falls within it, but the portions that are beyond the extent allowed to be completely uncontrolled?

 

[axym]

Hi All,

Great discussion!

3DDave,

I have also thought that some of these anomalies are caused by the part being modeled with sharp corners, which is fundamentally unrealistic. If the part is modeled with a tiny radius at each corner, this changes things a lot.

CH,

I'm not insisting that the tolerance zone can be extended to the degree shown in my sketch. I don't think that it should be allowed to extend that far, but I'm trying to come up with a reason why not.

You brought up the issue of tolerance zones being independent. I generally support the idea that tolerances are independent of each other, and that we should be able to assess conformance to a tolerance independently of the other tolerances. Unfortunately, I think that if I follow this principle with tolerance zone extension it will lead to a conclusion I don't like ;^). If tolerances are independent, then we should be able to assess conformance to the profile tolerance on the upper surface in my figure, without knowing anything about the side surfaces. Or in Burunduk's figure, we should be able to evaluate the profile tolerance on the curved surface independently of the profile tolerance on the upper planar surface.

Does this mean that when assessing conformance to a tolerance, we pretend that we have no idea what the extent of the feature is? So that we must extend the tolerance zone boundaries to whatever extent the feature might have?

Burunduk,

Your figures capture the issue quite well. If the arc tolerance zone does not extend, then the red segment would be nonconforming (which seems wrong). What would happen if the red section extended even further? It would technically be within the extended arc-shaped zone, but outside of the planar zone. Would we say that the part conforms to the arc-shaped zone but not the planar zone? I believe that we would.

We have to deal with extent for other tolerance zone types as well. For a position tolerance on a cylindrical hole, what is the extent of the position zone, when the actual thickness of the part could be greater than the as-modeled thickness? I believe there is some reference to the thickness of the part - we don't cut off the zone at its nominal extent.

chez311,

You could be right - we may just have to accept that "strange" things can happen if we specify tolerances that have insufficient constraints or apply in different datum reference frames.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[Burunduk]

What would happen if the red section extended even further? It would technically be within the extended arc-shaped zone, but outside of the planar zone. Would we say that the part conforms to the arc-shaped zone but not the planar zone? I believe that we would.

I believe so too. I think this supports the idea that tolerances are indeed independent of each other (apart from simultaneous requirements and patterns), and to allow them to be independent, the zones must be allowed to extend as much as necessary to allow the evaluation of the entire considered feature regardless of its as-produced span. This is also what the Y14.5 definition specifies (in the paragraph that was quoted in this thread while being misinterpreted to mean the opposite thing from what it implies).

I'm not insisting that the tolerance zone can be extended to the degree shown in my sketch. I don't think that it should be allowed to extend that far , but I'm trying to come up with a reason why not.

I wish to take this statement to mean that the situation should be prevented by the designer, who will not allow this by taking appropriate measures to prevent such uncontrolled variation of part features.

I wouldn't like to see the standard imposing any artificial limitations on this matter, which will only introduce new issues to be confused about. Instead, the designer should not be lazy and define the appropriate constraints for the part features that his drawing controls. Not thinking that it is necessary and skipping essential specifications as in this figure, that may lead to results such as these may be either a result of not thinking things through or apparently of misinterpretations such as that the span of a surface is somehow limited by the length of the true profile of that surface.

I also don't think that removal of sharp edges has anything to do with all that.
There are several very good reasons to remove sharp corners from parts (on the production floor, not on the model - unless it is a radius or a chamfer intended to be larger than what the general deburring achieves). Resolving tolerance zone-related ambiguities is not one of these reasons. Especially since no such ambiguities actually exist unless one creates them by misinterpretation of the standards.