There are irregular features of size that cannot be grabbed hold of with calipers. The three pin examples shown in Y14.5-2009 Figure 4-35 are good examples of this.
nry67 - Regarding your "many letter dimensions" drawings I prefer to say that B, G, H, K, & U all "may" be features of size... The standard only defines fully opposed planar surfaces as a feature of size, but of course if you lose 1% of that overlap, so the surfaces are not quite fully opposed any more, most designers will still consider, and use, the feature as a feature of size. If the overlap is reduced further, then at some point the designer may decide that the feature is no longer to be considered a feature of size. They could show this by choosing a profile tolerance for each surface, rather than a size tolerance for the width of the feature. This is a part of GD&T that is not clear cut, I believe, so it is up to the designer to decide how much overlap for partially opposed planar surfaces is "enough" to be a FoS.
The situation is similar for a partial cylinder. I would say that your "AA" dimension on the partial cylinder looks pretty good as a candidate for being a feature of size, but since it's not a full cylinder only the designer can make this call. Does it function as a feature of size by constraining a mating feature that is also a partial cylinder, or is it different, such as maybe acting as a contact surface as a feature on a rocker arm? If it's the latter, then a profile tolerance makes sense. If it's the former then size and position tolerances may be justified.
Whether that partial cylinder is an irregular FOS or not is a good question... Read the definitions again and you may say "it could go either way". The conservative approach would be to declare AA to be an irregular feature of size, for which rule #1 does not apply. That would then lead to a form control being applied separate from a size tolerance. I wonder what others think of AA in nry67's figure..? Irregular or regular? If it has insufficient arc subtended, then maybe is shouldn't be considered a FoS of any kind, but since I don't know how it functions, I don't know if I can make that call.
Good points, Dean. I guess I was thinking of the examples given in his sketch when I mentioned the calipers as an analogy.
However, I think you're being too skittish when questioning whether B, G, H, K, U are truly FOS. The requirement in para. 1.3.32 is that there be "a set of two opposed parallel elements." I think B, G, H, K, U meet that minimum requirement. There is no stipulation that the opposed parallel elements be the same length. (If that were true, then we could never use the GD&T symbol for parallelism because no physical surface could ever be "parallel" to a datum plane which extends to infinity.)
AA and V would fall into the category of irregular FOS, now that I'm thinking more about it. This is because they aren't full cylinders (which seems to be the mandate for regular FOS), but we could say that they may "contain or be contained by an actual mating envelope that is a...cylinder."
John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems
John-Paul,
Whether a feature qualifies as a feature of size or not doesn't really have anything to do with a parallelism control in any way that I can think of. With parallelism, a tolerance zone that is parallel to a plane or axis of a datum reference frame imposes a tolerance zone upon a surface, axis, or center plane. With a "width" type of size tolerance the two surfaces acting equally together must constrain a mating envelope.
Please see the attached file for a set of three "is", "is not",and "may be" figures.
For the middle case #2 in the attachment, with partial overlap, as with B, G, H, K, & U, it's a judgment call whether each is a feature of size. Those shown in nry67's figure would probably be OK as features of size, but unless there's nearly full opposition of the two parallel planar surfaces then only a discussion about function will lead to a proper determination of whether the feature should be considered a feature of size or not.
For the similar question of a partial cylinder, quite a bit more than 180 degrees should be subtended to consider the partial cylinder to be a feature of size. Otherwise the axis found may not be so repeatable in the "barely constrained" direction.
Hello again - Boy I wish we could edit our posts here... That second sentence should have read... "With parallelism, a tolerance zone that is parallel to a plane or axis of a datum reference frame is imposed upon a considered feature's surface, axis, or center plane."
Sorry for my premature clicking upon "submit" for that prior post.
Here's how I look at it, based simply on the definitions given in paragraph 1.3.32:
A regular feature of size MUST have opposing elements, so case 1 and 2 in your sketch meet that.
A semicircle can be "contained by an actual mating envelope that is a sphere, cylinder, or pair of parallel planes." So it is a FOS of the irregular type.
I mentioned the geometric characteristic of parallelism because I was highlighting that I wouldn't take the word "opposed parallel elements" as given in 1.3.32.1 and think that to mean "parallel over equal distances." In fact, all 3 of your sketches have parallel surfaces but #3 fails on the "opposed" issue.
Admittedly, I am taking the purely legalistic approach based on Y14.5! I agree that there are considerations such as repeatability that might put a chink in the wisdom of calling a semicircle a FOS. (Isn't it fun to debate the little stuff?)
John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems
I think I'll just use 'profile' and put you lawyers out of business! (Okay... Just shoot me. My brain is already fried trying to understand variable tolerances with their stress distributions, strength distributions, and what not.)
John Paul & SeasonLee,
I think you're treating this subject as one that can be classified as black and white, when good practice dictates that it must be gray.
A feature of size can be two opposed elements, but if a couple of planar surfaces have some overlap that does not necessarily make them a feature of size... If those parallel planar surfaces only overlap by 1% of their footprint, then can they properly constrain a mating envelope, just as a similar feature with 100% overlap can? Probably not, is the only supportable answer, I believe, so a size tolerance should not be applied to that pair of parallel planar surfaces. Where is line drawn then, regarding what is and in not a feature of size... 3% overlap, 7%, 15%, 55%, 80%? Since we can't say where the line is definitively, then that leaves the design engineer to decide whether a size tolerance (a directly toleranced dimension applied to the distance between the partially opposed parallel surfaces (which I know those most involved in this discussion are fully aware of, but which I add with the hope of adding clarity to anyone new here)) is reasonable to apply.
To say that a size tolerance can be applied when there is .01% overlap, so therefore "opposed elements", when in fact the surfaces are planar and mostly don't overlap, would be a poor practice. The mating envelope for that feature is two parallel planes that are fit outside or inside the entire feature, not just to the opposed elements... If the feature is slightly imperfect, with the features not parallel enough, and the .01% overlap doesn't do its job of orienting the mating envelope, then the mating envelope will rotate from the orientation of the feature, just as it would for a case with parallel planar surfaces with no opposition.
For partial cylinders, partial spheres, and partially opposed parallel planar surfaces, the designer has a "size or not size" judgment call to make. With fully opposed parallel planar surfaces, full cylinders, full spheres, and features comprised of only opposed elements with a directly toleranced dimension applied (such as the elements on each end of an elongated hole) the there is no question that the feature is a feature of size.
There is no point in having a black and white answer to this question if that answer yields a non-functional result for some cases.
Hi again Dean -- I would say that yes, even if there is only .0001% overlap of the footprint, it is considered "opposed."
However, I think I have a solution to our dilemma that I had overlooked before. One of the other requirements for a FOS is that it be "associated with a directly toleranced dimension."
I agree with you that for practical reasons it's not a good idea for a designer to think of a distance with only a small overlap (such as your example #3) as a feature of size. So here's the solution: The designer shouldn't give a directly toleranced dimension across there. Problem solved: it's not a FOS. But if he does put a directly toleranced dimension across there, then it instantly becomes a FOS since all conditions have been met.
I think this forces the designer to take those "may or may not" situations and make them "black and white" simply by the method of dimensioning.
One last thought: because of this thing about "directly toleranced dimension," there are many items that folks might instinctively call a FOS that are not really such. Using examples from the 2009 standard, we can say that these items are not a FOS:
-the height of the part in Fig. 4-7
-the width of the part in Fig. 4-9
-the diameter of the pin in Fig. 4-45
-the part thickness in Fig. 7-38
(On a real print these would probably be dimensioned, but as they are now, those items cannot be called a FOS, either regular or irregular.)
John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems
John-Paul,
Letting the design engineer make the call whether a partially opposed pair of parallel planar surfaces is a feature of size or not, with evidence of the outcome of that decision shown by either applying a directly toleranced dimension, or not, is exactly what I've been saying all along. By saying the designer must decide, what I meant throughout this discussion is the designer must either choose profile or a size tolerance (how else would they indicate the outcome of their decision).
More on the semantics side of this... parallel planar surfaces having opposed elements, does not make the planar surfaces "opposed". To say it that way paints a misleading picture of their relationship. They have opposed elements, or are "partially opposed", is a much better description in my opinion.
Sorry if I missed what you were saying all along. My hesitation was in your claim that there is a gray area, where I maintain that it really is black or white. (When looking at a print we can always say with certainty if something is or is not a FOS.) I think you were a few steps earlier in the process, where someone creating a new design has the power to decree something as a FOS, and that is true.
Regarding the "opposed" stuff, I agree ... I wasn't specific enough in my preceding post and simply said "opposed." Paragraph 1.3.32.1 mentions "a set of two opposed parallel elements or opposed parallel surfaces." But a footprint overlap of .001% does have two opposed parallel elements, right?
John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems
John-Paul,
Yes, I agree that a pair of partially opposed parallel planar surfaces have opposed elements even with .001% overlap. If I size tolerance is clearly applied to only the overlapping portion, but not to the rest of those planar surfaces, then I think only those opposed elements and the directly toleranced dimension between them are a feature of size.
And as I think we now agree, if the designer decides that the overlap for a partially opposed pair of parallel planar surfaces is sufficient, and the if they decide that the feature functions as a feature of size, then a size tolerance can be added and the entire footprint of both planar surfaces + that directly toleranced dimension are a feature of size.
Yes, well said... We probably went off track from the OP's question, but it helped me understand a FOS better by wrangling through the fine points of the defintion.
John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems
J-P,
I am quite supprised by your exclusion of those parts named above. I do not see the logic in any of them, sorry, you do mean 2009 right? This FOS exclusion has always been an odd issue with me, would you explain?
Frank
Frank ... while those figures I mentioned do have opposing elements, they do not show "directly toleranced dimensions" (such as height or diameter), which is one of the other required conditions of a FOS per the standard's definition.
John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems
OK, I can see that, you are taking them strickly as they are given. I assume those dimensions are given as they would be on a completed part drawing.
Thanks, for the clarification.
Frank
