Shortcomings of ASME Y14.5-2009?

I have been thinking it for a while and wanted to start a similar thread. Here are some of my thoughts:
Is a FOS a feature? Not defined. A cylinder FOS a feature, but width type of FOS requires two opposed flat surface features to be a FOS. Logically not consistent.
Is a point or a line a feature? Not defined. What about an apex or an edge?
Is cone a FOS? Not defined.
Is a circle a FOS? For example, root circle of cone.
A round hole is a FOS. A triangle hole, square hole, rectangle hole, pentagon hole, hexagon hole, heptagon hole, octagon hole, ..., obviously is not FOS. But a round circle can be viewed constructed by infinite number of lines. A centroid can be derived from all these shapes. Similar to oval hole or even any shape of hole. In cases you are looking at the centroid of these shape, can they all be FOS?
...
This is all for today.

Fig. 8-18... applying a profile tolerance to a toleranced shape; i.e., a "non-true" profile, which goes against the second sentence of paragraph 8.2.
But that's been discussed a little in the past -- where's Evan when I need him?

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

See this summary from Bill Tandler... I like his comments on what was done right and could be improved.
Paul

I know Bill Tandler's article. I agree with some comments. Some are based on his personal preferences (like the one with (S) modifier).
I have some sort of personal list, but I would prefer to wait with posting it, until more votes are given.
Hopefully there will be more votes

pmarc,

Shortcomings of Y14.5-2009? How dare you even suggest it! ;^)

I have an extensive list as well, and I've been waiting to post it until others have responded. What do you think - should I just go ahead?

J-P,

I'll back you up on Figure 8-18. I have major issues with that one (and Figure 8-17 as well).

Evan Janeshewski

Axymetrix Quality Engineering Inc.

Evan,
I was just going to post my list, but since you offered to post yours, please go ahead. I can't wait to see what I overlooked in mine

To keep this thread alive and also to add my 2 cents I would say the issues I know of are:
- Cannot find a true definition of a pattern
- Datum feature simulator on a pattern: no indicator whatsoever on how supposed to work/act—expand/contract in the same time or one feature at a time until one achieve this maximum and then the others, etc.
- Fig 8.27 (profile of a line in combination with direct tolerance dimension)—I have read that some GD and T experts, here on the forum, seem uncomfortable with the picture itself.
- Missing (intentionally or unintentionally) recommendation /rule for the thread callouts (MMC versus RFS). GD and T community is divided in two camps (one to recommend having the thread at the MMC and the other the thread at RFS). Yes, I know it’s case by case, application based, by the majority of them should fall into one of those 2 categories.
- Composite callout applied on a single feature (is it allowed, but not recommended or not allowed…)

greenimi,
Thanks for comments. Here is my list (not complete though):

- Feature of Size definition [para. 1.3.32] - still unclear and causing countless arguments even amongst GD&T experts.

- Actual Local Size definition [para. 1.3.54] - unclear as for "any cross section" term. What is the orientation of the cross-section? I assume it is perpendicular/normal to something? But what is it? The math standard seems to be of no help here.

- Derived Median Line / Plane definitions [paras. 1.3.30 & 1.3.31] - what if considered feature is nominally curved, like a garden hose for example. How would an unrelated actual mating envelope look then?

- Tolerance Accumulation example [para. 2.6] - directly toleranced dimensions used to control a non-FOS distances between surfaces. This seems to contradict the statement in the foreword of the standard: "This revision contains paragraphs that give a stronger admonition than in the past that the fully defined drawing should be dimensioned using GD&T with limit dimensioning reserved primarily for the size dimensions for features of size". I realize that the intent of this figure is to show purely the tolerance accumulation concept, but I am pretty sure that quite a lot of readers, when looking at fig. 2-4, may think this is reasonable and perfectly fine way of dimensioning the part. Especially that the note at the end of the paragraph about using basic dimensions together with geometric tolerances rather suggests that the choice between the two methods exists, and not that the latter is the correct one.

- Unclear procedure of how datum feature simulators behave when pattern of features is selected as a datum feature and is called out at RFS. If one simulator fully touches its feature, what happens with the other (if we just consider pattern of two features)? Does it stop or expand/contract further until maximum possible contact with the second feature is obtained?

- Figures for straightness of derived median line at MMC [fig. 5-3] and flatness of derived median plane at MMC [fig. 5-9] showing "axis" interpretation of the callouts. For an unaware reader the figures may imply that the tolerance zones are of uniform diameter (for straightness) or width (for flatness) regardless of size of toleranced feature along its axis / centerplane. In reality (and in my opinion in majority of cases) size of the tolerance zones may change depending on what happens in each individual cross-section. Boundary interpretation is insensitive to actual local geometry variation, plus it is much more intuitive, and reflects functional reason of applying such callouts much better.

- Simultaneous Requirements: RMB [para. 7.5.4.1] - does this paragraph imply that there is no possibility of defining Separate Requirements when two or more patterns of features are located relative to common datum features SUBJECT to size variation? After reading 7.5.4.2, where the choice between SIM and SEP REQT is clearly worded, I would interpret that such option in 7.5.4.1 does not exist, and SIM REQT is the only possible condition. But what if I took fig. 4-40 and imagined two keyways in each cylinder instead of one in each? Wouldn't I be allowed to apply SEP REQT then?

- Chapter on runout in general, but especially lack of clear information about validity (or not) of total runout callout applied to other than cylindrical surfaces of revolution (like cones, sine waves, etc.).

pmarc,

I'm looking at your 4th bullet. I remember I saw somewhere using the coordinate dimension and tolerance is appropriate on the stepped surfaces, examples could be a part like fig. 2.4, or bore holes. It's quite straightforward. Otherwise you have to use Profile tolerance on each surface and designate one end surface as local datum feature. Overkill.

I should probably clarify one thing. I somehow missed last sentence of the suplementary note in paragraph 2.6 saying that: "Locating features using directly toleranced dimensions is not recommended." It is good that the caution is there, but that still does not change my opinion that the concept of tolerance accumulation should not be explained using not recommended dimensioning method(s).

bxbzq,
My main objection to use coordinate dimensioning in such cases is that very often the dimensions are specified for non-feature-of-size type of features (and this is what is happening in case of some dimensions in Figs. 2-4a and 2-4c). Additionally lack of datum feature references causes that each dimension is measured as if it was not tied in space to other functional features - and in reality this should not be a case in this kind of products. Moreover, sometimes something opposite may happen - the directly toleranced dimensional requirements are verified as if the dimensions were basic and the profile callouts were applied (this can easily happen if an inspector places the shaft from fig. 2-4b vertically on a surface plate and uses a height gage to check each step of the shaft.)

As for overkill topic, I would say that in majority of cases GD&T application requires much more drafting work and is less straightforward in comparison to old-fashion coordinate dimensioning. It looks like overkill sometimes - that is true, but if this is the only way to clearly express design intent, why to avoid it? I am in the group of folks firmly believing it is much better (and less painful) to spend some extra hours on the product definition stage to clearly define what is really needed, than to fight neverending battles with fabricators making parts satisfying ambiguous or non-functional drawing requirements - the battle that often can't be won.

pmarc,

You really want to use Profile tolerance to control those flat surfaces perpendicular to hole axis in fig-1.37, 1.38, 1.39, 1.41, 3-15, 3-29, 7-24, 7-25, 7-26?

If there was a reason to do this, why not?
Take fig. 7-24 as an example and imagine that as-produced top surface of the part has a maximum possible form error in a way that the top surface in the place of left bore (as shown in cross sectional view) is lower than a portion of the same surface in the place of right bore by the amount of actual form error. What happens then? Per the figure, the depths of the bores must "follow" the actual curvature of the top surface, but since the surface contains ups and downs, the depths will also be distributed up and down. So if a requirement is to keep flat surfaces of the bores as coplanar as possible, profile may be a better option, because you will be verifying the depths from the same fixed datum (reference) plane. Do you see my point?

There is also the option of the dimension origin symbol, which is the non-GD&T cousin of a datum.

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

Pmarc,

There may be applications where coplanar or surface to surface relationship is important, but for typical bore holes that hold bolts, screws or multiple diameter shaft, would you still use Profile tolerance to control the surfaces?

Saying shortly:
- for stepped shafts - I would.
- for what you call "typical" bores - depth symbol or directly toleranced depth dimension have my green light as long as shortcoming described in my previous post does not conflict with real design requirements.

"Shortcomings of Y14.5-2009? How dare you even suggest it! ;^)
I have an extensive list as well, and I've been waiting to post it until others have responded. What do you think - should I just go ahead?"

Evan,
Would you mind to post your list too? I am just curious, what are your thoughts.

pmrac,
I think you really are hitting on the crux of the problem, industry wants to continue to take shortcuts, but, the whole reason these newer alternatives were developed was because the old methods are not sufficient. ASME (the committee) is also in the difficult position of having to be a salesman for the standard, there is no real enforecment so they have to sell it, like any good salesman then they have to sugar coat everything for public consumption.
Frank