Minor errors in Y14.5?

[Belanger]

Please don't ask why I'm looking through the standard when it's nearly midnight, but... I'm hoping the other GD&T regulars on here can confirm a couple of minor things I've found in the 2009 standard. I'm not trying to nit-pick them (a job well done, I must say!), but here are three things that don't seem right:

1-- Figure 7-45 shows datum feature A with the new "CF" symbol. Is that usage OK? I guess the concept makes sense, but what bugs me is that "CF" is defined in paragraph 2.7.5 as only for features of size.

2-- Check out Figure 4-32. Is it kosher to show the 20 mm basic dimension and just assume that it is centered around datum A? I'm sure this question was in another thread many moons ago. I think most of us said that you should at least have one of the holes dimensioned as 10 mm from datum axis A.

3-- This last one's pretty tedious: Figures 5-2 and 5-3 describe straightness on a FOS this way: "the derived median line of the feature's actual local size must lie within..." The definition of derived median line in paragraph 1.3.31 says that it is formed by segments which are normal to the unrelated actual mating envelope. My question: Is there an internal conflict in those statements? The last sentence of 1.3.31 seems to throw a wrench into the idea of having straightness find the center points for each local cross-section. Heck, if we're checking the straightness based off of the AME, it will always be straight!

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

http://www.gdtseminars.com

 

[MechNorth]

Yes. I see the "actual local size" as a discrepancy, or inconsistency as well. Will you submit it to ASME for clarification in the next edition, which they are working on now? The latest graphic makes the point effectively, I think.

Glad we were able to sort it out. Do you have a resolution for your client now?

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[DeanD3W]

The improvement to local size that I would like to see is to align its definition with the sphere swept definitions in Y14.5.1, usually referred to as the "LMC sphere" concept. This would take care of John-Paul's concern, I believe... The orientation of the local sizes, for an RFS or MMC case, would be defined by the sphere that could just be fit to the feature at a given location, where that sphere's perimeter fits just within the material (so at a given cross-section that is oriented very close to that which a micrometer would find, the local size would be a maximum inscribed circle for a cylindrical pin, or a minimum circumscribed circle for a cylindrical bore). The local size could be approximated with a micrometer or caliper measurement, just as local sizes are often currently measured, but a more rigorous measurement and some analysis software could do a much better job than the current definitions lead us to.

I'm not aware of another approach that would provide the proper orientation of cross-sections, and the LMC sphere concept is already in Y14.5.1. Two point measurements would still model the proper local size for slots or slabs (widths), since they model the size of the sphere that is fit to the feature, but for cylindrical features of size local sizes would then be defined as circles that are an element of a sphere that is fit to the feature. This more clearly defines local size, and also provides a single center point from which a median line can be derived.

Dean

http://www.d3w-engineering.com

 

[DeanD3W]

Too bad we can't edit these posts... The last sentence of my post above should be "This more clearly defines local size, and also provides a single center point at each cross-section from which a median line can be derived." (added "at each cross-section")

Dean

http://www.d3w-engineering.com

 

[pmarc]

Interesting is a fact that Y14.5 actually does not precisely clarify how the actual local size measurement should be taken. Definition 1.3.54 says it is "the measured value of any individual distance at any cross section of a feature of size". Fig. 1-1 offers explanation that explains almost nothing IMO. It looks like the orientation of cross sections in which actual local sizes are shown are not that whatsoever - the sections seem to follow the curvature of derived median line (stay normal to it), though nothing in the text supports this observation.

J-P's most recent graphic seems to be in line with this too. Assuming that long dash-short dash center line on blue and red pictures is a derived median line of the banana shape feature, the measurement planes of actual local sizes are indeed always normal to the derived median line.

But this would mean that actual local sizes can be found only if derived median line is known. Isn't it crazy for someone who would like to use caliper to measure diameter of the feature?

 

[Belanger]

Thanks Jim, Dean, and pmarc! I think I'm good with my customer issue; they were having concerns about checking size vs. straightness, and questioning whether both should be based on the same cross sections. (One other variable in my situation is that the pin is relatively stiff, but could flex with pressure -- that's a whole different ball game than this thread.)

I never noticed the sphere description that Dean mentions. That's a good way to at least verbalize one of the cross sections. (The other variety being cross sections based on the UAME.)

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

http://www.gdtseminars.com

 

[axym]

J-P,

I agree that there are several issues with the definitions for actual local size and the derived median line.

Starting with actual local size, there is still ambiguity over what it means. Some statements in Y14.5 indicate that the actual local sizes are based on cross sections, and other statements indicate that they are based on 2-point linear distances. In any case, the actual local sizes do not need to be oriented normal to the actual mating envelope. I agree with pmarc that in Figure 1-1 the actual local sizes appear to be normal to the derived median line, although there is no requirement for this. The derived median line doesn't even belong in that figure.

Now on to the derived median line. The DML Straightness characteristic is one of the deepest swamps in all of Y14.5 :^(. Y14.5-2009 defines the DML in the following way:

derived median line: an imperfect (abstract) line formed by the center points of all cross sections of the feature. These cross sections are normal (perpendicular) to the axis of the unrelated actual mating envelope.

My first problem with this definition is that the center point of an imperfect cross section is not clearly defined anywhere. Not in Y14.5, not in Y14.5.1, not in ANSI B89.3.1. We don't know if it's the center of the maximum inscribed circle, minimum circumscribed, minimum zone, some sort of average, or what. So there are various possible DML's that one could establish.

The second problem (in my opinion) is the requirement for the cross sections to be normal to the AME. This makes sense in certain Straightness applications, but not in others.

Let's start with DML Straightness RFS. It would seem that the cross sections for the DML should be allowed to follow the curvature of the as-produced feature, as the cross sections for Circularity are allowed to. But this would be inconvenient to inspect, and I suspect that the requirement to orient them to the AME was put in to simplify things. I have trouble envisioning applications for this characteristic in the first place. Can anybody else name any applications where DML Straightness RFS makes sense? Perhaps these would guide us as to how the cross section centers would need to be defined.

The requirement for the cross sections to be oriented to the AME makes sense for DML Straightness at MMC, where the purpose of controlling the DML is really to control the surface of the feature so that it doesn't violate a certain boundary.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[pmarc]

Evan,
I support most of your statements, though still have some comments / questions:
- Just for clarity, derived median line is shown in fig. 1-1, but not in fig. 1-2.
- What makes you think that cross-sections for Circularity are allowed to follow curvature of the as-produeced feature? Per 5.4.3(a) of '09 edition "Circularity is a condition of a surface where for a feature other than a sphere, all points of the surface intersected by any plane perpendicular to an axis or spine (curved line) are equidistant from that axis or a spine". Per 1.3.28 feature's axis is "the axis of the unrelated actual mating envelope of a feature" (unless otherwise stated). So in my opinion, if we are talking about nominally straight cylinder, we are again, like for staightness, forced to use UAME concept to be able to check circularity error properly. Does ANSI B89.3.1 state otherwise? (My apologies if, by asking this question, the discussion will drift away from original path).

 

[axym]

Pmarc,

The reference to the spine (curved line) is what allows the Circularity cross sections to follow the curve of the as-produced feature. We can choose any arbitrary spine, to arrive at an optimal Circularity value. We can choose a perfectly straight one for simplicity, but we don't have to.

For DML Straightness RFS, I think it would make sense to allow the cross sections to be normal to an arbitrary spine.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[ptruitt]

Evan's last statement caught my eye: "...the purpose of controlling the DML is really to control the surface of the feature so that it doesn't violate a certain boundary."

Peter Truitt
Minnesota

 

[axym]

Peter,

Yes, this seems to be true for the MMC case. The meaning of Straightness at MMC in terms of the DML is very complicated. The DML is difficult to establish in the first place, and then it must be compared to a tolerance zone that expands differently at each cross section. I don't know of any applications in which the designer really intends to control the DML to a locally expanded tolerance zone. In all the applications I know of, conformance of the surface to the virtual condition boundary is what really matters. Because of this, GD&T books tend to deal with Straightness at MMC exclusively in a gaging context and ignore the DML altogether.

The end result of a Straightness at MMC tolerance is to open up the Rule #1 boundary by a certain amount. That's it, it's a relatively straightforward concept. But because of the way the Y14.5 standard has evolved, the tolerance is encoded in an indirect way that is unnecessarily complicated and difficult to understand.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca