Sr GDTP Y14.5-2009 Exam Review Ch-7 Sep2021

[metrologic]

Sr GDTP Y14.5-2009 Exam Review Ch-7 Sep2021

Hi Everyone! I'm up to section 7, Tolerances of Location. As you may recall, I'm working through the Y14.5 standard in preparation for upgrading my GDTP certification from the tech level to the senior level. Here are some questions I had from this section:


Q1. 7.2.1.1 Dimensions for True Position. What is the origin of the use of true in GD&T? Wouldn't nominal make more sense?

Q2. 7.4.1 Projected Tolerance Zone. The language in this subsection discusses projecting the tolerance zones associated with positional requirements to ensure the orientation variation of fixed fasteners does not interfere with mating parts. Am I the only one that prefers to think of positional tolerance zones as infinity long, and hence their projection seems kinda strange and unnecessary? In my opinion, the focus should be on identifying the feature axis orientation and relevant length --and then assessing whether that segment is contained by a tolerance zone that is extended axially indefinitely. Fig. 7-20 provides a good example of the problem at hand for fixed fasteners installed after assembly. As can be seen, the segment of projected feature axis that passes through the maximum thickness of the mating part depends on the variation in orientation of the fastener. Projecting a positional tolerance zone is trivial, measuring and projecting a feature axis is not. But maybe I'm thinking about this all wrong?

Q3. 7.4.5.1 Noncircular Features of Size at MMC. Subsections (a) and (c) provide an interpretation of positional tolerances for noncircular features of size. Both paragraphs describe a theoretical boundary that can not be violated. They seem to be redundantly saying the same thing if I'm not mistake. What is the difference?

Q4. Fig. 7-34 Positional Tolerancing, Boundary Concept. So figure 7-34 depicts positional tolerancing for rounded slots. The long axis of the rounded slots run left to right on the page, parallel to datum B. The rounded ends are given a "left right" positional tolerance of 1.5mm at MMC. The long straight sides of the slots are given an "up down" positional tolerance of 0.25mm at MMC. When I try to interpret the situation I see two features of size, a length defined by the opposing rounded ends and a width defined by the straight sides. Each of these two features has its own given positional tolerance, which is defined by two parallel planes running either perpendicular or parallel to datum B. Up to this point, I feel the concepts of feature axis location and virtual condition boundaries have been reasonably well defined. In this case however, what step by step logic will lead one from the feature control frames to the theoretical boundary that is "equal to the MMC size of the internal feature minus its positional tolerance". At first glance, I believed this would be identical in nature to a simple hole or stud that is controlled with bidirectional position tolerancing. But after thinking it over, the rounded slot seems to really be two features, each with their own positional control, while the simple bidirectional toleranced hole is one feature with two complementary positional controls. It seems to me, that the positional requirements shown for the rounded slot in figure 7-34 would conflict with each other and make it quite difficult to derive a theoretical boundary that shall not be violated.

Q5. 7.6.4 Concentricity. Is it allowed to use secondary or tertiary datum feature references when applying a concentricity or symmetry control?

 

[3DDave]

A1. I have a book from 1964 that calls it "true." This predates the ANSI series and is somewhere within the following efforts:

USASI Y14.5-1966
ASA Y14.5-1965 Dimensioning and Tolerancing for Engineering Drawings Proposed.
MIL-STD-8C (1963)
MIL-STD-8B (1959)
ASA Y14.5-1957

Also "GD&T" is the marketing term. There are no "Geometric Dimensions" or "Geometric Dimensioning" as defined in ASME Y14.5 It appears to have originated with Lowell W. Foster calling his series "Geometrics" Lowell appears to have been a Johnny Appleseed of Dimensioning and Tolerancing so people took what he wrote/taught as representative of the contents of the standards.

A2. an infinite projection of an as-produced axis will leave any infinite tolerance zone. The only meaningful use is to either imply the extent of the tolerance zone as based on the amount of material it's embedded in or to be explicit about where the tolerance zone lies and whether the axis remains inside it/them. Beyond that it's unclear to my why it would be unclear to you.

A3. without looking more closely I think this is from the department of repetitive redundancy. Consider that '2009 was rapidly replaced with the 2018 version - the typical interval is 10-15 years and they did this in less than 8, suggesting major disappointment with the 2009 effort. I'd say, "Look at the development notes" but ASME forbids releasing that info.

A4. Yup. It's a polite fiction that the zone will be created exactly the same way that location variation of the feature would - which it won't. Basically they are only going to be checked independently if the two location tolerances are given. AFAIK the same would be the way that VSA would simulate that condition - the extreme width at the tangencies of the arcs and the width of the slot ignoring the rounded ends.

A5. If you mean the concentricity geometric characteristic symbol then, no. A similar effect with position tolerancing allows multiple datum feature references. Concentricity is removed in the 2018 version along with symmetry.

 

[Burunduk]

Regarding Fig 7-34,
You are right that what is shown is two separate features of size, both of the "regular feature of size" type (per subsection 1.3.32.1). One is a set of "two opposed parallel surfaces" (the slot width), the other is "two opposed parallel elements" (the slot length). Each has its own positional tolerance requirement modified MMC, and its own VC boundary, but those two boundaries are unified to one slot-shaped boundary from practical reasons, because this is the most reasonable way to make a hard gage or generate a virtual boundary for both requirements. The standard doesn't go into detail explaining this nuance, and I'm not sure that it should. In any case, the two boundaries or one combined boundary, whatever you prefer, are well-defined relative to the datum reference frame, their establishment does not depend on the actual feature and there is no internal conflict within that boundary or between those boundaries.

As a side note, the feature made of the two line elements defining the slot length, although being a feature of size, is not a good candidate to be controlled for position at RFS. Because, although being directly listed in the FOS definition and commonly controlled for position, it cannot constrain an unrelated actual mating envelope.

 

[Belanger]

Nominal simply means "named" as in choosing a number for naming convenience. Thus, a 2x4 is the "nominal" size of a board. Of course, 2x4 is not its "true" size!

 

[metrologic]

@3DDave
What does VSA stand for? (it's in your A4 answer)

 

[3DDave]

Variation Systems Analysis. It's one of the very few tools that took the definitions for variability allowed by the descriptions in Y14.5.1 and created matching mathematical models. It handled cases that the "vector loop" competitors didn't.

They got bought out by a big CAD company which probably tried selling it as a check-box item so they could say "we offer tolerance analysis" but big CAD buyers look at the license price and say "Our users already know how to do that" and take a pass. Sadly - they don't and won't and so usable parts get tossed, unusable parts get accepted, and most companies put ridiculously large tolerances on drawings and get away with it because fabrication technology is often far better than that.

All those costs are far higher than VSA cost, but they are hidden, appearing on no budget.

 

[metrologic]

@3DDave
Interesting... What would be a good example of a "vector loop" competitor?

 

[3DDave]

As far as I know

https://www.sigmetrix.com/history-of-sigmetrix/

which originated with ADCATS at Brigham Young University.

 

[metrologic]

@3DDave

Do you have any experience with measurement software, CMM software, etc? Everything I've had the opportunity to work with seems terrible in regards to geometric tolerance support.

 

[3DDave]

Most of my experience was explaining to inspectors why part features aren't the same thing as datum feature simulators. Also trying to explain that tolerance analysis isn't suitable to explain massive warpage of weldments when a .125 skip weld is replaced with a .63 multiple pass continuous weld on .25 material. But sure - let's open the tolerances up from +/-.030 to +/- 1.00 because it's a pretzel twisted piece of metal. p.s. - I gave that tolerance once when it really didn't matter and they whined it was too much.

 

[metrologic]

Back to rounded slots...

As a side note, the feature made of the two line elements defining the slot length, although being a feature of size, is not a good candidate to be controlled for position at RFS. Because, although being directly listed in the FOS definition and commonly controlled for position, it cannot constrain an unrelated actual mating envelope.

Yeah, it's specifically the rounded ends that throw a monkey wrench into my thought process. In contrast, if the Y14.5 example illustrated a rectangular slot, I would think it would be a text book application of geometric tolerancing. But as a sanity check, we would expect a rectangular BOUNDARY for a rectangular slot, right?

 

[Burunduk]

Yes, a rectangular slot toleranced similarly would produce a rectangular boundary. That would be a more well-behaved, although much less common (at least for machined parts) application of the same principle. I suppose that if the slots' application would be trivial per the concepts established elsewhere, it would not get a unique coverage in the standard, which not many applications have.

 

[metrologic]

@Burunduk

As a side note, the feature made of the two line elements defining the slot length, although being a feature of size, is not a good candidate to be controlled for position at RFS. Because, although being directly listed in the FOS definition and commonly controlled for position, it cannot constrain an unrelated actual mating envelope.

Burunduk, wouldn't two opposing points or two opposing line elements be able to contain an unrelated actual mating envelope composed of a pair of opposing parallel planes? As an example, I'm thinking of how one would measure a gauge block with a micrometer. Theoretically on a perfect part, I don't have a problem with deriving a center plane from a pair of opposing points or a pair of opposing line elements. But in practice on a real world part, I find the concept is almost impossible to apply. No one seems knows which exact pair of actual elements should correspond to the pair exemplified on the blue print.

 

[Burunduk]

metrologic,
Actually as I think of this now again it is more of a problem for external features of similar shape as the slot than it is for a slot.
With two opposed line elements or two opposed points that do not lie on planar surfaces, such as the line elements along the rounded ends of an elongated tab, the physical feature itself can't constrain the orientation of the two parallel planes Unrelated AME. The UAME will tend to rotate as it keeps contracting to satisfy the condition of minimum separation between planes.

 

[metrologic]

Back to rounded slots...

So I found a copy of Y14.5.1M-1994 REAFFIRMED 2012. Section 5.6 covers position tolerancing for elongated holes at MMC. "Such tolerancing is always considered to be bidirectional in nature, even if a single tolerance is applied. Only a surface interpretation is provided." As in Y14.5, a virtual condition boundary is defined, but there's no proper derivation from position tolerancing principles. I think I am with 3DDave here, most likely the shape of this provided boundary is a polite fiction.

 

[Burunduk]

metrologic, If you think rounded slots are an unconventional application of position tolerancing, wait till you get to figure 8-24 from the profile section where profile in conjunction with MMC position is presented. The surface application of MMC opens many surprising possibilities.

 

[metrologic]

Yeah, I'm studying section 8 right now. I thought about bringing figure 8-24 into the present discussion, but perhaps it's best to come back to it in my next study review thread.

 

[metrologic]

@3DDave

A2. an infinite projection of an as-produced axis will leave any infinite tolerance zone. The only meaningful use is to either imply the extent of the tolerance zone as based on the amount of material it's embedded in or to be explicit about where the tolerance zone lies and whether the axis remains inside it/them. Beyond that it's unclear to my why it would be unclear to you.

Forget projected tolerances --let's use the context of regular position tolerances. As far as I know, Y14.5 always depicts regular cylindrical position tolerance zones as running the length of the material. But the as-produced hole axis are always shown over extended. See Fig. 7-8 for an example. If the extended axis of the as-produced hole pierces the end face of the tolerance zone cylinder, and not the tolerance cylinder side wall, then it passes inspection in accordance with the axis definition. Essentially what I was saying up above, is that I'd rather check conformance with the infinities flipped around. Let's not over extend the hole axis. Let the hole axis be limited to the length of the hole. Then check if this 3d line segment is contained by an infinitely long cylinder tolerance zone. If it breaks the cylinder wall it's out of tolerance, if not, it's in.

 

[Burunduk]

metrologic,
There is nothing in the Y14.5 standard that suggests that the feature axis has to be extended any amount beyond the length of the feature. Centerlines shown slightly longer than the features they are associated with is just a drawing representation practice, and not something that may influence conformance to tolerances. That said, there is nothing that prevents anyone from extending either the tolerance zone or the axis beyond the actual length of the feature. So any of them can be extended as long as you don't extend both, because you don't want to detect a violation of the tolerance boundary or report a measured value at some point in space that doesn't affect the function of the feature. But as far as the standard is concerned, both the tolerance zone and the controlled axis can be just as long as the actual as-produced feature, or as specified by the projected tolerance zone notation.

 

[metrologic]

eh I think I've got the flu Burunduk. I'm going down.[pre][/pre]