The "purpose" for definition of FOS and AME (ASME Y14.5 2009) Conical /Tapered features 3

[dtmbiz]

Unfortunately the ”purpose” for the need to define certain and specific terms and concepts", which would be helpful in determining the “intent” of the standard’s definition regarding inclusion and /or exclusion for features considered in respect to those definitions and concepts. More specifically regarding FOS and AME for this discussion relative to conical and tapered features.

Here are my general points for discussion. I hope others will post theirs.

Purpose of FOS: (Feature of Size) My understanding of its purpose is to identify features that have center planes, axis and center points in order to locate and orient tolerance zones.

Purpose of AME: (Actual Mating Envelope) My understanding of its purpose is to establish a produced FOS’s actual location and orientation of it’s center plane, axis or center point by use of a AME Datum Simulator (gage) in order to verify compliance with the defined tolerance zone by comparing the true feature's vs. the produced feature's location and orientation of axis, center plane or center point.

It is also my understanding that applied geometric controls to an FOS, “other than size” can only be verified after the produced feature’s has been verified to be within size limits.

Conical and tapered feature’s: In a previous thread (thread1103-460248) there are arguments that conclude that these type features cannot be classified as features of size because in the case of a cone (conical feature) there is a limit to expansion or contraction about the apex, and similarly a limit to intersection of tapered surfaces beyond their intersection.


Disagreement with argument that 'conical and tapered surfaces are not FOSs because an AME cannot be defined":
Conical and tapered surfaces can and do have an AME in the physical world (vs a purist mathematical theoretical world) which can identify a produced FOS axis or center plane.
The AME’s Datum Simulator would not expand or contract beyond the limit of the apex of a conical surface or intersection of tapered surfaces whether or not they actually occurred within the size limits of the feature’s extent. Concluding that the apex of a conical surface and the intersection of tapered surfaces would be to one side of the tolerance limits and would be the minimum or maximum allowance which therefore would be the minimum or maximum limits of contraction or expansion of an AME. (internal / external feature). Expansion or contraction is limited to be within limits of size vs. infinite or unlimited

**** Also would like to mention in relevance to the FOS definitions, specifically around the standard's use of “may”, and the definitions of words ; may, must , will, shall in the engineering environment (English that is). “May” is a permission word that “allows” and is not mandatory. “Shall” is a “mandatory requirement” which is not used in definition of FOS

 

[chez311]

Plus-minus toleranced dimension applied on a feature in the Y14.5 standard is a clear indicator that the feature is treated as a feature of size

Y14.5-2009 includes several examples of nonFOS dimensioned with +/- tolerances. This in and of itself is not an indicator of whether its a FOS or not. For example:
fig 2-1 angular features
fig 2-4(a) heights
fig 2-22 fillets

For that matter, I do not consider sliding, as in a conical envelope that "keeps contracting", a basis for rejecting a UAME. The envelope can as well stop contracting as soon as the slightest normal force is exerted, the mechanism that would allow it is irrelevant.

The amount of contraction would be arbitrary. For a primary datum feature it can "contract" 0mm, 1mm or 100mm and provide identical simulation.

I believe Norm’s posts refuted Axym’s posts nicely.

I believe quite the opposite and have been unconvinced of anything to the contrary.

In practicality his post admits that these features “work in practice”

I think thats taking the statement out of context. He says specifically simplified (+/-) dimensioning can work in practice, there is no deviation from the statement that a cone is not a FOS. Basically that the ambiguity of measuring an angle (or combination of angle and circular element) with +/- dimensions is reduced significantly because of the precision with which tapered features are often produced (ie: collets, toolholding tapers, etc..) results in very little form error. Apply the same concepts to a feature with increasing form error and the ambiguity increases proportionally.

If directly toleranced dimensions are used (such as a combination of an angle tolerance and a size at one end of the cone), I don't think that stackup is the only complication. The allowable variation isn't rigorously defined. The only reason that these methods work is that cones are generally manufactured with very little form error (such as the tapers used in the molding and tooling industries). Because of this, the results of the ambiguities are reduced to an insignificant magnitude. This makes the simplified dimensioning work in practice, but let's not pretend that it's rigorous.

If one wants to go off into “theory land” without regard for reality

Why is asking what the size of a cone is considered going off into "theory land" ? This is a perfectly valid question for which I have not seen a sufficient answer. The typical answer is that the size of circular elements can be controlled - if this is the case then without location specification a specified size tolerance can be satisfied anywhere along the length of a cone and is mostly meaningless without location specified to another feature. Additionally the swept spheres definition of size per Y14.5.1 is not supported for a conical feature.

The gotcha is that if a single number represents all conical surfaces of every extent then any other description is not of the cone but of the extent of the cone relative to some reference. It's why a single countersink tool can produce conical surfaces with different truncations, but the tool fits them all exactly; they are all the same.

Excellent example, couldn't have put it better myself.

 

[Burunduk]

The amount of contraction would be arbitrary. For a primary datum feature it can "contract" 0mm, 1mm or 100mm and provide identical simulation.

Not sure I understand the point of this. The amount of contraction is always arbitrary, also for a cylindrical simulator.

As dtmbiz noted, the only thing that matters is the final result - the UAME definition describes an envelope that ended up being "expanded" within or "contracted" about the feature. A feature of size ended up being contained by it or containing it. The process that led to it is not important. Also insignificant is how much offset was during the simulation from any reference/origin, and what mechanism caused the process of contraction/expansion to stop.

The very definition of a conical taper includes the concept of size.
If you read para. 2.13 in Y14.5-2009 you see that a "conical taper" is not only a feature of specific form but also a numeric value defined by the change in size per unit of length: (D-d)/L. If the definition would intend to detach conical features from the size characteristic it would probably define a cone by a single parameter of the included angle, without any "d's". That would be a very poor definition, that doesn't tell anything about the cone's unique geometry.

 

[3DDave]

The contraction for a cylinder starts at MMC/MMB size. There is no MMC/MMB size for a conical surface.

(conjecture 1) A cone with a specific base diameter (maximum size by tolerance) is overlaid by and positioned to another cone with a specific base diameter (minimum size by tolerance),
both have the same perfect surface angle (perfect as defined by Y14.5 via gage tolerances), and both have the same cone height.

(conjecture 2) You can't be saying that both cones would have the same size truncated diameter ?

Conjecture 1 reply: They cannot both have the same cone angle, the same height and different truncations.

Conjecture 2 reply: What I wrote and what you concluded are significantly different. I mentioned two cones of different angles that have the same truncation diameter and asked if that meant they were the same size. Not sure how that can be read any other way, but here we are. It's not as if that BT drawing isn't hopelessly double dimensioned.

The response to your fourth question is, what makes that angle so perfect in real parts that the angle is identical?

The BT taper is used RFS, depending on elastic deformation of both parts to negotiate a mutual location and orientation not precisely predictable from the geometry as given.

If it's a feature of size, what determines the value that will be reported on the inspection report and used in the engineering analysis of actual part variation?

 

[chez311]

Not sure I understand the point of this. The amount of contraction is always arbitrary, also for a cylindrical simulator.

As dtmbiz noted, the only thing that matters is the final result - the UAME definition describes an envelope that ended up being "expanded" within or "contracted" about the feature. A feature of size ended up being contained by it or containing it. The process that led to it is not important. Also insignificant is how much offset was during the simulation from any reference/origin, and what mechanism caused the process of contraction/expansion to stop.

So you would say an external cylindrical primary datum feature specified at RMB which comes in at 10mm dia (assuming no form error) would be sufficiently simulated by a simulator which is arbitrarily contracted to only 12mm or 11mm dia?

If you read para. 2.13 in Y14.5-2009 you see that a "conical taper" is not only a feature of specific form but also a numeric value defined by the change in size per unit of length: (D-d)/L. If the definition would intend to detach conical features from the size characteristic it would probably define a cone by a single parameter of the included angle, without any "d's". That would be a very poor definition, that doesn't tell anything about the cone's unique geometry.

(D-d)/L is the equivalent of slope for a feature of revolution.

The inclination of any line can be described by its slope or angle. The latter is not "poor" and tells you everything you need to know about its orientation. Interesting that you should bring up the example of slope, as it results in a non dimensional number of which the units of "size" (neither in diameter or length) do not appear as they are cancelled out and yet still fully describes a taper/cone.

 

[Burunduk]

So you would say an external cylindrical primary datum feature specified at RMB which comes in at 10mm dia (assuming no form error) would be sufficiently simulated by a simulator which is arbitrarily contracted to only 12mm or 11mm dia?

That's not what I said. The simulator would have to contract to whatever diameter value the feature's UAME was produced at. That doesn't mean that the amount of contraction matters. It could have started contacting from 11, 100, or 1000000 millimeters diameter.

(D-d)/L is the equivalent of slope for a feature of revolution.

The inclination of any line can be described by its slope or angle. The latter is not "poor" and tells you everything you need to know about its orientation.

Then the "slope equation" of a feature of revolution is size-dependent. It is the change in size per unit of length. The units of size are canceled out in the final calculated value yet size is a parameter in the calculation.

Describing a cone only by the included angle is indeed a very poor "definition" because an angle can "define" any geometry that can be dimensioned with an angular dimension. It could as well be a corner between two flat surfaces. It doesn't tell you that we are dealing with a cone.

 

[Burunduk]

They cannot both have the same cone angle, the same height and different truncations.

If by different truncations you mean different truncation sizes, then yes they can and they must. Because the base sizes are different.

 

[3DDave]

It's like Through the Looking Glass.

a1 = a2, h1 = h2, d1 <> d2.

(being very pedantic)
a = included angle,
h = distance from vertex to truncation,
d = diameter at truncation

Is this possible for two cones, 1 and 2?

Burunduk says these equalities are consistent.

Anyone else?

 

[Burunduk]

By cone height, I meant distance from base to truncation. I am pretty sure that this was also the meaning in "conjecture 1" to which you responded. What makes you measure the height from apex to truncation? Do you measure material or "air"?

 

[3DDave]

How do you measure to an axis? Same way.

 

[Burunduk]

I don't think you understand. "Truncation" implies a truncated cone (see image below).
The "truncation" is the small end opposing the base.
You wrote: "h = distance from the vertex to truncation" which is the height of the "missing" piece removed by truncation. I understand "cone height" in dtmbiz's example ("conjecture 1") as frustum height. The question wouldn't make sense any other way.

 

[3DDave]

Truncation in solid geometry presupposes a finite cone with the top removed. Since it doesn't describe how the rest of the infinite cone was removed to create the base, I adapted the word to describe that as truncating the infinite part. Sorry for the confusion.

That doesn't change my mind in the least. His example was a statement of wrong reasoning. It still failed to establish that they are different "sized" cones. If I drop them into a perfectly angle-matched fixture, there is zero gap with the fixture and therefore zero mobility. The only thing that changes is the distance of the cut surfaces to the vertex. So the size doesn't change.

And, as in my example, where the angles aren't the same, then either the base or the smaller diameter will be an exact fit and the exact same distance of that feature from the vertex; no mobility either.

This is not even touching on the case where neither the base nor the smaller surface are at right angles to the axis of the cone. What diameter of which ellipse could govern the "size"? And what if those cuts are not planar? Such a thing happens with conical features that intersect cylinders and other shapes. What height is found for those? That's why the vertex location and angle are the key features of a conical surface and not any other mix of dimensions that happen to work out some way and interact with other surfaces to determine.

I know I asked someone to produce a document with their reasoning on this matter, preferably with diagrams and examples of how size might be determined, but so far no luck. Instead it's discussion of a particular subset of a subset of a subset to try to justify a really bad generalization without a demonstrable adequate application.

Look at this as another example: three surfaces of a cube sharing a corner. Are they a feature of size? They also fit into a matching fixture and manage to control 6 degrees of freedom where a cone only does 5. They form a taper, right? But just like the cone the only sure way to measure the "expansion" or "contraction" of the idealized mating surface is to measure the linear distance between the vertex of the ideal surfaces to the cube surface vertex, which will always be coincident when the surfaces are coincident. Those three surfaces on a cube are the same in that respect as all of the similar surfaces on all cubes. Which either means that tiny cubes are the same size as giant cubes or that those three surfaces do not represent size. I'm voting for the latter. Oddly, it will also match any three surfaces sharing a vertex of any rectangular parallelepiped, so the feature is not unique to cubes.

I get that there's a desire to do -something, anything- to somehow account for variations in certain volumetric features subject to external restraint in order to make up some palatable story, but I haven't seen a generalized way that is simple to describe or any work on that desire at all. Plenty of claims that it just applies in some nebulous manner, but no solid work to prove it. Had the work been done it should have been included in the updated math standard; maybe they forgot?

Just like any upset of using "truncation" to independently describe each end of a conical body, I seriously object to calling something without a definable size a "feature of size" even if "irregular" gets stuck on it to hand wave the fact that it isn't one.

 

[Burunduk]

Just like any upset of using "truncation" to independently describe each end of a conical body, I seriously object to calling something without a definable size a "feature of size" even if "irregular" gets stuck on it to hand wave the fact that it isn't one.

If by "upset" you mean me, there was no "being upset" from my side, only a will to get things clarified. Now you understand dmbiz's example (even if you don't agree with the intended point of it), and I understand what you tried to say in response: "They cannot both have the same cone angle, the same height and different truncations (base diameters)".

 

[dtmbiz]

None of "solid geometry’s” defined shapes have any particular size until values are assigned to its generic geometric definition nor is it’s location and orientation in 3D space defined until dimensions are used to position and orient.

To argue that a conical feature doesn’t have size when values are added to that particular conical feature’s solid geometry definition, is to argue against solid geometry’s definition of a cone

“Perfect” in Y14.5 is accepted for Datum Features, basic dimensions or boundaries when simulated by processing equipment. (In reference to my post for a point with cones having the same angle)

To argue against a smaller volume cone being “overlaid by” or “positioned perfectly to” a larger volume cone which both share the same height, axis, and base center point which create the extents for a tolerance zone, is to argue against Y14.5’s concept of boundaries and tolerance zones; Boundaries defined for an imperfect produced conical feature’s surface to lie within or a positioning tolerance zone for it’s axis to lie within.

 

[3DDave]

Smaller volume, not size? So if two cones have the same volume they are the same size?

Write a paper when you have it all figured out.

Circular cone; elliptical cone; skew cone; cones having non-planar truncations; three planar faces sharing a vertex; three planar faces sharing an intersection but with intervening chamfers; multiple non-intersecting, non-parallel holes.

These all meet the requirements for the IFOS definition, but I see that the committee really meant, by way of their examples, that an IFOS stands in place of a regular feature of size that has only one controlling dimension. They just f'd up horribly, leading people to think it meant cones as well, which is why they didn't show any examples of cones.

 

[Burunduk]

but I see that the committee really meant, by way of their examples, that an IFOS stands in place of a regular feature of size that has only one controlling dimension

All examples of irregular features of size except fig. 8-24 have only one controlling dimension. As a matter of fact fig. 8-24 has only one controlling dimension too. That is if you pick one and don't count the others.

 

[3DDave]

Figure 8-24 was chosen by a different sub-group on the committee as an IFOS, not by the profile group, which is why IFOS is not mentioned in section 8.

No response at all about the cases I mentioned tells me there is no firm argument for cones.

Picking a nit instead.

Fig 8-24 was not designated as an example of IFOS, it's titled: Fig. 8-24 MMC Principle Used With Profile Controls., not Irregular Feature of Size of a Boundary. It also has no directly toleranced dimensions, violating the IFOS definitions.

 

[pmarc]

Fig 8-24 was not designated as an example of IFOS, it's titled: Fig. 8-24 MMC Principle Used With Profile Controls., not Irregular Feature of Size of a Boundary.

Para. 2.8. Y14.5-2009:
"RFS, MMC, and LMC may be applied to geometric tolerance values on features of size. See Figs. 7-34 and 8-24."

Para. 7.4.5.1(c), Y14.5-2009:
"NOTE: This boundary concept can also be applied to other irregular shaped features of size - such as a D-shaped hole (with a flattened side) - where the center is not conveniently identifiable. See para. 8.8"
Para. 8.8 is the one which Fig. 8-24 belongs to.

It also has no directly toleranced dimensions, violating the IFOS definitions.

IFOS definitions do not require use of directly toleranced dimensions. They require features or collection of features to be directly toleranced. Direct tolerancing methods are defined in para. 2.2. Subparagraph (c) seems to fit to what is shown in Fig. 8-24.

 

[Burunduk]

Figure 8-24 was chosen by a different sub-group on the committee as an IFOS, not by the profile group, which is why IFOS is not mentioned in section 8.

...

Fig 8-24 was not designated as an example of IFOS, it's titled: Fig. 8-24 MMC Principle Used With Profile Controls., not Irregular Feature of Size of a Boundary. It also has no directly toleranced dimensions, violating the IFOS definitions.

I don't know which sub-committee included what figure, and I don't think that it matters how the figure is titled in this case. All that matters is that there is a position tolerance applied to the feature in question. The only conclusion from that is that it is treated as a feature of size:

" 7.2 POSITIONAL TOLERANCING
Position is the location of one or more features of size relative to one another or to one or more datums. "

As for directly toleranced dimensions, the other IFOS examples do not have those either, see figures 4-33, 4-34. It's not a violation of irregular FOS definition (but of regular FOS definition - which they are not).

Smaller volume, not size? So if two cones have the same volume they are the same size?

Just like with any other 3D tolerance zone I don't see why it can't be described in terms of volume. For example, if profile tolerance is applied on a basically dimensioned cylinder, the tolerance zone for the size and form (also location and orientation if applicable), of that cylinder is captured in the gap between the volumes of two coaxial ideal cylinders. I see no issue with dtmbiz's description of the equivalent conical tolerance zone. Feature size and tolerance size is not measured in volume units. I understand it as an auxiliary to describe how a tolerance zone is "created" and where it is located.

 

[dtmbiz]

3DDave

elliptical cone; skew cone; cones having non-planar truncations; three planar faces sharing a vertex; three planar faces sharing an intersection but with intervening chamfers; multiple non-intersecting, non-parallel holes.

No response at all about the cases I mentioned tells me there is no firm argument for cones.
Picking a nit instead.

Fig 8-24 was not designated as an example of IFOS, it's titled: Fig. 8-24 MMC Principle Used With Profile Controls., not Irregular Feature of Size of a Boundary. It also has no directly toleranced dimensions, violating the IFOS definitions.

1) I believe you will find 2.13 CONICAL TAPERS excludes most if not all of your examples as complying with the accepted definition for conical features in Y14.5 2009.

2) You are saying Fig 8-24's internal feature does not have "directly toleranced dimensions" because the dimensions are basic dimensions ? Excluding it from a ""Linear Extrucded" I-FOS?
Then in Fig 4-33 "Datum B" is not a FOS because it is defined with a basic dimension ?

Smaller volume, not size? So if two cones have the same volume they are the same size?
Write a paper when you have it all figured out.

a smaller volume cone being “overlaid by” or “positioned perfectly to” a larger volume cone which both share the same height, axis, and base center point

Discard what doesn’t support your argument ?
Disregard context ?

You don’t believe volume of a shape is a means to define size.
In the prescribed example for a smaller volume resulting from a particular geometric shape which relates to another larger volume of the same type shape,
with specificallly defined shared geometric characteristics, cannot define a Y14.5 geometric concept of boundaries resulting in a tolerance zone ?

It is not my intent to change your views, rather to explain mine.
No value to misrepresent my posts.

A single cone at a specific taper angle has a segment cut some distance above the apex (maybe 2”) and is slid along the axis to a height of (maybe 4”) from the apex.
A larger base diameter is cut at that height.
Both conical features are cut 1” below the coincident base diameters.
They then have a consistent space between the two conical features, a tolerance zone.

That example is not geometrically possible?


Burunduk
It appears that 3DDave was referencing the manufactuing processs of "upsetting" which is in a general sense amounts to the decreasing of height by "hammering"
to increase the diameter in a certain area of a cylindrical shape. The process doesnt necessarily intend to or result in a defined conical feature.

 

[3DDave]

I think you'll find that all of them qualify under your concept of "irregular feature of size" and if you have no general purpose explanation for evaluating them, then you are hand-waving your current explanations for right cones.

I ask you questions. I get arguments about the questions, not answers.

"A single cone at a specific taper angle has a segment cut some distance above the apex (maybe 2”) and is slid along the axis to a height of (maybe 4”) from the apex."

There is no cone "above the apex" so it cannot be cut there. The apex of a cone is the highest point. Just draw a picture.

No - he was upset that I was misusing "truncation" so much so that he went and searched for pictures to hammer the point. People who aren't upset just put in a link.

LOL. dtmbiz, I guess Hammering on FOS for 10 years has still not gotten you the answers you like. When I proved that the Fig. 4-16 case (c) analysis in the standard was garbage, I did not argue with people that they were wrong, I had direct proof. Since you know you are right, why does it bother you otherwise?

"You don’t believe volume of a shape is a means to define size." Congrats. Got that right. I don't. But you have not given a means to describe size variations in cones and, since I have asked and there has been zero answer, I guess there is not one. That means that you cannot tell me when two cones are the same size if they are not perfectly identical.

Best of luck in your future endeavors.