Spinoff: Using position to only control perpendicularity 1

[Belanger]

I know this has been discussed several times as an offshoot of other threads, but I can't recall if it's had its own thread. The dilemma is whether a position tolerance can be applied to a traditional single hole through a plate, but the only datum referenced is the face of the plate. The objective is to simply let it control perpendicularity. I say, nay nay

The ASME standard tells us that position is "the location of one or more features of size..." (paragraph 7.2, emphasis added). And each of the subparagraphs of 7.2 indicate location as the main objective. So by definition, a position tolerance must control location.

We all agree that in the classic hole-through-plate example a position tolerance will have perpendicularity come along for the ride. But that's just it -- it comes along for the ride! If you want perpendicularity to be the vehicle, then you should use the perp symbol (or even angularity or profile!).

An analogy: if I say to you, "Please stop by the grocery store. We really need some bread for tonight. And if you want to get some cookies, that's OK too." Suppose you come home with only cookies. The main objective hasn't been met.

Sometimes folks will toss out the profile symbols as examples where different aspects can be opted in or out. But at its base, a profile tolerance is a form control. That's it. If you add datums, then you get orientation and maybe location and size (all around can do that too). But notice that profile is "upwardly mobile." That's fine.

But to use position to achieve only perpendicularity is "downwardly mobile," and the symbol fails to do its inherent job, which is location. It's akin to using parallelism to control ONLY flatness -- a higher-order symbol can't be used to achieve a lower-order symbol while cutting out its own meaning.
Thoughts?

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

http://www.gdtseminars.com

 

[MechNorth]

Technically, yes, because a location control also gives you the orientation. Practical use? Again, no.

Sorry, J-P, but we're going to have to agree to disagree here. BTW, it's still debated at the committee level too, so it's not like we're unique here.

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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[Belanger]

Not sure what you're saying yes and no to. But OK to end it here if you wish.

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

http://www.gdtseminars.com

 

[MechNorth]

To the 34deg example.

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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[Belanger]

Legal to use parallelism on a 34 degree angle? Please tell me you read the question wrong!

If you ponder the parallelism scenario for 34 degrees (in light of paragraph 6.3.2) I think you'll change your answer. Then apply that same logic to position in light of paragraph 7.2: Position, by definition, controls location.

At any rate, it's been good to hear views from all sides on this.

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

http://www.gdtseminars.com

 

[axym]

J-P,

I agree with you that the standard forbids the use of Position in your single-perpendicular-hole example, but only because of the statement in 7.2 requiring either multiple features or at least one datum reference. The other argument that Position is by definition a location tolerance, and so it must locate, is on shaky ground because Y14.5's classification of characteristics is not completely consistent. For example, Concentricity is classified as a tolerance of location and Circular Runout is not. There is no question that Circular Runout locates the feature to the datum axis to at least the same degree that Concentricity does. So I try not to read too much into the "intrinsic" meaning of words and terms that Y14.5 has chosen to label things with - we can argue about that forever (and often do). We have to look past the words and get to the facts of exactly how the tolerance zones work, and then see how they would apply in certain situations.

So can Position be applied to the single perpendicular hole even though the location is not controlled? Yes, it can. The tolerance zone and DOF constraints are well defined, and there is no ambiguity. Does the standard allow this application? No, it does not, because the application violates the criteria of 7.2. Would I recommend that Position be applied to a single perpendicular hole? I would say no, because the Perpendicularity tool exists and is more straightforward. I would say this even if the use of Position in this application was legal.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[Belanger]

Evan... When I say that position controls location, I say so not because position falls into the category of location, but rather because the very definition of position says that it controls location. (I hope that doesn't sound too confusing.)

I can't help but keep pushing the parallelism analogy: the standard says that parallelism is for features "equidistant" to one or more datum references. Can I use parallelism on another angle? You and Jim imply that I can (while acknowledging that it violates the standard), because I would have a basic angle -- making the tolerance zone and the DOF well defined!

So while we all agree that perpendicularity on a single hole is the best way to go, I can't see why so much energy is expended in justifying position (even obliquely) for a situation that fails to meet its definition per the standard.

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

http://www.gdtseminars.com

 

[MechNorth]

Sorry, I had misread the question on the angle, as position. No, parallelism cannot control an angle other than 0/180-deg.
Parallelism & Perpendicularity are specific cases of angularity, and are applicable only at those conditions. Not the same as position, which controls all orientations as a subordinate control.
J-P, as long as the English language (or any other for that matter) allows the reader to provide their own personal emphasis on the text, there will be differences of understanding. Understanding that difference, and understanding the other person's perspective without agreeing with it is a critical ability and the precursor to a successful debate.

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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[Belanger]

Agreed on your observations about language.
But it's interesting that we all understand what "equidistant" means when discussing parallelism, but we don't all understand what "location" means when discussing position.

I gave it the old college try, though

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

http://www.gdtseminars.com

 

[MechNorth]

And they were good arguments, J-P. You should register your comments with ASME re the lack of clarity / ongoing disagreement. It will be addressed by the Y14.5 committee, though there's no guarantee of resolution.

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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[axym]

I didn't agree that Parallelism could be used on features at angles other than 0/180, sorry if I somehow implied that.

I would say that a better analogy would be specifying Angularity on a feature with a nominal 90 degree angle to the datum. The tolerance zone and its constraints would be well defined and unambiguous, but as far as I remember this was not allowed in the '94 standard. Presumably the use of Angularity for 90 or 0/180 angles was considered confusing because the Perpendicularity and Parallelism tools are available for those cases. This has been revisited and the '09 standard allows the use of Angularity for all basic angles. Nothing has changed with the meaning of the tolerance zones, only the legality of certain applications has changed.

I'm not really surprised that we have more trouble discussing location and Position than equidistance and Parallelism. Parallelism only deals with single special-case features that are nominally equidistant to a datum plane or datum axis. Position covers the location of features that are more complex and may not be equidistant from anything. When we're dealing with location in a Position tolerance context, the concept of equidistance is not sufficient to describe what's going on.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[axym]

I would agree with Jim that Profile is not a form control at heart. In its default state, with basic dimensions defining the true profile, a Profile tolerance controls more than just form. It also controls the size of the feature as applicable, which form tolerances never do (e.g. a Cylindricity zone can expand and contract to adapt to the as-produced size of the feature). I hesitate to use the word "size", because this control is not confined to features of size and doesn't require any opposed surface points. When applied to any curved surface, Profile will control the curvature. This goes beyond controlling form.

The applications in which Profile controls only form are few and far between. One possibility would be to apply Profile to a single planar surface, which would give the same end result as Flatness. Not really a best practice. The standard also contains an example in which Profile is made to control only form, by combining it with directly toleranced dimensions (the conicity thing). This is a questionable practice at best - the mixing of directly toleranced dimensions with geometric tolerances is non-rigorous.

I would say that Profile has much more in common with Position than it does with form characteristics. I would go as far as to say that Profile is to surfaces what Position is to features of size. When applied to multiple features, Profile controls relative location (and relative orientation) in the same way that Position does. When applied with datum feature references, Profile controls location (and orientation) to the datum reference frame in the same way that Position does. The tolerance zone mechanics are very similar.

It's true that Profile controls form and size, and Position does not. This is because Profile is applied to surfaces and Position is only applied to resolved geometry (axes, center planes, center points) which by definition have perfect form and zero size. If we bring in the surface interpretation of Position, the distinction between Profile and Position starts to blur. It is possible to envision Position at MMC as a Profile tolerance with a single boundary.

This is why it seems very inconsistent to me to have Position treated with a "must locate" directive and Profile treated with a "form plus options" flexibility. If we have a special case in which an orientation tolerance would suffice, I would say that it would be equally inadvisable to apply Profile as it would be to apply Position.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca