FOS defined with basic dimensions... legal?

[dtmbiz]

If I have a cookie-cutter shape that I want to define with basic dimensions and use a profile tolerance, can two parallel and opposing flat surfaces in that profile be used as a Feature Of Size?

It seems to be that it could be, as the size it the basic dimension with the tolerance applied from the profile FCF.

Is this correct?

 

[fsincox]

I believe so.
Frank

 

[KENAT]

I don't see why it couldn't be considered an FOS.

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[axym]

dtmbiz,

This is a ticklish one, particularly in Y14.5M-1994.

I would imagine that the purpose of this is to use the two opposing surfaces as a datum feature at some material condition. In '94 material conditions were only defined for features of size and to meet the '94 definition of a feature of size, the feature must be associated with a size dimension. So the profile tolerance alone won't get you there.

Could a size tolerance be applied to the parallel-plane feature and coexist with the profile tolerance? I don't see why not. The profile tolerance would just have to be a refinement of the size tolerance. If the size tolerance had the same bandwidth as the profile tolerance, then it would work because the size tolerance would allow more variation at LMC.

Jim, JP, others - opinions?

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[dtmbiz]

Yes Evan this is the exact phraze that got me to wonder if it is or isnt?

What is is?

Anyway what is "size"? A basic dimension is a theoretically exact value, where as, in a sense, it could be argued that a non-basic dimension without a tolerance is theoretical exact value. The traditional "size" dimesion would have a tolerance attached to it or default to a general tolerance, while the basic dim has its tolerance in a FCF.

Just thinking about possibilities.

 

[Belanger]

This is ticklish even in the 2009 standard, Evan! It drives me nuts, but the definition of a feature of size says that it must be "a directly toleranced feature" (para. 1.3.32).

I can't recall if this came up in another thread, but because of this definition, you can't just point to a hole or pin and call it a FOS. In order for it to be a FOS, it must be directly toleranced!

And that brings up the definition of "directly toleranced." If I give a distance of 20 mm but leave it to the title block for the tolerance, I would say that it is not "direct." Same for the OP where the distance is a basic dimension. Or do you guys think that the profile all around qualifies as "directly toleranced"?

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

 

[Belanger]

Sorry, one other comment...

Evan, I would say that you can't have a size tolerance be applied to the parallel-plane feature and coexist with the profile tolerance. That's because the standard requires profile to be applied to a "true" profile, which in turn mandates basic dims for the surface.

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

 

[dtmbiz]

Unfortunately we are stuck on the 1994 standard.

 

[MechNorth]

I don't see a problem using opposed surfaces as a datum feature ... note NOT a datum fos. As indicated above, the tolerance must be directly applied (i.e. applied to the dimension, not the feature). In this case, referenced at RMB (RFS) would get you the center-plane based on the two opposed surfaces, though which one would dominate without a reference to a higher datum precedence could become a point of contention. '94 does not specify that only FOS can be used to establish a datum center plane, though it doesn't give much guidance otherwise.

Using an irregular datum feature controlled by a profile control is now included in the standard ('09) as an irregular datum feature (See Fig. 4-3). You can also now reference a planar datum at MMC in the datum reference frame ... another topic for another day.

As for the question of whether a general linear tolerance block would make an undimensioned (non-basic) dimension a FOS, it would depend on whether or not the feature has directly opposed points; the tolerance is considered to be applied directly to the dimension and so that aspect is already met.

I believe that the missing link here is that the standard requires a directly toleranced "dimension", not a directly controlled feature which is what you must have when you use basic dimensions to describe the nominal geometries.

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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[axym]

JP,

Yes, you're right. We need basic dimensions for the true profile.

I hate to bring this up because I object to it, but there are examples in the standard in which profile is used along with a directly toleranced dimension. There's a figure where the line profile tolerance ends up acting like parallelism, I think. The usual location-controlling aspect of profile is magically overridden by the presence of the directly toleranced dimension. It's horrible.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[MechNorth]

Evan, can you give us some references? Tks

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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[Belanger]

Ah yes... Just as position used for orientation only is my pet peeve, Evan has this one. It was brought up in a thread many months ago and I think we all just kinda threw our hands up and said that the picture violates the letter of the law for profile.

It's Figs. 8-17 and 18 in the new std.

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

 

[axym]

Jim,

I don't have the standard in front of me, so I can't give any specific figure or page numbers. I think it's near the end of the profile section. The part has a flat bottom and a curved top, and there's a line profile tolerance. The directly toleranced dimension lets the height float within a larger range than the profile tolerance would otherwise allow.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[MechNorth]

Not sure it's really a conflict. The toleranced diameter is at one point only (rather useless in my opinion) while the profile controls the entire surface. It's grossly along the same lines as two single-segment position controls wherein the second is a refinement of location as well as other stuff.

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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[axym]

Jim,

Are you talking about the figure with the conical part? That isn't the one I was thinking of, but it has a similar problem. The profile tolerance would usually control the "size" of the cone, but the presence of the directly toleranced dimension somehow overrides that. The profile tolerance ends up acting like a total runout tolerance. This example is even more horrible than the other one I was thinking of.

The other example is right near the end of the profile section in 2009, I think.

JP,

I wish that this was my only pet peeve with Y14.5. It is a big one though. I also have the self-referencing datum feature, the "datum axis" established by hole pattern, the wildly inconsistent use of the term "datum plane", two "interpretations" for position tolerances, the list goes on ;^)

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[Belanger]

Evan,
You might be thinking of Fig. 8-27 of the new standard, with a flat top and flat bottom. But what you describe is actually OK, because profile is not controlling the location; it's only controlling orientation to the datums (and form, of course). IOW, FIg. 8-27 is exactly equivalent to using angularity with "each element" noted below it. Thus, no conflict with the size dim.

But I had thought you were referring to the examples of profile on a cone...

Jim,
The standard says that a profile tolerance is applied to a true profile. A true profile is defined as perfect, and given by basic dimensions (or CAD data). So are you saying that in Fig. 8-17 the "profile" is still true because profile is translated as "shape only"? (Yet the size is not "true.")

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

 

[dingy2]

I will put my 2 cents worth in this discussion.

Fig. 8-25 of the 2009 standard has an irregular feature of size with a profile tolerance controlling the form and a positional tolerance controlling the location. The complete irregular shape is a feature of size.

Could we separate 2 parallel surfaces in this example and call the enclosed space a feature of size? I think that 1.3.32.1 covers features of size of "set of two opposed parallel elements or opposed parallel surface, each associated with a directly toleranced dimension. See par. 2.2"

Par. 2.2 (a) Limit dimensioning
(b) Plus and minus tolerancing
(c) Geometrical tolerances directly applied to features.

To sum it up - I think that 2.2 (c) covers the question and a yes may be appropriate here.

Dave D.

http://www.qmsi.ca

 

[MechNorth]

Re Fig. 8-27, J-P is right that the profile of a line controls the orientation & form as a refinement of the location already established by the fos dimension. It's a hierarchy-thing; first thing, a feature must be located if it can be, then it can be refined for orientation & subsequently for form.

Fig 8-17 has a basic angle defining the conical surface, and a linearly-toleranced dimension controlling the size at the one end alone. From whatever acceptable size that diameter is produced at, a "perfect" cone of 15-deg included angle is established and the surface is then controlled based on the profile control. Leaving them as a basic dimensioned width though ...

Fig 8-18 is a little bit different, but similar. The toleranced diameter sets a zone within which the single cross-sectional diameter must lie. The basic 15-deg conical profile zone can float within that linearly toleranced zone at that specific location. I have seen this used (poorly) as a datum target line; they way it was done didn't work very nicely.

Dave...hmm. Definitely the entire outer path establishes an irregular feature of size. If you changed the 50.8 dimension from basic to linearly toleranced, then it is mostly a feature of size (the two sides are not 100% directly opposed as they are of different lengths); perhaps again an irregular feature of size? From 2.2, indeed if basic 50.8 dimension remains, then the profile control is now considered a directly applied tolerance ('09) and the width can indeed be classed as a feature of size. Tks for pointing that one out.

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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[axym]

JP,

Yes, it is Fig 8-27. I still don't like it. To me, profile controls size if it can and controls location if it can. This is one of the things that distinguishes profile tolerances from orientation tolerances and runout tolerances.

Perhaps what irritates me is the "oh by the way, here's how it works in this case" way in which these examples are presented in the standard. Here's the explanation of Figure 8-27 on page 176:

"Figure 8-27 illustrates a part with a profile of a line tolerance where size is controlled by a separate tolerance. Line elements of the surface along the profile must lie within the profile tolerance zone and within a size limiting zone. In this application, the datum references only orient the profile of a line tolerance."

So in this case the datum references only orient and we don't need basic dimensions to define the true profile? Where the heck did that come from? Now I have to hunt for a dimension to see if the profile tolerance controls location relative to the datum feature (they call it size)? Good grief. See, now you got me started on the "profile with size" thing.

I wish that Y14.5 would explicitly adopt underlying concepts, and stay consistent with them. Instead of implicitly adopting underlying concepts, and subverting them by including non-rigorous exceptions for specific cases.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[Belanger]

I've never seen a problem with profile floating within a ± tolerance if we only desire form and orientation. Just seeing that the location is not basic is enough to realize that profile doesn't cover that quality.

But this cone picture has me bothered -- how can a circle/cone/sphere have "true" shape but retain a toleranced size? Wouldn't you rap the knuckles of someone who attaches profile of a surface to a hole if there's a size of Ø10 ± 0.2 on the same hole?




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