Examples of Equivalent Dimensioning & Tolerancing Schemes

[pmarc]

Hi,

Throughout the years I have come to a conclusion that in general there are not many examples where changing dimensioning and tolerancing scheme from one to another would keep the geometric requirements for the system unchanged.

One example where this conclusion would not be true is changing from perpendicularity wrt A to total runout wrt A when applied to a flat face normal to datum axis A.

Another one would be a simple bushing where its ID and OD are controlled with the same +/- tolerance, and then it does not really matter which of the features will be datum feature A and which will be controlled with position or runout relative to A.

I have some more, but I would like to see what others can offer. So could anyone share some examples?

 

[pylfrm]

chez311,

For the scenarios with the RMB datum feature reference, I considered a simplified case where both datum features have a size tolerance of 20 +/- 0.1. I also assumed the datum axis would be derived as if the datum features were a single continuous feature.

With the simplification and assumption described above, I don't think #3 and #4 are equivalent. Consider an example where the two datum features are perfect cylinders of identical diameter and length, and are parallel but slightly offset from each other. The datum axis is derived from the smallest cylinder that circumscribes both datum features, which would involve four points of contact in this case. So far, the datum axis is coaxial with the smallest cylinder that circumscribes the individual feature axes. Now imagine the cylinder that establishes the datum axis is slightly shifted and reduced in diameter, and the material that protrudes beyond it is removed from the datum features. This shifts the datum axis without affecting the individual feature axes, and could result in a part that scheme #3 accepts but scheme #4 rejects.

I agree that #2 is equivalent to #1/#5. The same goes for #6 with the simplification and assumption described above.


pylfrm

 

[greenimi]

Pmarc and all,

Are those two schemes equivalent:

Consider a cylinder Ø50 ± 0.2.
Are there any differences between the mathematical definition of this cylinder in ISO GPS (as default, as shown) and ASME Y14.5 with explicit Independency symbol used?
ISO GPS: size versus ASME size and I

Also on the reverse: there are any differences between the above cylinder defined for size and having E symbol in ISO GPS and the same cylinder (same size) in ASME Y14.5?
ISO GPS size and E versus ASME size

I am interested in the nitty-gritty details of those schemes (if you can call them schemes in the tru true sense of the word)

 

[pylfrm]

greenimi,

Although it doesn't answer your question, you may find this article interesting:
Size tolerancing revisited: A basic notion and its evolution in standards


pylfrm

 

[greenimi]

Pmarc and all,
I would like to propose another equivalent dimension and tolerancing scheme. Please validate it and let me know if you agree or disagree with my statements:

Fig 8-25 from 2009 (page 177) has shown parallelism within 0.12 to A primary and B secondary for EACH ELEMENT in the right view. If this parallelism callout is to be replaced with profile of a line (not profile of a surface) within 0.12 to the same datum reference frame (A primary and B secondary), wouldn’t this new scheme (profile of a line) be equivalent with the original shown in the standard?

I am thinking this, because in the explanation of fig 8-27 states: “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.”


Also fig 8-25 has the following explanation : ” The surface must not only be within the profile tolerance, but each straight line element of the surface must also be parallel to the datum within the tolerance specified.”

So, in my mind, based on these two explanations, my proposed equivalency works. What do you think? How far in the weeds am I?

 

[chez311]

greenimi,

The profile of a line tolerance in 8-27 can float because it has a directly toleranced size dimension, the tolerance in 8-25 has basic dimensions but can float within the profile tolerance because it utilizes parallelism. If one were to replace the parallelism in 8-25 with profile of a line it would be fixed with basic dimensions to the specified DRF, right? Am I missing something? These don't seem equivalent..

And based on the previous discussions I've followed here, if you were to replace the basic dimensions in 8-25 with directly toleranced ones, I think you would have trouble developing a scheme that creates a valid "true profile" to utilize with a profile tolerance. Not saying it can't be done though.

 

[chez311]

That being said, I think if you were to replace the profile of a line callout in 8-27 with parallelism then I would agree those two are probably interchangeable.

 

[Belanger]

chez311 -- I will agree with greenimi on this one. Profile does not require that the location from the referenced datums be basic. Such was the case in Fig. 8-27. In Fig. 8-25, the distance from the datums is not directly tolerance with plus/minus, but there still is a location tolerance from profile of a surface. You mentioned this fact, so I'm not sure why you wouldn't extend the same courtesy to profile of a line.
It's akin to a surface which might have two FCFs stacked together: one showing profile of a surface with 0.4 to A and B and the other showing profile of a line with 0.12 to A and B. The top one controls location and the bottom one controls form and orientation.

edit: Just saw your last post. Still wondering why you think that parallelism (each element) can substitute for profile in 8-27 but that profile of a line can't substitute for parallelism each element in 8-25. Is it the plus/minus vs. basic thing for the distance from the datums?

 

[chez311]

one showing profile of a surface with 0.4 to A and B and the other showing profile of a line with 0.12 to A and B. The top one controls location and the bottom one controls form and orientation.

JP,

Could you perhaps explain to me how a FCF with profile of a line to 0.12 [A|B] with basic dimensions in the view shown with parallelism in 8-27 8-25* would not be constrained in location to A and allowed to float within the larger profile of a surface? Maybe 0.12 or datumless profile of a line, but that wouldn't really be equivalent would it?

Is it the plus/minus vs. basic thing for the distance from the datums?

Yes, that is the issue that I have. As I said, I may be missing something and need someone to point it out to me, but I just don't see it.

*Edit sorry! Wrong figure

 

[chez311]

JP - I edited my posted. I referenced 8-27 when I meant to say 8-25, critical mistake my bad.

 

[Belanger]

Profile is not automatically required to be locked in at a basic distance from any datum. Profile's main job is form; it must always control that (unless you use circled T, I suppose). In addition to its main job of form, a profile tolerance may also reference one or more datums. If a datum is referenced, then profile takes on the additional job of orientation. But it could still move up or down within a larger location tolerance. I think that's a fair statement about Fig. 8-27.

In Fig. 8-25 the profile of a surface tolerance of 0.4 is certainly a fixed, locational tolerance. But I would read the profile of a line tolerance (substituted for parallelism) as controlling form and orientation, since assuming it to be a location control would make the profile of a surface tolerance meaningless.

Think of it as you would a composite profile tolerance: The upper and lower callouts may reference the same datums, but the lower one is a refinement that controls orientation from the datums. (I know this is not strictly composite, but I'm trying to think of an analogy.)

 

[chez311]

Profile is not automatically required to be locked in at a basic distance from any datum.

Of course not automatically, but if it is located with basic dimensions to a referenced datum feature, would it not be fixed in location? If not, why not? I agree with your statement in reference to 8-27 as it includes directly toleranced dimensions.

But I would read the profile of a line tolerance (substituted for parallelism) as controlling form and orientation, since assuming it to be a location control would make the profile of a surface tolerance meaningless.

Without a note saying otherwise, I don't think I would be assuming anything. There are basic dimensions to [A] with [A] in the DRF - this tells me there is a location constraint. I agree that would make the profile of a surface somewhat redundant (somewhat because I know that controlling a surface with line elements can have unexpected consequences - thought maybe not so much with a complex surface with profile controls) and instead of assuming that this means that there is no location constraint, I would probably point out to the designer that they may have created a somewhat redundant control.

Think of it this way, if it were profile of a surface instead of a line would it not constrain location to [A]? Why would profile of a line be any different?

FYI the guys over at tec-ease seem to agree with me.

https://www.tec-ease.com/gdt-tips-v...3R-p44sQIHbuUoVglUWjG_DfVUgNcrb8j-K-RtUv7qbi8

Pulled directly from the tec-ease example:
"On the second drawing datum references have been added to the profile of a line tolerance. The result is that the profile of a line tolerance overrides the profile of a surface tolerance. Since every line in the surface must be located with respect to the datums, the entire surface is being controlled for location by the profile of a line tolerance."

 

[pmarc]

For what it's worth, I agree with chez311 on fig. 8-25. Changing parallelism to profile of a line will not create equivalent scheme, because in the profile of a line scenario 0.12 wide tolerance zone will be located at basic dimensions (fixed relative to A|B), which is not the case in the parallelism scenario. In fact, such profile of a line would override profile of a surface.

However, I don't agree with:

That being said, I think if you were to replace the profile of a line callout in 8-27 with parallelism then I would agree those two are probably interchangeable.

Profile of a line is 2D control, parallelism is 3D control, and so parallelism would impose more stringent requirement. For example, imagine that the top surface, as seen in the left view, has a valley of 0.4. With that, profile of a line error could still be very small (even zero), while for parallelism the actual error would be 0.4.

 

[chez311]

pmarc,

My apologies - I agree that a 3D parallelism control would not be equivalent to a 2D profile of a line. Since we were discussing surface elements, I glossed over and neglected to mention that a similar callout to 8-25 would have to be applied to 8-27 noting "EACH ELEMENT" along with parallelism to be equivalent. This is what I meant in my initial statement, I should have been more thorough. Would you agree with that statement, modified as I have noted?

 

[pmarc]

I agree that with the additional note the schemes would be equivalent.

 

[greenimi]

So, I got the things backwards. (parallelism callout from 8-25 with EACH ELEMENT replacing profile of a line makes an equivalent scheme and not what I wrote in my original post, which basically is in reverse)
Maybe the things seen from China are truly up-side down. Well, hopefully only temporary.

Thank you chez311, pmarc and Belanger for your contribution for my own education. I appreciate your continuous help.

 

[Belanger]

Sounds good -- I too agree now with the consensus that's been arrived at!

 

[aniiben]

Pmarc, Chez311, Belanger:

6.4.3 states :"Although orientation tolerances are only constrained in rotational degrees of freedom
relative to the referenced datums, the notation of EACH RADIAL ELEMENT adds a requirement for the tolerance zone(s) to be constrained in location relative to the axis from which the radial elements emanate. Tolerances for individual elements may also be specified using a line
profile tolerance."

Are you thinking that having EACH ELEMENT instead of EACH RADIAL ELEMENT a reasonable extension of principle? Otherwise why 8-27 with parallelism (and EACH ELEMENT note) instead of the original profile of a line, would be an equivalent dimensioning schemes?

 

[chez311]

aniiben,

I mean, its not much of an extension of principle when its clearly stated as a supported note earlier in 6.4.3 as well as explicitly shown exactly how to utilize it and what the tolerance zone looks like in examples like 8-25.

Where it is a requirement to control only individual line elements of a surface, a qualifying notation, such as EACH ELEMENT or EACH RADIAL ELEMENT, is added to the drawing.

 

[greenimi]

Pmarc, chez311, Belanger, aniiben and all,

When I posted my equivalent scheme (which has been discussed that is not quite equivalent) I was under the impression that if EACH ELEMENT note (in its essence or by its definition) bring the location control aspect with it by default. So, in fig 8-25/2009 since EACH ELEMENT is already shown, the location (aspect) is there making it a refinement of the overall profile of 0.4. By replacing the parallelism with profile of a line the location aspect is kept unchanged.
Not sure where is the disconnect.


I also would like to bring another potentially equivalent scheme: fig 4-4/2009 if existing control for datum feature C is replaced by composite profile? Are those equivalents? Basically, remove parallelism and flatness symbols and the applicable two small horizontal lines (and bring the profile symbol in the middle of the first column)

 

[pmarc]

In fig. 8-25 EACH ELEMENT doesn't add any location control to the callout. It is there because without it the parallelism callout wouldn't really make sense (it couldn't be applied to a non-planar surface).

The intent of the parallelism callout with EACH ELEMENT is to refine orientation of elements seen as lines in the right view. Profile of a line would do more than that - in addition to tightening orientation to A|B, it would control location of the linear elements relative to datum A.

As for the idea of modification of fig. 4-4, I would agree that these would be equivalent schemes (although if this was real example, I would definitely recommend using the original scheme as much more straighforward).