0 @ MMC vs. 0 @ MMB 2

[weavedreamer]

 

[3DDave]

mkcski - It's fine if you read the entire callout into words, to wit:

Does the axis of this feature align perfectly with the axis of the referenced datum feature that is allowed to shift within the tolerances allowed on that datum feature axis?

The exact same condition exists when a 0 position tolerance at MMC is verified when the feature is at MMC.

If the second callout is legal, then the first is also.

All that is necessary for a part to be accepted is for an inspector to find a case where the axes are aligned.

 

[Belanger]

While I agree with mkcski that 0 RFS with datum MMB is not legal, perhaps it could be thought of this way: The feature toleranced as 0 actually creates the datum, and the item identified as datum feature A will be the one either accepted or rejected.
In other words, the datum is supposed to be the "driving" factor, but in this case the datum is "driven," simply because the feature that is being tagged with position is stated as being perfect, while the datum feature is the one feeling the variation.
I wouldn't say that this is an extension of principles, but a misuse of the principles. (Makes for a good academic discussion, though.)

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

 

[3DDave]

JP - this is the same case for all RFS feature references to datums with MMC/MMB conditions. The Datum simulator will always be backdriven by the RFS simulator. It's just in some cases there is an additional flexibility when the RFS reference is not '0'.

Tolerance simulation software, such as VSA will have no trouble with this.

Too many are looking at their cookbooks without understanding the math.

Would it be much better if it was .000000000001 instead of 0?

 

[Belanger]

So I need a straight-up answer from you Dave
Is that right-hand picture acceptable per Y14.5, or not?


John-Paul Belanger

 

[3DDave]

Already said they were both legitimate. I was still considering at the time that there was some other difference but, after more consideration, I think they are identical.

Is there something in the standard or the interpretation that indicates otherwise?

 

[pylfrm]

I'll say acceptable as well. Shame on me for being noncommittal earlier. This time I have some supporting evidence:

ASME Y14.5.1M-2004 provides two interpretations for positional tolerancing. Para. 5.2.1 defines the surface interpretation, and para. 5.2.2 defines the resolved geometry interpretation. Para. 5.1.1 states that the surface interpretation shall take precedence if the results are not equivalent.

For both interpretations, the tolerance specification is defined in such a way that a zero tolerance value applied on an RFS basis does not create a problem, and still allows conforming parts to exist.

Additionally, the actual value of position deviation is defined as the smallest tolerance value to which the feature conforms.

For any part that meets the requirements of OP's left-hand picture, can we agree that the actual value for the position tolerance in the right-hand picture is zero? If not, then what else might it be?


pylfrm

 

[Belanger]

It's true that nothing in the standard says that you cannot use zero RFS, but there's still something weird about it. However, I agree that the net effect is the same so I'll accept it at that.
pylfrm -- for your last question, I would say that you don't even need the qualifier that forms the first part of your sentence. This is because the actual value for the position tolerance in the RH picture is always zero for an acceptable part, simply because that's what the callout prescribes.

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

 

[weavedreamer]

With a little more assistance, I was able to get clarified where the weirdness lay.

In Paragraph 7.3.4 is found "the positional tolerance allowed is totally dependent on the unrelated actual mating envelope size of the considered feature, as explained in para. 2.8.3." This is reiterated in the referenced paragraph 2.8.3 also.

When I earlier stated:

Even though the shift acts like the bonus in this singular relationship, they are different ways of regarding the relationships between the features in question.​

The difference is: the positional tolerance comes from the unrelated actual mating envelope of the feature of size, not the the datum feature of size.

 

[mkcski]

I like your analysis of the difference in affect of size change between the feature vs the datum. Well expressed! The difference is "hard" to see when the datum and feature are both F.O.S. and have a coaxial relationship.

Certified Sr. GD&T Professional

 

[Belanger]

Weaverdreamer, I'm not sure I see that as being any different. Yes, the position tolerance is looking at the unrelated actual mating envelope, but the datum is also derived from the unrelated actual mating envelope.

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

 

[weavedreamer]

Belanger,

As each dimension is supposed to have a tolerance, direct or indirect, the 0@MMC gets the tolerance zone from the unrelated actual mating envelope size of the feature of size.

Specifying it as 0@RFS, effectively leaves it without a tolerance zone. This is the violation of the standard.

Adding the datum feature of size in acted as a diversion.

 

[Belanger]

I mentioned in my first post that it was a violation, but subsequent comments swayed things to make me agree that they are effectively the same.
FYI, all of your initial pictures involve unrelated actual mating envelopes (UAME). The only difference is that MMC creates a constant-size UAME and RFS creates a variable-size UAME. But does that really make the two pictures different?

So allow me to rephrase your original post: An UAME has to be found for Zone A. And an UAME has to be found for Zone C.
The left-hand picture in your OP asks us to compare the axis of Zone A's constant-size UAME to the axis of Zone C's variable-size UAME.
Then, the right-hand picture asks us to compare the axis of Zone C's variable-size UAME with the axis of Zone A's constant-size UAME.
Notice how the words come out the same!

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

 

[mkcski]

I agree with John-Paul's conceptual statements. However....realistically is makes no sense to me to have the position tolerance on the Zone C feature in the right-hand sketch totally dependent on the size change of the Datum B. This just seems backwards and contrary to design practice using GD&T.

Additionally, the short length of the datum feature Zone C in the left-hand view compared to the feature length Zone A is questionable for creating a repeatable datum simulation.

Certified Sr. GD&T Professional

 

[3DDave]

The only requirement for RFS is that one orientation can be found that meets the requirements. In practice there is the potential for multiple, equally valid solutions.

 

[pylfrm]

Specifying it as 0@RFS, effectively leaves it without a tolerance zone. This is the violation of the standard.

The tolerance zone does not disappear, it collapses to a line. If the feature axis is coincident with the "tolerance line", then the requirement is met. At least this is how it's described in Y14.5.1-1994.


Side note: In my previous post, I mistakenly wrote Y14.5.1-2004 (which does not exist). That should have been Y14.5.1-1994 (which was last reaffirmed in 2012).


pylfrm

 

[weavedreamer]

The tolerance zone does not disappear, it collapses to a line. If the feature axis is coincident with the "tolerance line", then the requirement is met. At least this is how it's described in Y14.5.1-1994.

My wording didn't convey this well. Rephrase as:

Specifying it as 0@RFS effectvely leaves the zone for the feature axis without a tolerance. This is the violation of the standard, per my ongoing struggle to clarify and articulate my understand of it here. The tolerance zone for the feature axis is supposed to come from its unrelated actual mating envelope (paragraph 7.3.4). Invoking MMB makes it appear to be solved via the related actual mating envelope.

In essence, this concurs with Mr. Belanger's first reply:

I'd say it's a violation, since the hole has zero tolerance, regardless of its size. The datum's shift just says that you might be able to jiggle the part around, but the hole's axis must still be perfectly aligned.

 

[Belanger]

Specifying it as 0@RFS effectively leaves the zone for the feature axis without a tolerance.

No -- it only leaves the zone without a tolerance with respect to the datum axis derived from a MMB datum feature.

In reality, the actual datum feature might deviate (most likely it will deviate) from the MMB. Thus, the zero tolerance is following a moving axis -- which really isn't a zero tolerance anymore.

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

 

[mkcski]

My head is spinning. In the real world, who would specify position like this when even those in-the-know cannot agree on an (the) interpretation. Please see Y4.5 2009, paragraph 1.4 (d)

Certified Sr. GD&T Professional

 

[pmarc]

This discussion is really interesting, but could we go back to the beginning of it for a moment?

My understanding of weavedreamer's second post is that the idea of changing datum feature from zone C to zone A was a result of concern about repeatability of datum axis B derived from such a short portion of the cylinder in zone C. So what if the repeatability problem is gone? In other words, perhaps in reality the cylinder in zone C works as a secondary datum feature, not primary? Any chance to learn more about part mating relationships with other components in assembly?

 

[weavedreamer]