Pattern as a Datum; dimension origin

[dtmbiz]

(Pls see attached drawing)

Balloon 1:
Datum B is the pattern of 4 bosses
(delineated in this way for illustration only)

Balloon 2:
Dimensions originate from Datum pattern axis

Balloon 3:
Dimensions originate from “features” of pattern Datum B

Balloon 4:
Dimension is interpreted as equal distant from the
axis of Datum B; or is “half dimension” required
as shown 29.00 BSC in Balloon 2 area?

Question 1:
In our company there are some who say dimensions relating to a Datum pattern
originate at the axis of the pattern. Others say dimensions originate from the features
of the pattern. I don’t see any example figures in the standard (1994) that make this clear.
Where should basic dimensions originate from in relationship to the referenced pattern as a Datum;
Pattern axis, or feature ?

Question 2:
Are dimensions that are related to an axis of pattern Datum assumed to be equidistant?
Or are half dimensions required?

“4.2.2.1 Mutually Perpendicular Planes. The
planes of the datum reference frame are simulated
in a mutually perpendicular relationship to provide
direction as well as the origin for related dimensions
and measurement.”


I believe the dimensions should originate from (or "trace back to") the axis of Datum B in this case, because it is part of the DRF.
I also believe that dimensions should not be “assumed” to be equidistant and should have
“half” dimensions that relate back to the DRF.

Comments?

 

[dtmbiz]

Posted the wrong drawing again! Hopefully this is the correct one.....

 

[Belanger]

Question 1: The basic dims should be tied back to the actual features, not the axis of the pattern. (You'd also want to relate those bosses to each other with a position tolerance, and most likely then reference that datum B with the "M" modifier when positioning the other items back to B.)

The reason I say not to dim to the axis: think of another four-boss pattern that is not a nice rectangular pattern; maybe they are scattered at different locations. We can still call them datum B, but it would be nearly impossible to dimension to the "centroid" of the pattern.

I don't have the standard handy today, so I'll have to look later for an exact statement to back me up...

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

 

[pmarc]

As long as all dimensions that define spacing within the pattern and its orientation and/or location relative to higher precedence datums are basic, there is no difference where would be the origin from which other features are located. Basic dimension values are non cummulative.

I would also try to use pattern center as an origin whenever possible (just for clarity), but as J-P said if irregular patterns are considered it could be really hard to find centroids of those.

Unfortunately, as you noticed, Y14.5M-1994 gives very few examples of defining pattern of features as a datum feature so it would be quite hard to find proper example.

 

[dtmbiz]

The drawing isnt near complete, just trying to emphasize the areas of question.

John-Paul,
What about the DRF origin. The center of the pattern is the axis with 2 planes intersecting at 90 deg.,
right? Wouldn’t the center planes be relatively easy to find by touching of the virtual pin locations?

Are you saying that you would measure from "a" boss and then add a "half the distance between bosses to get back to the DRF origin?

Pmarc,
I do agree that this has been my way of thinking also.

Since a DRF should be the result of datums selected because of part function and interface; my view point is that if the dimensions come from the pattern axis, then it would be apparent that the part features are being centered in this case. Vs. selecting edges for example that serve little function relative to the mating parts.


Your comments are appreciated

 

[axym]

I agree with what JP and pmarc said, but I would also go even further. For the purposes of evaluating the conformance of the feature to the geometric tolerance and calculating an actual value, it doesn't matter where the basic dimensions originate from. In fact, they don't have to originate from anywhere - there doesn't have to be a well-defined "origin" for the basic dimensions. It's all relative - we just need to be able to calculate the locations of the considered features relative to each other, and their locations relative to the datum features. Any layout will do, as long as these relationships are indicated or can be calculated. Some layouts are more ergonomic than others, but it doesn't matter as far as the geometric tolerance is concerned.

The same thing goes for the DRF origin. It doesn't really matter where it goes, to me it's purely arbitrary. I realize that this doesn't completely agree with what Y14.5 says, particularly in '94. Some of my biggest pet peeves with Y14.5 are in this section, particularly the concept of the "axis of a pattern". This is a flawed concept that doesn't work (IMHO) and has led to a lot of confusion.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[Belanger]

Well said, Evan. This notion of three mutually perpendicular planes is true, but for the OP's example, that 3-plane system can be superimposed over any of the four bosses. As long as we know the basic dims tying them together, it doesn't matter. So that's why I discourage bringing another item into the discussion (an imaginary axis) that itself would have to be generated from the four bosses anyhow.



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