Limits of perpendicularity related to thickness of part

[310toumad]

I looked around and couldn't seem to find anything in the Y14.5 standard that gives any sort of guideline on applying perpendicularity to very thin features. I would think at some point it would become non-sensical to do so because it would become impractical to try and measure it. But how thin is too thin? If I have a cylinder that is 1/16" in height, or a hole in thin sheet metal there is nothing in the rules that states I cannot apply perpendicularity to these features, but it just seems like it would be bad practice.

 

[chez311]

310toumad,

I'm sure there are practical limits (I don't know of exact figures but I can envision points at which perpendicularity doesn't make sense in the bulk of applications) however with the establishment of whats "practical" there also comes very highly specific applications for which the normal rules (or rather - rules of thumb) do not apply. For example - the ratio of the relative height/width as well as to the rest of the part is important (1/16" high boss thats 1/16" in diameter and/or is on a part which is 1/8" x 1/8" x 1/8").

Regardless the standard doesn't typically define what is NOT allowed, as the committee cannot envision all the possible applications of the different controls. It is up to the designer to decide whether the applied control is necessary for part fitment/function, whether it is measurable, and to work with quality/production to determine if the process is capable to achieve the tolerances applied.

 

[Belanger]

Also keep in mind that we must have datum features toleranced among themselves (some people call it "qualifying" the datums). So even on a paper-thin part, if you want datum A derived from the main surface and datum B derived from a hole in that thin part, then you still must put a perpendicularity control on datum feature B. This is stated in paragraph 7.9 of ASME Y14.5-2018:
"To make it possible to calculate the true geometric counterpart boundaries of each datum feature in a datum reference frame, a relationship between the datum features shall be specified." (Emphasis on "shall.")

I know it doesn't seem important from a practical point of view, but without perpendicularity there is incomplete definition of the datum, which will goof up other downstream tolerances tied to that datum.

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

 

[310toumad]

Belanger, do you know where I could find a similar statement in the 2009 standard? That is the one I have access to, thanks.

 

[greenimi]

4.9 /2009 states:
"Datum features shall be controlled directly by applying appropriate geometric tolerances or indirectly by dimensions such as the size of a primary datum feature of size. This in turn makes it possible to calculate the datum feature simulator boundaries of each datum feature in a datum reference frame."

The paragraph mentioned by J-P has been added in 2018 (true geometric counterpart not being a defined term in 2009)

Now, I've seen some brown bag sessions and videos from late prof Don Day and he was asked about the very same condition (adding perpendicularity on a very thin parts) and his answer has been something like "some common sense" does not hurt. Enough data cannot be gathered to validate the perpendicularity.

"Datum Feature Selection" video

https://www.tec-ease.com/premium/login/brownbag-sesions.php

 

[chez311]

Additionally there is the following in 4.9 of Y14.5-2009:

*The relationships between datum features
to be considered are the
(a) form of the primary datum feature(s) (see Figs. 4-2
and 4-5) and/or the location between features in a pattern
used to establish the primary datum. See Figs. 4-24
and 4-25.
(b) secondary datum features’ orientation and/or
location as applicable, to higher precedence datums. See
Figs. 4-2, 4-5, 4-26, and 4-30.
(c) tertiary datum features’ orientation and/or location
to higher precedence datums as applicable. See
Figs. 4-2 and 4-5.

However this does not have as strong a wording as the preceding statement in the 2018 version as pointed out by JP.

Enough data cannot be gathered to validate the perpendicularity.

This is I think the sticking point for me too, I understand a secondary datum needs to be oriented relative to the primary, but if its not measurable - or even if it is but theres not much value in actually measuring it, one could get into a little bit of trouble.

JP - what would your solution in this case be? Would it be to put an arbitrarily large perpendicular tolerance on the feature so that anyone looking at the print (ie: a supplier quoting a job) knows that it does not need to be directly inspected as it would be well within any reasonable process limits or maybe actually putting such a note ("NOT REQUIRED FOR PART ACCEPTANCE") on the perpendicularity tolerance?

*Edit: added to the quote from 4.9 in Y14.5-2009 as it didn't seem to make sense without context

 

[Belanger]

Ideally the dilemma could be solved by using a functional gage -- no real number needs to be spit out; the functional gage post would simulate the datum as intended by the function, even if the part itself is paper-thin. Thus the beauty of MMC...

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

 

[aniiben]

Or add zero at MMC on the "pretend" - theory driven perpendicularity

 

[310toumad]

Sorry, I guess I don't understand, but wouldn't zero perpendicularity at MMC be adding an even more stringent requirement (at least for part of the tolerance band)?

 

[Belanger]

Not really. First, it depends on the thickness of the part. If it's paper-thin, then the perpendicularity is not necessarily serving the function of the part directly; it's more about the interfacing gage (see my first comment above). But also realize that a zero tolerance at MMC is not tightening the overall envelope (the "virtual condition") because having a zero often allows the size of the hole itself to have a larger spread on the diameter tolerance.

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

 

[greenimi]

Let me ask a quick question here (reference fig 7-4/2009, or 10-4/2018 )

If perpendicularity callout (for example .005 at MMC) is added for “B” within some tolerance to “A” and
a perpendicularity callout (for example .010 at MMC) is added for “C” within some tolerance to “A” primary and “B” secondary
then what would be the maximum perpendicularity error between the SIDES of the shown plate to datum feature A? Is it limited to the perpendicularity error of the middle plane (.005 or .010) ?

The main question: is how do I relate the perpendicularity of the middle plane with the orientation error of the sides?
What is controlling the orientation error to the primary datum feature A?

 

[axym]

greenimi,

I think that the worst case perpendicularity error for the sides is the same as the worst case for the center plane. So for the B sides it would be 0.55. This is after giving it 1 minute of thought.

Your question also brings up an interesting point, that the MMB's for B and C in that figure cannot be calculated (because the figure does not include perpendicularity tolerances for B and C).

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[greenimi]

Thank you Evan,

Evan and all,
May I ask one additional question:

Is the orientation relationship between the side surfaces (“parallelogram shape-wise” in the lower picture, 7-4 figure/ 2009) -their mutual orientation—influenced by this perpendicularity of the middle plane or the out of square-ness allowed is only driven by the width sizes. Again, I am not talking about the parallelogram (of the lack thereof) in the cross section (upper figure), but in the lower one.

Thank you

If it's paper-thin, then the perpendicularity is not necessarily serving the function of the part directly; it's more about the interfacing gage

J-P,
Isn’t your statement the perpetual dispute: the product definition functional driven or inspection driven? If no serving the function then why perpendicularity of thin feature is placed on the drawing, in the first place?

 

[axym]

greenimi,

The width tolerances on B and C would not control the relationship between the side surfaces of B and C at all. B and C could be extremely parallelogram'd and the width tolerances could still conform.

If there was a perpendicularity tolerance on C that referenced A and B, then this would indirectly control the parallelogram effect to some degree. But even if this perpendicularity was perfect, there could still be a significant squareness error between the B and C sides (because the width tolerances allow B and C to be significantly tapered).

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[Belanger]

greenimi -- I didn't mean to make it sound like inspection/gaging should be the driver. But it's about completeness of the drawing, and getting the part locked into a coordinate system.

Function is the main driver in GD&T. But the GD&T is only as good as the coordinate system (DRF) that it's tied to. Having perpendicularity on a thin part might not seem functional but it completes the definition of the DRF.

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

 

[greenimi]

But even if this perpendicularity was perfect, there could still be a significant squareness error between the B and C sides (because the width tolerances allow B and C to be significantly tapered).

Evan and all,

That is interesting (at least for me).

So, if you compare now (2009 standard pictures):

7-3 figure (assuming a perpendicularity is shown on datum feature B to datum “A” AND also a perpendicularity is shown on datum feature C to datum “A” primary and “B” secondary) I would say that the squareness error is directly controlled which is not the case in fig 7-4 (with same assumptions that some perpendicularities –of the middle plane in this case of fig 7-4—are added).
Am I correct?

My current understanding is that perpendicularity of the middle plane (fig 7-4) does nothing to controlle the parallelogram’d shape of 7-4.
Again: just to reinforce a bit:

On both figures 7-3 and 7-4 some perpendicularities are to be added (just for the sake of this discussion to make the drawing “completed)

On 7-3:
- on datum feature B: perpendicularity .xxx tol to A primary
- on datum feature C: perpendicularity .xxx tol to A primary and B secondary

On 7-4:
- on datum feature B (middle plane): perpendicularity .xxx tol to A primary
- on datum feature C (middle plane): perpendicularity .xxx tol to A primary and B secondary

Should I understand that is a significant difference between the squareness allowed (parallelogram effect) between these two schemes?