Reporting A Unilateral Surface Profile Tolerance

[DJ Nelson]

We have a surface with a profile of .008 U .000 to ABC.
we measured the surface with a scanner after aligning to the Datums.
We ended up with a MIN of .0072 and a max of .0121 the software calculated the measured Profile as .0322
I am not sure if I do not understand how this is calculated and reported but it seems questionable see Attachment
If correct or incorrect were can I find were it is that explains how to record this.

 

[Burunduk]

It makes sense. Since there is a deviation outside the tolerance zone of .0121, your measured value will be the sum of the specified tolerance and twice the deviation outside. That gives: .008+2×.0121=.0322.

 

[pmarc]

DJ Nelson,

You may want to take a look into ASME Y14.45-2021 standard for measurement data reporting. In the profile section, you should find the answer you are looking for.

 

[3DDave]

It makes sense to double the excess/shortfall for a bilateral tolerance zone which it seems to have done for the second case.

There is no Unilateral tolerance zone, no matter how large, that can encompass the negative deviation, which the bilateral calculation is supposed to represent.

This appears to be the software writers finding themselves unable to find a way to report this and making something up.

 

[pmarc]

The .0322 measured value is in line with the approach shown in Y14.45, which basically follows the method of obtaining the actual value for profile tolerances defined in Y14.5.1-2019.

 

[3DDave]

If the zone in the FCF is changed to [0.0322(U)0.000] the -0.0072 will not be in that zone.

I don't see how any computation can cover the negative value when the lower limit is guaranteed to be zero.

 

[pmarc]

To begin with, we don't know if the -.0072 is in the tolerance zone in the OP case, because we don't know what max dev and min dev really mean.

But since we know that .0122 is surely outside the tolerance zone and that .0122 is greater than |-.0072|, -.0072 cannot be used to calculate the measured profile value.

 

[3DDave]

You are right - the (U) notation has been endlessly confusing. I voted against it, but the vote at the committee meeting was a pre-arranged deal. +/- values for the limits would have been perfect, but the goal isn't clarity, it's some other need.

In any case, there is no zone that is suitable to encompass that offset. The example shows how an acceptable part is accepted, but does not show how an unacceptable part is rejected. There should be no value shown if changing the base tolerance cannot accept the deviation.

 

[3DDave]

I do get it - if the user doesn't understand the reporting it is up to them to specify what they want. It is not the reponsibility of the Y14.45 committee to make understandable or directly applicable computations.

In the public review draft example - changing the 0.4 tolerance to 0.32 would not allow the part to be accepted because now the -0.06 amount would be outside the tolerance zone that has only 0.02 left into the material from the original 0.1 zone. One would have to change both the zone width and amount of unequal distribution and clearly the Y14.45 group didn't think that was necessary to mention.

 

[greenimi]

But since we know that .0122 is surely outside the tolerance zone and that .0122 is greater than |-.0072|, -.0072 cannot be used to calculate the measured profile value

So in this case why the reported profile is NOT 2 X 0.122? I am sure I don't know the reported values per Y14.45, but I am trying to improve my knowledge in this area.

 

[DJ Nelson]

Thank You Burunduk that makes a scene were is there a place in one of the standards that defines that calculation?

 

[pmarc]

I am not exactly sure what you mean by 'offset', but I would again recommend grabbing Y14.5.1-2019 or Y14.45-2021 to see how this all works for any type of uniform profile tolerancing (symmetrical or nonsymmetrical) according to the latest standards. I agree the methodology is not the most straightforward, but is understandable.

Back to your [0.0322(U)0.000] example, again -.0072 will not play a role in the calculation of the measured profile value, regardless if it represents the distance from true/nominal contour or the distance from the profile tolerance zone LMB. That's because the .0121 is still greater than the absolute value of -.0072.

 

[pmarc]

greenimi,

It is not 2 X .0121 (.0122 was my typo) because the formula for the measured profile value is T = t + 2g, where t is the specified profile value (.008 in this case) and g is the worst-case distance of the measured profile from one of the tolerance zone boundaries (.0121 in this case).

Imagine that .0121 was for example .003 instead. You would know by looking at the measurement data that the measured feature is over max, but multiplying .003 by 2 would give you .006, which would be less than .008 and so you could incorrectly conclude that the feature is conforming.

 

[3DDave]

Offset - as in, not symmetric. As in "The tolerance zone has an offset."

Leave off the -0.0072. The part fails on the +0.0121

Try my notation for this [Profile of Surface| +0.00/-0.008|...]

So now there are two values, same as before

+0.0121 <- clearly out of tolerance
-0.0072 < clearly in tolerance.


No zone calculated for the offset / not bilaterally equal tolerance zone can encompass the clearly out of tolerance value.

But inspectors need a simple calculation that they don't think about and doesn't give useful information to anyone, so Y14.45 group used a simple calculation they didn't think about that doesn't give useful information.

 

[DJ Nelson]

Will It be the same calculation for a Bilateral Tolerance?

 

[pmarc]

3DDave,
I don't disagree.

But we are where we are and it doesn't look like the notation defined in Y14.5 is going to change any time soon.

 

[pmarc]

DJ Nelson,
The formula is the same for a bilateral tolerance, both symmetrical and unsymmetrical.

 

[DJ Nelson]

Thank You everybody you all cleared it up for us.

 

[Burunduk]

It is not 2 X .0121 (.0122 was my typo) because the formula for the measured profile value is T = t + 2g, where t is the specified profile value (.008 in this case) and g is the worst-case distance of the measured profile from one of the tolerance zone boundaries (.0121 in this case).

Worth noting for the OP that the sign of g becomes negative when the feature is entirely inside the tolerance zone.

 

[pmarc]

Burunduk,
Yes, that is true.