Datum Feature Simulator - definition Vs. examples

[Burunduk]

ASME Y14.5-2018:

"3.18 DATUM FEATURE SIMULATOR
datum feature simulator: the physical boundary used to establish a simulated datum from a specified datum feature.
NOTE: For example, a gage, a fixture element, and digital data (such as machine tables, surface plates, a mandrel, or a mathematical simulation) are not true planes, but are of sufficient quality that the planes derived from them are used to establish simulated datums. Datum feature simulators are used as the physical embodiment of the true geometric counterparts during manufacturing and inspection."

Can anybody clarify how digital data/mathematical simulation can be an example for a "physical boundary" and how can they be considered a "physical embodiment" of a true geometric counterpart?

Thanks.

 

[3DDave]

I'm more concerned that a machine table is considered digital data - though that really is just a reflection of how poorly the sentence is constructed.

For example, a gage or a fixture element (such as machine tables, surface plates, a mandrel,) and digital data (such as a mathematical simulation) ...

is the correct way to form that sentence.


As a practical matter all mathematical simulations within inspection equipment are imprecise. There is a limit to the precision of the values used and the calculations performed on them. Technically they are represented by the movement of electrons within solid materials - stretching the "physical embodiment" claim, but it's not purely conceptual. Actual transistors are taking on actual states and represent conditions that don't perfectly duplicate the supposed theoretical datums.

 

[axym]

Hi All,

I agree with 3DDave that it's the wording that is the issue here. By "physical" I believe they mean "established using an imperfect solid object or derived from an imperfect measurement". The concept is that the datum feature simulator is an approximation of the true geometric counterpart, which is theoretical and would be established using a perfect solid object or derived from a perfect measurement.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[Burunduk]

3DDave and Evan,
I think that the most common application of "digital data/mathematical simulation" in the context of datum simulation is operations performed within the inspection program on a virtual representation of the part generated by the measurement process - such as a cloud of points in an STL file. It could even be a result of a non-contact scan. Surely imperfect and imprecise, and not only conceptual, but as far from "physical" as any measurement technique can be, isn't it?

According to my understanding, a "true geometric counterpart" is a purely theoretical concept - it is something that can be attempted but can never be achieved. It's something to base theoretical principles on when not considering imperfections of the inspection equipment or process. Then, limiting the "simulator" within the definition to being physical seems questionable.

 

[3DDave]

Photons are physical enough; I keep seeing really attractive deals on photon wood cutters.

I'm trying to figure out what reasoning has led you to believe the train has left the tracks, seriously.

The hierarchy should be:

1) Theoretical three datum planes
2) (Theoretical) True Geometric Counterpart that is aligned to the three planes.
3) (Not theoretical) Datum feature simulator that is the imperfect not-theoretical best effort approximation as an approximation of the True Geometric Counterpart
4) (Part) Datum feature that is aligned to the applicable Datum feature simulator

Note that step 3 can either use best-effort gauges or best-effort measurements.

I preferred the previous term - "theoretical datum feature simulator" as it bridged the gap from datum feature to theoretical datum. True geometric counterpart doesn't help cement that place in the hierarchy.

From the foreword: "The use of the “true geometric counterpart” term is limited to datums."

That clarification would not be required if "datum" was still in the term.

This however, is a place it runs off the rails from deep in 7.6:
"When magnified surfaces of manufactured parts are seen to have irregularities, contact is made with a true geometric counterpart at a number of surface extremities or high points." Being theoretical there can be no contact seen by magnification.

And 7.11.2:
"If irregularities on a datum feature are such that the part is unstable (i.e., it rocks) when it is brought into contact
with the corresponding true geometric counterpart,"


It's like someone was paid by the word and no one made a map.

 

[Burunduk]

I get the logic in the hierarchy you listed, but for me, the transition from the imaginary to the physical doesn't take place from step 2 to step 3.

The simulation process in step 3 is non-physical unless a fixture is used. Just because there were photons involved at some step doesn't make a computerized operation performed on a scan-generated file "physical". Again, non-physical still doesn't mean ideal or perfect, but it is nevertheless virtual and doesn't suit to the definition. That's where it goes off tracks, for me.

 

[axym]

Hi All,

This is one case in which Y14.5's old hard-gage descriptions haven't been modified enough to accommodate modern non-contact methods. I agree that "physical" isn't really the right term to describe an optical scan. It's interesting because the geometric counterparts established by coordinate metrology devices are perfect - the software defines perfect planes, perfect cylinders, perfect parallel-plane widths etc. that have perfect form and perfect relationships to each other. The imperfection is in the relationship of these perfect entities to the actual part feature.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[3DDave]

CMMs aren't perfect - they are limited by the numerical precision of the software and the precision of the measurement taking tools. If there is a slope the surface evaluated by the software will have a stair-step to it even though the steps are not noticeable to the typical user. They may be far more precise in some functions than metal gages, but they are not perfect.

For example, decimal 0.1 units in binary representation is an infinite repeating series that is truncated/rounded in the typical IEEE 754 format for floating point arithmetic.

https://stackoverflow.com/questions/588004/is-floating-point-math-broken

0.1 + 0.2 -> 0.30000000000000004

Is it close enough? Usually.
Is it perfect? No.

Even more math:

https://docs.oracle.com/cd/E19957-01/806-3568/ncg_goldberg.html

What Every Computer Scientist Should Know About Floating-Point Arithmetic, by David Goldberg, published in the March, 1991 issue of Computing Surveys. Copyright 1991, Association for Computing Machinery, Inc.

 

[Burunduk]

It's interesting because the geometric counterparts established by coordinate metrology devices are perfect - the software defines perfect planes, perfect cylinders, perfect parallel-plane widths etc. that have perfect form and perfect relationships to each other.

I would say that the simulators established by the software have perfect form and perfect relationship to each other, but an indirect and imperfect relationship to the actual surfaces of the datum features. Therefore they are not perfect as TGCs. A plane simulator or a datum plane tangent to a portion of the scan considered as the planar datum feature may contact the 3 high points on the cloud of points, but it may not properly represent the plane which would contact the 3 highest points on the actual surface that was scanned. To fully represent the TGC, an infinite number of points would need to be scanned with perfect precision. Likewise, a "true" True Geometric Counterpart of a cylindrical actual feature would have to be of a diameter value which is precise to infinite decimal places.