Terms: "Best-Fit" & "Least Square Fit" & "Adaptive Positioning&quot

[dtmbiz]

The part is a complex surface DLD (Direct Laser Disposition" with machining similar to material removal from a casting.

I have heard of "best-fit" in the past for castings. Recently I had access to ASME Y14.8 standard for castings .

Word search did not find this term defined in the standard as a definition or principle. There was reference to "best-fit" but not defined.

I have never heard of Least Square Fit in ASME standards.

The part has two complex surfaces at opposing ends of the part, each identified as a Datum Feature. H and G respectively.

Profile callouts use H-G as a single primary Datum.

General Notes references "Least Squares Best Fit" for "Adaptive Position"

I have never heard of "Adaptive Positioning" either.

Is anyone familiar with this ?

Is this Voo -Doo ?

 

[3DDave]

Analytically you are supposed to jiggle the part until it has the closest fit on both ends at the same time, however you want to define 'close.' This is usually the result of the person doing the design looking at their analysis based on perfect parts and throwing on some requirements to make figuring out what variations are acceptable onto the the QA/QC group.

 

[axym]

dtmbiz,

I haven't heard of "adaptive positioning" either.

Least squares fits are not defined in ASME dimensioning and tolerancing standards, that I know of. Y14.5 allows a different fitting routine to be specified, if the default of high point contact is not desired.

The primary datum H-G could be calculated using CMM software, if points were measured on both surfaces. This seems to be the intent. There will be some variability in the datum, however, as it will depend on where the measured points are and how many there are.

When I see complex surfaces specified as datum features, in many cases this doesn't follow how the part fits. Does the part contact corresponding complex surfaces on the mating part?

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[dtmbiz]

Axym

These surfaces do match the mating surfaces. (t appears to be an almost line to line fit)

It is an airfoil that the outboard vane contour (datum feature H) mates with outboard platform; the inboard vane contour (datum feature G) mates with inboard platform.

A plane derived from either contour would more closely follow (be parallel to) the engine axis (At least with my very limited knowledge of deriving a plane by Least Square equations);
in contrast, a G-H plane would more closely align with the stacking axis to make any kind of sense, as I see it. (I have know idea what datum plane is trying to be achieved)

I do not believe that it is at all legal using the hyphenated concept in this configuration according to ASME Y14.5M 1994.


Axym, 3DDave

Have no idea how this "jiggling" can result in "reliable and repeatable" measurements...

Thank you for your feedback...

 

[3DDave]

It doesn't, but guess what, if that isn't a workable method on a CMM, it won't be any better in the actual assembly. 'Reliable' and 'repeatable' are vague in the sense that inspection is pass/fail. If you can find an orientation that passes then it passes. If you give up or the part can't meet the requirements then it fails. If there is room to jiggle the part, then there are an infinite number of 'passes' configurations for the setup and there will be an infinite number of alignments in the assembly; trying to find the 'best' is a waste of time. If the desire is for a repeatable/reliable answer, that time could be better spent changing the design to eliminate the ability to have more than one alignment.

The hyphenation says that both datum features are used to align and locate the part. The datum planes are oriented to the frame of reference, not the part. It's the goal to align the part to the planes. Without a picture it's hard to say for sure what is being controlled.