Hope this is the right forum to ask this question. If not, please feel free to move it to the appropriate one.

I have been requested to develope a measurement method for a part (see part print in attached .pdf file). The problem is in measuring the two hole positions (7a and 7b) as called out on the print. When the C datum has a profile tolerance, AND the position has an MMC modifier for the C datum, how do you set up the measurement system for variable measurement data? Equipment available (beyond the basic) is a Starrett Contour Projecter with edge detector and QC4000 software, and a B&S CMM with WTutor software.

I have also attached the nominal print geometry and actual part measurement, as two .dwg files.

If anyone can describe the 'correct' method of aligning the part for layout, I would be happy to hear from you.

After your responses, I will also share the method that I have been urged to use by one of our Project Engineers; not sure if it is correct or not, hence the questions.

Thanks to all who respond.

Alignment for inspection with the contour projector:
• Align Surface A (parallel to the ways of the staging surface)&(perpendicular to the light source)
• Rotate the view screen to the basic angle 53° and align the edge of surface C to the view screen cross hair line.
• Move reference to the center of diameter B and zero the coordinate measuring scales.

Alignment for inspection with the B&S CMM:
• Establish a plane either by probing surface A directly or by probing the surface that it will lie upon for inspection (the datum feature simulator) and align the coordinate system to that plane.
• Construct a line from points gathered on surface C and rotate the coordinate system 53° relative to the line.
• Probe diameter B and move reference to its centroid.

Measurement of the two holes: (Position allowances for datum shift disregarded)
• Probe each hole for size and location
• Keep track of the individual X & Y coordinate measurements and the feature sizes of the two holes and create control charts for that data. The coordinate mean errors from target can be used to correct the fine blanking or powdered metal die tools.
• Subtract the actual size from the MMC size of each hole to figure its bonus for position tolerance.
• Add the individual bonus to the minimum allowable position tolerance @ MMC for each hole.
• Subtract the individual computed position error from the individual variable position tolerance and record the results as pass/fail for discreet data attribute predictions.
• If you wish to generate variables data process capability predictions for the computed position errors to the variable tolerance limits you can do that by modifying the process capability equation to acknowledge a variable specification limit and assuming normality for the distributions for both size and the position deviation. Ppu(Variable) = (minimum variable tolerance + mean bonus – mean position deviation) / ( estimated variance of size + estimated variance of the position deviation)^.5 ...edit "sorry about that its the square root not the square".

Note: Datum feature bonuses must be applied simultaneously and unidirectionally to all features sharing the same datum structure and modifier designations. They cannot be applied individually to each hole!

• If you have nothing better to do and you would like to attempt to apply the bonus tolerances for datum-shift to each individual part you can do that graphically by creating scale-sized overlay of the part with the virtual condition of diameter B and both fastener holes and lines for the boundaries surface C matching the amplification of the shadow graph. Then shift the overly visually within the datum feature boundaries to attempt to place the virtual condition boundary of the two holes within their shadow boundaries on the view screen.
• Applying the datum feature bonuses mathematically in a CMM algorithm is quite involved would have to be a iterative process where the residual or remaining variable tolerance for each hole is computed along with its deviation vector... then candidate datum references can be translated and rotated along a common resolved vector with magnitude limits that do not permit the datum features to exceed their boundaries. Re-compute the position deviations and residual tolerances and repeat until the residual tolerance in both holes,diameter B, and surface C is maximized.

Or just make up a hard-gage with virtual conditions for diameter B and both holes and boundary limits for C and check them with that because it does the same process physically. And tell your STA and ISO auditor that you can't do Cpk's with variable tolerances. It is a lot easier that way.

Paul

Who have this standard of following?
The Metric System Applied to Geometric Tolerancing (reference: ANSI/ASME Y14.5M)

Wangxingde
The American National Standard is available at the following source,
http://members.asme.org/catalog/ItemView.cfm?ItemNumber=N00594

So Ron,
I'm curious what was the other method that you were urged to use from the project engineer?

Yes. I'm curious, too.

Sorry for the delay guys, my internet was down for 2 weeks (withdrawal pangs are SLOWLY subsiding).

In a nutshell, using the contour projector, I would measure the part 'as-is', creating both the B and C datums, plus a series of discrete points. The points and features would then be exported to .dxf file.

This .dxf file would be overlaid onto the nominal part geometry, and rotated around B, so that all the points of C were within the tolerance bands, while finding the 'best fit' for the remainder of the points.

I'm still not convinced that this is 'correct'; the C datum is shown as a specific rotational angle from B.

Opinions?