Reporting Requirements for Geometric Tolerances 1

[axym]

Hi All,

Here's a question (more of a survey, really):

When a first article inspection report is produced for a part with geometric tolerances (position, surface profile, flatness, perpendicularity, etc.), what information do you feel is necessary?

1. Measured values for each geometric characteristic

2. Measured values for each basic dimension

3. Other



Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[CheckerHater]

This forum has several threads starting with question: "please, tell me how to measure that".

Strangely, they are left unanswered the longest.

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[axym]

drawoh,

The profile inspection methods you describe sound reasonable, and I have done both myself. I've measured XY coordinates of points on the surface, imported them into CAD, and fitted them to the model. I've also inspected profiles of flat sheet metal parts using a 1:1 Mylar plot with the profile zones added in.

One interesting thing is, that I believe is lost on many people, is that at a certain level of "nastiness" the profile zone becomes much easier to inspect than the dimensions would be. Once you're willing to move away from the idea of directly toleranced dimensions, things get much more clear.

The profile requirements in Fig. 8.5 and 8.6 would both be relatively straightforward to inspect if a coordinate-based inspection or template-based inspection method was used. The complexity of the basic geometry does not become an impediment - there is just a basic shape with a zone, and you verify the part (or measured points) to that zone. Each point is either in the zone or not in the zone.

The point that John-Paul and I were making, and that others have alluded to, is that trying to inspect each dimension directly (i.e. not the profile zone) would be very difficult. Many (most) of the dimensions would be difficult to inspect repeatably on an actual part, and some of them would be impossible to even define on an actual part.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[KENAT]

As to the OP regarding information I want to see on an inspection report.

Well, I'd like to see the inspector actually interpret the GD&T vaguely correctly.

This having just dug into a failed 'circular position' on an open ended slot where the actual FCF was a bilateral position only on the width.

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[3DDave]

There's a temptation to want some sort of deviation information, to see how far from nominal features and dimensioned happened to be.

But in many cases this isn't as helpful as one might like because often the frame of reference those deviations are calculated from is not fixed with respect to any other frame of reference on the part from which other measurements are made. This is certainly the case for any tolerance that has an (M) or (L) modification to a datum reference, but any unstable datum feature will cause problems.

There should be two sections; one in which it is reported on whether or not a solution was found to accept each feature. The other part will record where all features are located and oriented from a fixed frame of reference so that engineering can create a model that is representative of the as-inspected part and use that model to check where the part is relative to the original design goals.

All that is required to accept the part is the first portion, which is the same as if hard gages were used to accept the part.

Analyzing what the results mean in terms of meeting current goals is a complicated effort that is difficult to represent on a simple report on gathered data based solely on the dimensioning and tolerancing, because, by design, using FCFs is to allow variations that cannot be explained by single numbers.

The latter part of the report can be used to analyze if the parts are being made consistently and pin-point consistency failures in the production system. High-definition point cloud capture can be handy for cast parts subject to heat treat and straightening or injection molded parts, while a simpler geometry descriptions can serve for simple prismatic items.