The problem exists in the shaft itself. It had a callout of 2.000" diameter +/-.005, and it was out of round by .0015. The shaft itself was in tolerance, but its out-of-roundness condition was going to make it difficult to determine if the spud was concentric to it since it, datum A, would have a varying centerline.

My question is: When a feature is called out to a daturm as described above, is there an assumption that the datum should be as good or better than the feature(s) dimensioned to it? Is there anything in the standard that supports this?

Consider a hole perpendicular to a planar datum. If the part feature that corresponds to the datum is very lumpy - it doesn't matter. The part is placed onto a suitably flat datum simulator - regardless of any flatness callout for the datum feature - and then the perpendicularity is checked to the flat datum simulator.

This will cause trouble if the mating feature is not held as flat as the datum simulator. Few D&T analysts think to look for the effect that such contact creates. Some guidance is held in the ...14.5.1... Mathematical portion, but even that leaves many questions open for the planar case.

Any thoughts on this?

ASME Y14.5M-1994, consider the following:

2.8.1 Effect of RFS "... referencing a datum feature on an RFS basis means that a centering about it's axis or center plane is necessary..."
-Concentricity is always applied on an RFS basis.

4.5.3 Specifying Datum Features RFS "...the datum is established by physical contact between the feature surface(s) and surface(s) of the processing equipment. A machine element that is variable in size (such as a chuck, mandrel, vice, or centering device) is used to simulate the true geometric counter part of the feature and to establish the datum axis..."

So those are the rules. A mandrel or chuck would contact the shaft in your example (on the hight points) only then does the theoretical datum axis exist. The datum axis is not the axis of the gaging mandrel or chuck, nor is it the axis of the datum feature (part feature) in contact with the gage, although they may be very close in location and orientation there are three seperate and distinct axis.
1 the axis of the gaging equipment,
2 the feature axis,
3 the Datum axis, which is the resultant of how the feature fits into the gage.
Measurements must be made from the Datum axis.

---

LGW (Apr 06, 2005 04:05 p.m.):
I agree with both Dave and ASME-MAN, theoretically speaking. However, I still have an issue with how one would establish the datum from an imperfect datum feature. In other words, how does the customer communicate to the manufacturer, using GD&T, the expectation of the designer? I really don't see where ANSI Y14.5 addresses how a datum centerline is derived from an imperfect datum feature such as a shaft. There is no rule-of-thumb that I am aware of.

Any thoughts on this?
LGW

---

LGW, Please take a minute and carefully read 4.2.1 on application of datums. You will see that the tooling that simulates the datums is used to establish the axes, planes, points, and centerplanes. Also take a careful look at figure 4-10. This shows how the irregularities in the part can interact with the irregularities in the datum simulator to produce a result that is not perfect. The intention is to show that one should make the gaging/simulators sufficiently accurately so as not to contribute in a large way to measurement errors.

This same section should be familiar to both the inspector and the manufacturing engineer and/or the machinist. The entire point of ANSI and ASME Y14.5... was to have a single description that all participants would refer to so that no further explanation is required.

What isn't in either standard are algorithms or processes or gaging standards. It is left to the various participants to decide how much to spend and how great an effort to make. Usually, this decision is swayed by the holder of the cash, who should ask his supplier just exactly what plan is in place to manufacture to the requirements of the drawing. Materials, plating, painting, marking, and dimensioning and tolerancing should all be reviewed for conformance.

Thanks for your input. I believe what it comes down to is this: the customer needs to specify how the part is going to be used or gaged in order for both the customer and the manufacturer to agree on the results of the concentricity callout.

The ANSI standard only goes so far in explanation, and the rest is left for the designer to fine-tune. Gaging techniques and methods can produce varying results which the standard cannot anticipate.

Thanks again,