Paul,
The purpose of the composite positional control is that you don't have to overly restrict the general position of the pattern while allowing the features to be more tightly positioned wrt each other, whether a planar layout of features or coaxial. So, by using only the second level of the composite control, with the tighter tolerance, you would be unduly restricting the position of the pattern as a group. The "uninterrupted" or continuous feature point is valid in this particular application because the holes will likely be fabricated at the same time as a single tool pass, but what if they were different sizes? While they could be fabricated using a stepped drill, the probability of such tooling drops with the decrease in production batch size. As a result, you would need to use a composite control to first locate the two holes wrt the DRF, then second to control the coaxiality of the two holes to each other.
Dave, Dave, Dave my friend... every part needs GD&T.
![[soapbox]](/data/assets/smilies/soapbox.gif "[soapbox] [soapbox]")
Without invoking Y14.5, you don't even get Rule #1. With GD&T, you get the size controlling the form at least. With Y14.5, you get documented implied conditions; without it you make assumptions based on "common practice". With GD&T, you have a documented means of controlling size, form, orientation and location of every feature wrt a datum structure. With GD&T, you know how to set up the part for inspection. With GD&T, you minimize tolerance accumulation, and you control the feature under consideration, not the dimension to a point in space.
I've frequently heard people say that they only use GD&T on critical features,; those that they care about. My questions then are:
1) so you don't care how the part is set up for inspection?
2) if you don't care about the feature, why have it?
3) if the feature broaches another feature but meets the +/- spec, is it ok?
4) a square or rectangular position zone is ok, with its biases?
5) you don't care if a rod or pin is bowed, or a rectangular feature is twisted and warped?
6) do you want to use a general tolerance that doesn't conflict with the geometric controls?
7) which is the overriding control ... a conventional +/- tolerance or a geometric control, when both are applied on the drawing?
8) do you think that you don't have to verify any of the conventionally-toleranced features when geometric controls are present?
I agree that designers shouldn't be tossing controls onto the pile just because they're there; they should be conservative in using only the controls necessary to attain the functionality they need. That, however, does not mean not using GD&T, it means judicious application of controls. My perspective is that, as an engineer, I want to fully define all relationships whether they have a tight tolerance value, or a loose tolerance value. By defining all relationships, I can avoid wasted time explaining what the drawing means, and avoid civil action to resolve open interpretation of an engineering print. Now, that being said, there are certain "junk" features that are appropriate applications of conventional tolerances; fillets, rounds & chamfers are the most prevalent.
Jim Sykes, P.Eng, GDTP-S
Profile Services
TecEase, Inc.
