ewh,
Yes, others here may have differing interpretations!
SeasonLee,
I like your GD&T drawing. There might be a couple of things on it that need to be tweaked, but it's pretty good. It's already vastly superior to the plus/minus drawing and I don't think any of the GD&T is unnecessary or overkill. Keep doing what you're doing.
I'm wondering about the details of how the part is constrained in assembly or fixturing, that led to Datum B being referenced at MMC. This suggests that Datum B's mating feature is a fixed size and may have some slop, leading to uncertainty as to which datum feature really stops translation in the Y direction - B or C?
Dave,
You're right, this is a can of worms! The anecdotes that you mention are unfortunate, but I don't think that the fact that GD&T was used is the problem. If a company has default tolerances that "don't mean anything", then that is a problem, but it is separate from GD&T. I'm not a fan of block tolerances because they tend to be nonfunctional, whether they're plus/minus or GD&T. If someone is making dedicated gages to inspect lightening holes, then again it's not GD&T that's the problem. Something else has gone wrong in the communication of design intent. The designer probably specified a position tolerance that was much tighter than necessary, or the lightening hole was covered by a default position tolerance just because it was a hole. Possibly the inspection planner thought "it's a position tolerance, I'd better make a gage even if the tolerance is really big" which isn't right either.
Regarding whether or not profile should be used on the perimeter surfaces of a part that are just "in space", i.e. not contacting anything, I say why not use profile? It's not that those surfaces have no functional design requirement - it's just that they don't touch anything. The requirement is that the form and location be close enough to nominal, for other reasons - rigidity, cosmetic appearance, who knows?
I realize that the standard does not mandate that all, or any, features be controlled with GD&T. But the argument that plus/minus dimensions reflect functional requirements better than GD&T seldom holds water, if ever. Take a look at the plus/minus tolerances on SeasonLee's first drawing and see how functional they look. .541 +/- .005 between an imaginary "vertical" line and the imaginary center of a fillet rad? 30 +/- 1 degree between a planar surface and an imaginary "horizontal" line? Functional, I don't think so.
I think that the opposition to GD&T is often a result of a quest to "keep things simple". The designer wants a quick and simple drawing, the manufacturing guy wants simple dimensions to program from, and the inspector wants simple characteristics that can be inspected with hand tools. Plus/minus tolerancing satisfies all of these things. This works fine for certain parts, where the form and orientation error of the features is small compared to the location tolerances that are required.
But on a piece of formed sheet metal like SeasonLee's part, things are not that simple. Those plus/minus dimensions and angles are a nightmare to inspect in a repeatable way on sheet metal, believe me I know. It doesn't matter if hand tools are used or if a CMM is used - the linear distances and angles become ambiguous on a real part with form error. This is why we have GD&T, for parts like this!
Once the hope for "simple" hand tool inspection is abandoned, and coordinate measurement technology is embraced, GD&T is generally much easier to inspect. If I were the CMM programmer, I would take SeasonLee's GD&T drawing over the plus/minus version in a heartbeat.
I think that one of the main reasons for "supplier apprehension" and accompanying cost increases is that they won't be able to measure the parts with just calipers and a protractor. If this is a limitation that the designer is forced to work within, then making a GD&T drawing won't add any value.
Evan Janeshewski
Axymetrix Quality Engineering Inc.
