How Literally Should We Take the GD&T Standard?

I hope everyone has had a great summer.   Here’s a topic that will be helpful even to seasoned experts in GD&T, and it kind of follows the previous blog entry…

Often, when discussing the finer points of GD&T with others, we end up going to the official standard (or standards) that pertain to dimensioning and tolerancing in order to seek guidance.  But if you’ve been in the real world, you know that a technical document can’t capture every possible scenario.

So we naturally look for the example or description in the standard that is closest to our real-world situation.  However, we have to make a decision whether we can make the leap of logic to say that a proposed design is still within the “spirit” of the standard.  This can sometimes be a sticky point!

An aside:  The two major standards when it comes to GD&T are ASME Y14.5 and also the ISO series of standards (ISO has an umbrella of several standards that embrace GD&T, not just one book).   The predominant standard in North America — and the one I’m most familiar with — is ASME.  In some ways the two standards have different philosophies about the depth of coverage:  in some areas ASME tries to nail down every option, and in other areas ISO takes the harder line.

There are those who would say that we must make our designs conform to the exact letter of the law, and any practice which is not described in the standard (or an obvious modification of one given in the standard) is not to be used.  But others espouse more leeway and say that the concepts given in the standard can be extended to many other areas that might have been unforeseen by the standard writers.

I put myself into the latter camp.  The key is to look carefully at the wording of the standard: if the words “shall” or “must” are used, then the door is pretty much closed to bending that principle.  But if no prohibition is given — or better yet, if the verb used is “may” — then there are probably other ways to practice the given idea and still be in conformity to the standard.

One concept that can serve as an example is the “tangent plane” modifier.  When introduced in the 1994 ASME standard, this modifier was shown for use on the three orientation symbols (when applied to surfaces).  Eventually, someone was bound to ask whether the tangent plane idea could be used with profile of a surface.  The standard never said this cannot be done, and so my vote would be that it’s OK.  Others said, nope, it wasn’t in the standard, so tangent plane wasn’t to be used on profile of a surface.   In the 2009 standard, the notion of extending “tangent plane” to profile of a surface is now clearly allowed; although no specific example is given, they added a footnote saying that the concept of tangent plane is equally applicable to “other geometric characteristic symbols where the feature is related to a datum(s)” (see page 103 of the current edition of Y14.5).

Bottom line:  GD&T is a language.  And like any language, there are certain rules that must always be followed.  However, there are many ways to patch together different parts of that language and still say something clear.  We shouldn’t be too legalistic and limit ourselves to designs that are only identical to the examples given in the standard.  Obviously, this is where training and knowledge of the fundamentals of GD&T are necessary in order to know when the envelope is being pushed to far!


  1. Hi,

    Is it okay to call a datum surface (ie DATUM A) and a flatness tolerance( with a leader) on one side of a plate and a parallelism tolerance with reference to the specified datum on the other opposite surface or would you only call out the datum on one surface and the parallelism with respect to the datum on the other side. Both surfaces are sealing surfaces.


  2. The situation you describe in the first portion of your question is indeed legal, and in fact it would be the desired way to go if they are sealing surfaces. Reason: Just because a surface is labeled as datum A doesn’t mean that it will be flat. The true datum that is taken during measurement will actually be a plane formed from the high points of that surface A. Think of a bumpy surface that sits on a flat gage table … the gage table actually becomes the datum that we want the opposite face to be parallel to. The flatness of the surface contacting the gage table is not part of the measurement (unless the print adds a flatness tolerance as you mention).

    Hope that helps!

  3. When using TOP, must the features always be perpendicular to the Primary Datum surface? In this case, a mounting pad with 4 threaded holes is being welded into a sub-assembly and the surface that the holes are perpendicular to is the relatively thin edge of another member of the assembly – in other words, the mounting plate is being welded edge-on to another flat surface. Can that flat surface be identified as the Primary Datum?

  4. My previous comment was badly worded!
    The 4-hole mounting pad is not directly welded to the other flat surface but we do need to locate the holes from that surface. Probably, this would normally be considered the Secondary Datum but can it be designated as the Primary?

    • The face perpendicular to the holes is often called out as the primary datum, but it is my no means a requirement. The function of the part drives the datum selection, but sometimes we may have to bend that rule if a certain surface is not accessible, or for other reasons.
      For your situation, anytime you mention threaded holes, my thinking is that there will be some adjacent piece that will be forced (by the torquing of the fasteners) to become flush with the face that is perpendicular to the holes. So even if it’s a relatively thin part, the narrow face could be the primary datum. Obviously, that’s subject to some caveats but those are my thoughts about your question.

Submit a Comment

Your email address will not be published. Required fields are marked *