Tangent Plane Modifier

In my opinion, one of the most underutilized tools in the GD&T toolbox is the tangent plane modifier. It was introduced in the 1994 ASME standard, yet some people still think of it as a new concept.  Shown in the example below, the tangent plane modifier (circled T) can save money by only controlling the high points of a surface, rather than every point:

To understand the drawing above, first realize what parallelism controls if no T modifier is given.  Regular parallelism requires that every point on the top of the part be within a tolerance zone of 0.2 mm. This means that regular parallelism inherently controls flatness to the same specification.

But there might be times when a designer does not need to control flatness. Perhaps another mating part will contact the top of our part, and we only care about the angle at which the mating part sits.  In that case, we don’t need the surface to be flat within 0.2, since our mating part will only feel the high points anyhow.

In that case, the tangent plane modifier makes sense.  It does not give us a “bonus tolerance” as the MMC modifier does with features of size, but it does have the advantage of being more forgiving of the surface’s form error.  The following is an illustration of the possible error that would be allowed by the tangent plane modifier:

Notice that portions of the surface can actually go below the tolerance boundary; this is because the tolerance is imposed only on the imaginary tangent plane — this can easily be inspected by placing a gage block or a flat plate on the top and then measuring the gage block’s parallelism.

Of course, there are times when this is not desired, such as when there are concerns about fluid leaking between this part and a mating part. And it should also be noted that the tangent plane modifier can be used not just on parallelism, but also on perpendicularity, angularity, and even profile of a surface.


  1. Thank you so much for the “tangent plane modifier” explanation, i am currently a student in a GD&T class, this helped immensely. book marking site, thanks again!

  2. I got knew what is the importance of tangent plane. where to use thanks, keep posting this type of knowlege sharing on GD&T concpet

  3. Thanks for this article. I thought I’d seen or used nearly everything interesting for my industry in the 1994 standard, yet here’s something that could be eminently useful in allowing more tolerance on parts in the typical mechanical assembly, especially combined with the suggestion that it may be legally usable on a profile of a surface requirement in the other article.

    The only problem is that this will probably blow the average metrologist’s mind as much as zero tolerance at MMC usually does. Even once you explain it they are clearly still skeptical…

  4. Thank you, this article is been very helpful.
    Does a flatness control form on a surface will be unnecessary when the surface itself has a parallelism control without the tangent plane modifier?
    Thank you.

    • Yes Eric, that’s right. Parallelism on a surface (without the T) means that all points on the surface must fall within the tolerance zone. So it inherently controls flatness at the same time, to the same tolerance value.

  5. Something to think about when using Tangent Plane modifier: What if your three high points form a straight line? Your mating part will rock back and forth making the “angle at which the mating part sits” ambiguous. Or, if the three points form a triangle, but are very close together, this would also cause rocking potential. Using a Flatness callout could have the same issues of course, but it seems to me it would be more controllable.

    • Indeed a good point, Dan. There’s been a similar discussion about establishing a datum plane (which is really the same thing: a tangent plane). The funny thing is that the 1994 edition of the Y14.5 standard addressed it, in paragraph 4.5.1. They said that if such a situation arises, the “part may be adjusted to an optimum position…” In other words, you can rock the plane however you want in order to make the thing pass!
      However, in the 2009 standard, they back off from this position a little (see paragraph 4.11.2). This time, they simply appeal to a “default stabilization procedure” as outlined in Y14.5.1M.
      But I agree, if the rocking scenario happens, that might have been a good candidate for flatness, to at least minimize the effect.

  6. What is maximum flatness deviation on the top surface in this example ?

    • Well, that often depends on whether the drawing is using the ASME or ISO rules of GD&T.
      The ASME understanding is that flatness is automatically controlled by the size tolerance, so in our case the maximum flatness deviation across the top would be 1 mm (from the height tolerance of ± 0.5).

      The ISO understanding of size (height) is that it controls only the “actual local size,” or all individual two-point measurements between the top and bottom. So ISO doesn’t see size as automatically controlling flatness at all.

      However, ISO doesn’t have the “T” modifier (maybe in the future they will). So we know that this print is using ASME; thus the answer to your question is 1 mm.

  7. Fantastic article, yet very simple to understand. Thank you so much. I also thank all the commentators here, very good discussion followed.

  8. May I ask for your opinion: According to Y14.5 can (allow) tangent plane modifier be used for any other surface than planar surfaces (for example a curved surface)? Are there any restriction in the standard itself or does not make sense per the physical realities? Any text in 1994 or 2009 to solve this dilemma

    • I would say that the use of “T” on anything other than a nominally flat surface was not envisioned by the standard. Therefore it’s mainly used on parallelism, perpendicularity, and angularity. It could also be used on profile of a surface if it’s designed to be a flat surface and will be referencing one or more datums.

      Suppose we have a nominal geometry that has multiple curves (such as a sine-wave shape, with multiple humps). You might say that all the peaks of that sine wave do create a tangent plane, but again I don’t think that is where the standard would want the tangent modifier. The statement in the standard that alerts me to the standard’s intention is in paragraph 6.5: “When a tangent plane symbol is specified with a geometric tolerance, the flatness of the toleranced feature is not controlled by the geometric tolerance.”

      So the solution — like many of these non-typical applications — would probably be to add a note beneath the feature control frame if you want a curved surface to be measured in a tangent-plane manner.

  9. I’m considering using this control, but I’m wondering two things. 1)
    If using a CMM to inspect, would this make the inspection more difficult because now the probe would have to search for and locate the three highest points on a surface? and 2) Is there a way or recommendation for specifying a “contact pressure” to inspect at? For a bolted interface, I’m thinking the tangent plane found during inspection could easily be different than the actual tangent plane because the three high points will compress under the load of the screws.

    • Good questions, Scott — and I’ll divide them into two main things:
      1) When using a CMM there is always a question of how much sampling of a feature is adequate. But for the tangent plane modifier, it’s actually easier for the CMM person! This is because they could simply lay a gage plate on the surface in question, and then probe the gage plate. Thus, any form variation on the actual part is sort of washed out, which is the intent of the circled T.
      2) Regarding a contact pressure, if the part in question will function with a load (such as bolted parts, or especially flexible parts that will conform upon assembly), then it might be wise to add a restraint note so that those actual forces are simulated during the inspection. The Y14.5 standard tells us that parts are — by default — check in the “free state.” But you mentioned a load coming from screws, so I’m thinking that a written note (near the GD&T callout, or perhaps a general note for the entire drawing) should be used.

  10. I have some fairly ‘new’ engineering models, the Mechanical Engineers have put these Tangent profiles to every stiffener on this sheetmetal part. The only thing that makes these slightly different, is the tangent profiles are to 2 datums. Making me think that they want to control every aspect of the stiffeners, except up and down movement of the part. Some of the tangent profiles are .020 and others are .060. Do you think this a feasible application of the tangent profiles.

    • I’m not sure I can envision what you describe, so I’m not sure.
      But the fact that the profiles are referencing 2 datums is not a problem at all. And the tolerance values wouldn’t seem to be a problem — almost any number is possible, assuming that it doesn’t violate any other tolerance applied to the same surface(s).

Submit a Comment

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