Flatness (theory) 1

[Kedu]

Are there any definitive statements from Y14.5 (1994, 2009, 2018 or any version) to reflect that flatness shall be applied to a (nominally) planar surface only?
Or, in other words, what prevents the application of flatness to a curve?

 

[Belanger]

Are there any definitive statements from Y14.5...to reflect that flatness shall be applied to a (nominally) planar surface only?

Not in the way you phrase it ("surface only") because the 2009 and 2018 standards permit flatness to be applied to a FOS, in addition to surfaces.

What prevents the application of flatness to a curve?

To use the 2009 version, Paragraph 5.4.2 states that "Flatness is the condition of a surface or derived median plane having all elements in one plane." The idea of "one plane" seems to preclude its use on curves, at least in the intent of the flatness symbol.

But are you thinking about a basically defined curve with a radius so large that it might fit within a flatness tolerance's allowable zone?

 

[Kedu]

Belanger,
Your provided scenario (a very large radius) is one of the way I am thinking to use flatness for a curved surface, but another case could be a sin wave or gaussian curve.

Why flatness could not be applied to such surfaces? What prohibition could you find in Y14.5 for such of practice? Otherwise stated, if not forbiden then it could be used, in my opinion.

From the theory point of view I cannot find in Y14.5 why is illegal to use flatness of the described cases. Just some statements about flatness definition, but again not a direct prohibition.

 

[3DDave]

How would flatness apply to anything besides either a nominally planar surface or a plane derived as a median to a pair of surfaces?

In the latter case it functions approximately as a symmetry control about the derived plane so it could apply to a wedge or roughly mirrored geometry of any kind that isn't cylindrical (special rule against that.)

Other than that there are an infinite number of ways to not use the standard correctly and it would be an imposition to require the committee to document all of them.

 

[chez311]

The excerpt JP pointed out in 2009 would seem to explicitly restrict use to nominally flat surfaces. This verbiage is removed in 2018 though.

I'm with 3DDave, we could invent all sorts of ways to use the standard in incorrect ways for theoretical reasons. We could apply a large runout tolerance to a rectangular prism, aside from some references to surfaces of revolution I don't see any explicit requirements that the cross section of the considered feature must be circular.

For practical purposes, if you have a surface which you truly don't care if it was flat or curved within a prescribed flatness tolerance, what practical functional reason would you have to not modelling it as nominally flat?

 

[Dean Watts]

Kedu,

I think the sentence about a surface having all elements in one plane is an explanation about the concept about what "flat" is, rather than a restriction regarding the type of feature that a flatness tolerance may be applied to.

I think you could apply flatness to a feature such as you're proposing, as long as you're willing to accept any shape within that flatness tolerance zone. I also think most every case it would be better to use profile, but since you're asking from an academic point of view (as you stated in another thread), I think the answer can be yes.

Dean

http://www.validate-3d.com

 

[3DDave]

Hi Chez - thanks for the link.

Dean - are you saying that the surface could have some amount of texture applied? Like a surface with a specified undulation with some small amplitude?

 

[axym]

Hi All,

I agree with 3DDave that we can apply tolerances in many ways that were not intended or foreseen, and we can't always determine their "legality". I could model a feature at 89 degrees to another feature and specify a perpendicularity tolerance. Is that legal? Someone might say that it needs to be angularity, but what if I actually wanted the zone to be perpendicular to the datum? Sometimes we just need to weigh the potential benefits of an unusual application against the confusion that it will probably cause.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[Belanger]

So maybe the flat-earthers are right: We can legitimately say that the Earth is flat, as long as we accept a tolerance of 7,918 miles!

 

[TheTick]

You can't handle perfect flatness!

https://youtu.be/-MwCJpEuC44

 

[Ajith NJ]

Hi
Flatness can be applied only to flat surfaces. May be you are thinking with respect to actual produced part. Because if you say flatness of 1mm , then within this 2 imaginary plane of 1mm gap the surface can take any form like concave or convex or wedge or zig zag etc etc. So ultimately at design stage the call out was for flat surface itself. But later at manufacturing because of more flatness tolerance it may look like curvature. I am just exaggerating.
Profile tolerance is preferred method of controlling the form of curvature.
And also one additional point about flatness. Flatness can be given to median plane also as per 2009 standard. This is how flatness deviates rule #1 of Gd&t.

Thank you

 

[mfgenggear]

I agree with @Ajith NJ
I had to think this carefully, flatness symbol are denoted on flat surfaces, however
a cylinder or a shaft spud can be lapped w/I x number of light bands.
however a different symbol call out is required, for example cylindricity. profile all around.
and a foot note would be required, example lap W/I 3 light bands.

 

[Belanger]

Ajith NJ -- what you write is indeed the intended use of flatness. But the OP was stretching the thinking to ask whether it's absolutely, always required to be applied to a nominally flat surface.
According to the way you describe it, "within this 2 imaginary plane of 1mm gap the surface can take any form like concave or convex or wedge or zig zag etc etc." Think about it: if the surface "can take any form" then could the surface even have been curved by design, perhaps with a defined radius of 1 mm, and technically still be within the tolerance? The letter of the law might actually allow for an answer of "yes."

I don't really like that idea, and I reiterate that flatness was intended for surfaces that are nominally flat (or median planes, since 2009), but it does make for a good discussion!

 

[Ajith NJ]

Hi
Good discussion!

No doubt that flatness is applied to nominally flat surface only.
Gd&t is mode of communication how designer tells his intention to the manufacturer. So how come designer can take liberty in telling his intention with the highly sophisticated 3D softwares available today.
Designer should always use nominal way of representing his design which off course is ideal one. Tolerance is for manufacturer which he definitely needs to produce the part. So shapes I was mentioning is just some cases that actual produced part can take within tolerance zone.
Coming to your point. Showing 1mm radius to the surface and asking for flatness tolerance at design stage itself, is like designer himself taking tolerance to show his intention. Eventhough experienced manufacturer understands what he has to produce by seeing this kind of call-out, I am sure he will ask for revision of this drawing in coming days to correct that.

But I definitely agree with you that GD&t can't overcome these kind discrepancies and these situations are very common because of human error.


Thanks you.

 

[3DDave]

I expect a flatness can apply to an embossed aluminum sheet, particularly where one does not care to detail the profile of the embossed finish, which might be a semi-random pattern.