Profile tolerance : Offset surface

[Madhu454]

Hi All,
Have one question regarding profile of a surface tolerance. The tolerance zone is formed by sweeping a spherical ball of size equal to tolerance value over the toleranced surface to have the upper limit and lower limit surfaces.

1. Can’t we simply offset the toleranced surfaces to get the tolerance zone.
2. Want to know the exact idea of using spherical ball concept to define the tolerance zone. (Is there any manufacturing considerations)?

 

[CheckerHater]

Pmarc,
I am trying to point out the fact, that if you go by ASME mathematical definition ALONE you will have round corners, hence illustration to my post from 30 Oct 13 6:51.
Then I notice, that ASME made arbitrary decision that offset surfaces have to be extended ‘til they intersect, BUT did not provide mathematical definition, HOW EXACTLY they have to be extended.
That makes it open to interpretation, or, in other words, AMBIGUOUS.
This is what I meant saying that BOTH ISO and ASME are not clear in their definitions.
Axym,
I’ve already answered your question in my post from 29 Oct 13 13:16. It’s from the book by Dr. Nielsen who happens to be Chair of ISO technical committee ISO/TC 213.

 

[pmarc]

CH,
I took a look to Y14.5.1. Of course it is true that it says that:
"A profile tolerance zone is an area or a volume generated by offsetting each point on the nominal surface in a direction normal to the nominal surface at that point".
But to be honest, after reading this statement, I do not really see a problem with identification of a direction of the offset at the sharp corner.

If 2 profiled surfaces are nominally flat, like shown on your picture:

http://files.engineering.com/getfile.aspx?folder=fe4ab54d-5bf7-433c-a675-69cf1dc7b20d&file=Draw1.JPG

the sharp corner point, which is common for both surfaces, is offset in two directions:
1. Direction normal to the true profile of the horizontal surface.
2. Direction normal to the true profile of the vertical surface.
This is my understanding of the quote.

The math standard is silent on what the shape of remaining area of profile tolerance zone is, but this is where the Y14.5 standard comes into play with some figures showing that offset surfaces have to be extended until they intersect, and that they form sharp, not rounded, tolerance zone. Granted, this is Y14.5's arbitrary decision, but I think it is much better than nothing offered by Y14.5.1.

Having said that, on the attached picture I still do not see a justification for presence of two radial lines emanating from the corner.

Of course I am also interested to hear why you think that ISO definition of "all-around" leads to ambiguity (maybe a reason to start another thread?)

 

[CheckerHater]

OK, could you help me to extend the tolerance zone?

 

[pmarc]

If this drawing is governed by Y14.5 - no, I could not.
If the drawing was governed by ISO, then I would say the tolerance zone would be rounded at the tip.

But that does not mean your interpretation shown in the previous sketch is right. It just means that ASME definition has certain shortcomings - like some other definitions throughout the standard.

 

[CheckerHater]

So, do you agree that ASME definition of all-around profile fails without proper mathematical backing?

Does it make it open to interpretation?

 

[pmarc]

I admit that ASME definition of all-around profile is not mathematically backed up for each and every case.

However, I also claim that the interpretation presented in your previous sketch is not in accordance with Y14.5 and Y14.5.1. One does not really have to go deep into details of the math standard to find out that the tolerance zone in your graphic can't be rounded - figs. 8-8 and 8-12 show that without any ambiguity.

And finally, if, despite of all of that, you still claim that your interpretation of shape of profile tolerance zone is correct (based on your understanding of the quote from Y14.5.1, and somehow ignoring what Y14.5 shows), read the last sentence of paragraph 1.2 in Y14.5-2009: "In the event of a conflict between the text of this Standard and the references cited herein, the text of this Standard shall take precedence".

 

[CheckerHater]

OK, I repeat one more time, the picture shown in 30 Oct 13 6:51 based on 14.5.1 definition alone, without taking 14.5 into consideration.
The picture shows normals drawn to edge. If it’s difficult to visualize, imagine (by extension of principle) that sharp edge is actually re-e-e-e-e-e-eally small radius.
If you insist that 14.5 supersedes 14.5.1 you must come up with interpretation that will allow you to extend the tolerance zone on my picture from 31 Oct 13 10:01 based on figs. 8-8 and 8-12.
But enough of that.
Pmarc, axym, season,
24 hours ago I politely asked to produce reference to any reliable source showing how ISO all-around tolerance zone behaves when going around the sharp corner.
The only response I got so far was “you are wrong”.
I suggest we stop this discussion where it is.
Instead we could try and discuss to finer detail, as pmarc suggested, ISO definition of all-around.
Unfortunately I won’t be able to present my case ‘til at least late afternoon.

 

[axym]

CH,

Sorry to repeat the same question - I overlooked your initial response about the textbook. If my response came across as "you are wrong", then I apologize - I am just reluctant to take any GD&T textbook as a "reliable source". This is from experience with textbooks written by various members of the Y14.5 subcommittee, that have content that does not completely agree with the content of the Y14.5 standard. I suspect that similar differences might exist on the ISO side, so I wanted to see how the ISO standards define ALL AROUND. I still have not looked this up.

I agree with you that the Y14.5.1M-1994 definition is lacking, and we cannot determine the corner geometry based on this definition alone. It does not even mention the preservation of sharp corners, that Y14.5 mentions. So I suppose that the rounded corner interpretation does not conflict with the Y14.5.1M-1994 definition, but it would conflict with the Y14.5 definition. The definitions in Y14.5 and Y14.5.1 both have gaps and need improvement. So yes, I would agree that the ASME definitions for all-around profile fail for certain types of geometry, because we don't know exactly what happens at the corners.

Your picture from 31 Oct 13 10:01 is quite compelling. This geometry completely defeats the ASME Y14.5 definition - I can't see how to extend those boundaries in a meaningful way. As the angle between the two surfaces at the tip approaches zero, the extension of the "outside corner" of the tolerance zone approaches infinity. I agree with pmarc that the zone would be rounded at the tip, if the ISO definition was used. This part seems straightforward to me, based on the idea of the sphere that is swept along the true profile.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[CheckerHater]

Thank you Evan,
It’s even funnier if angle does not approach zero (see picture)
This is why I suggested to stop arguing about shortcomings of 14.5 and check our (mis)understandings of deeper underlying concepts like what is happening to “all-around” in ASME and ISO worlds.

 

[pmarc]

In my reply from 29 Oct 13 13:48 I said:
"At the end of the day, I am afraid we will not get consensus, until some new/revised ISO standard(s) clearly show(s) how to deal with sharp corners controlled by profile tolerance."

I said it, because as far as I am aware there is no official ISO document specifically addressing this issue. At the same time I think that swept-sphere approach must result in a rounded corner of one of profile tolerance zone boundaries. As Evan said, this seems quite straightforward.

As for the figure 7-3 from Mr. Nielsen handbook, I have a feeling (but this is just a feeling based on the title of the figure), that its intention was not to show in details how the profile tolerance zones look like in vicinity of the corners. It was rather to show that profiles with and without CZ modifier have different meanings according to ISO GPS philosophy.

 

[CheckerHater]

I agree with your doubts about fig 7-3, BUT…
So far it’s the only picture that actually shows what’s happening when several separate zones become one “common zone” and it’s coming from the closest possible source to ISO committee.
So I keep asking if someone can produce reference from… anything reliable. Draft of the new standard. Textbook written by committee member. Anything.
Until then I believe, we still have reasonable doubt.

 

[DeanD3W]

CH - Your "Now what" figures made me laugh out loud... Good discussion, and very good points. It would be very nice to have one very good standard (set of standards) in the World. Instead we have two separate sets that both need some work. I still like ASME standards much better than ISO, but I would prefer CZ over CF and I think we need a sphere swept profile tolerance zone too. ISO should be restructured with fewer standards, their size tolerances need a default perfect form boundary, they should adopt ASME's feature axis and center plane definitions, and they don't need Concentricity, in my opinion.

Best Regards,
Dean

http://www.d3w-engineering.com

 

[pmarc]

Our discussion in this thread "made me" ask Mr. H.S. Nielsen, Chair of ISO/TC 213, for clarification about this entire topic. We had an e-mail discussion, and this is how I would summarize it:

Say we have a simple square defined with height/width basic dimensions and a profile tolerance without datum feature references...

If profile tolerance is applied between points A & B (where A and B are corners #1 and for example #4 of the square), according to Mr. Nielsen profile tolerance zone will be rounded at corners #2 and #3. This is because the "between" profile specification, as defined in clauses 10.1.1 and 10.1.4 of ISO 1101:2012, makes 3 sides of the square - between A & B - single contiguous feature. The tolerance zone ball is swept along that contiguous feature just like it would be swept if the contour between points A & B was a single line, arc or spline.

If, however, the profile tolerance was applied all around the contour, he admitted that the specification could be read two ways:
1. That profile tolerance frame applies all around, but each of the sides of the square are still independent features, or in other words, there is no orientational and locational relationship between them. (This is in line with upper-right picture of figure 7-3 from his book - profile is applied all around the contour, but each tolerance zone is independent of the others). So if they are independent, there is no issue with the corners, because each independent tolerance zone contains no corners.
2. Similar to the "between A & B" scenario - the all-around specification makes the entire contour single contiguous feature, and then the tolerance zone will have 4 rounded corners.

And there is one more case - profile tolerance applied all around with additional CZ (common zone) modifier inside tolerance frame. This drawing specification "does not create one feature, instead you lock the tolerance zones for all the toleranced features together in nominal location and orientation. Since all the tolerance zones in principle >>go to infinity<< and the workpiece have to fulfill all the tolerances, the practical result is that you get sharp corners, as I also illustrate in my book".

I will refrain from any comments about this at the moment (especially about the part regarding profile tolerance without CZ applied all-around). I just want to give brief and impartial description of what he told me.