Least square circle (GG) 2

[naah]

Hi
I am trying to understand the tolerance requirement of GG. I have a hole dimension according to the image below.
I need to know following point.
1. Should I add Roundness (0.15) requirement in dimension when i have GG requirement+
2. What GG means in the dimension. Does it refere to the +- toleranse in dimension? I need your experties help to interperate the GG requirement.

Thanks in advance for all help

 

[greenimi]

I think GG is the global size in ISO GPS so it is applicable only to the size dimension.
Not sure (but I might be wrong) why you would want to add roundness to the size dimension. There are independent from each other and should be achieved separately.

 

[greenimi]

ISO14405-1 I think has some more details

 

[pmarc]

naah,
Like greenimi mentioned, the SCS and (GG) modifiers are defined in ISO 14405-1 "Linear sizes" standard.

The combination of the modifiers means that in a Single Cross Section (SCS) (in your case there seem to be 3 cross-sections labeled D, E, and F), the size of the least-square (Gaussian) circle derived from the actual contour must be within the specified size limits. The requirement applies to both limits because (GG) is specified as the only size modifier.

The roundness tolerance has been added most likely to control form of the feature in cross-sections. Without it, all sorts of fancy/out-of-round shapes could "produce" a least-square circle falling within the specified size limits.

 

[naah]

Hi pmarc

If I interpret the GG tolerance in cross section . See image below. The blue line is Ø51.05 and the dashed line is the plus and minus tolerance (51.05+0.5 and 51.05-0.04)
The green line is the GG tolerance which is whitin the upper and lower tolerance. It should not go outside the tolerance limits. Have i understood this correctly?

 

[greenimi]

The combination of the modifiers means that in a Single Cross Section (SCS) (in your case there seem to be 3 cross-sections labeled D, E, and F), the size of the least-square (Gaussian) circle derived from the actual contour must be within the specified size limits. The requirement applies to both limits because (GG) is specified as the only size modifier.

Pmarc,
Just me trying to learn something from this discussion: are you saying that GG+SCS belongs to the local size category or to the direct global size one?
Not sure I understand that correctly and the differences between them (global size is a nebulous term for me)
Initially, I said that GG (With least-squares association criteria) is the global size, but after reading your replay looks like I am wrong, and, because of SCS indicator, the combo GG+SCS is just the local size….

Could you, please shout some light around those terms….
Thank you very much

 

[TheTick]

It seems the GG modifier stands alone and does not require additional circularity or concentricity.

Always quadruple-check the moment you find yourself reaching for circularity or concentricity. Most likely, you don't need them. (Especially when you really mean "runout")

 

[pmarc]

naah,
I don't see a GG circle in your sketch. If we assume that the green line is the actual/extracted contour of the feature in the cross-section, a GG circle would be a least-squares circle derived from it and the diameter of that circle would have to be within the 51.01-51.10 limits.

Notice that that a GG circle falling within the size limits might be established from the actual contour similar to the one you drawn, but it might as well be established from an actual feature which would be... a square, as an example. This is why a circularity tolerance is needed.

greenimi,
Per Table 3 in ISO 144405-1:2016, the GG+SCS specification is a "local size with least squares association criterion". Figure 29 in the standard shows a very similar example to the one posted by naah.

 

[greenimi]

If we assume that the green line is the actual/extracted contour of the feature in the cross-section, a GG circle would be a least-squares circle derived from it and the diameter of that circle would have to be within the 51.01-51.09 limits.
....
Per Table 3 in ISO 144405-1:2016, the GG+SCS specification is a "local size with least squares association criterion". Figure 29 in the standard shows a very similar example to the one posted by naah.

So, the direct global size it is not defined nor that should have been defined?
Is my understanding correct?
Not all the features defined in ISO should have the direct global size fully defined?

Thanks again

 

[naah]

Hi Pmarc
I have updated the sketch with the GG counters See the yellow line in the image below. As I understood the GG circle could be outside the +- tolerances.
Is it correct that the average of GG value should not exite the nomnial diamter value?

 

[greenimi]

I have updated the sketch with the GG counters See the yellow line in the image below. As I understood the GG circle could be outside the +- tolerances.
Is it correct that the average of GG value should not exite the nomnial diamter value?

I am not pmarc and I am sure he will come with additional explanation, but I would like to say that my understanding is that the GG circle should be within 51.01- 51.10....as pmarc stated: " a GG circle would be a least-squares circle derived from it and the diameter of that circle would have to be within the 51.01 - 51.09 limits."


By the way, why you removed the initial / original specification? Do you have any reason(s)?

 

[greenimi]

naah,

And one more question: What means "Is it correct that the average of GG value should not exite the nomnial diamter value?"

 

[pmarc]

So, the direct global size it is not defined nor that should have been defined?
Is my understanding correct?
Not all the features defined in ISO should have the direct global size fully defined?

greenimi,
I am not sure where all these conclusions are coming from. The direct global size has its definition in ISO 14405-1 and purpose. I would have to paste a solid chunk of standard to explain this. Instead, here is a snapshot of Annex B to the standard on which I marked how the size specification originally provided by OP fits in the global scheme of things.

 

[pmarc]

naah,

As far as I see, the yellow line in your picture is still not a circle, but rather an oval/ellipse. Circle, by definition, is perfectly round.

The specification you had in your original post basically says that once you establish the least-squares circles from the actual contour in cross-sections D, E, and F (a single circle per cross-section), the diameters of the circles must fall within the 51.01-51.10 range.

 

[greenimi]

I am not sure where all these conclusions are coming from.......

Easy answer: from my lack of understanding and education with the ISO GPS system.....that's why I came here to meet you and learn as much as possible.

And since I still would like to learn more about this subject, I would post this follow up question:
GG+SCS combo (from the original post) in my "very uneducated" opinion is a more relaxed than the default specification.

What do I mean?

Lets say, if from the original specification, the “GG+SCS” spec. is removed (keeping the size and concentricity and roundness) then what would be NOW the requirements:

Local size is still controlled within the same size limits, don’t it? And the default (unspecified) association criterion is still the least-square for the local size.

I am saying that because, I think, if only the local size specification is shown, the default association criterion is the total least square algorithm.
I know that for the direct global sizes there is NO default specification, but for the local size it is (least squares).

I hope you do understand what I am asking. If not, please let me know. Again, the scope of my extended conversation is to learn from you……

 

[pmarc]

greenimi,

Let's try this way.

The attached picture comes from ISO 14405-1:2016 Annex D and its original version explains the construction of an opposed pair of points on a cylindrical feature. To help discuss different cases here, I have added a few features (in red) to the figure:

1. If the size specification is ∅50 ±0.1 only, this means that all two-point sizes labelled 10 must be within the 49.9-50.1 range. This is a default interpretation of size specification in ISO GPS.
2. If the size specification is ∅50 ±0.1 (GG), this means that the diameter of the cylinder labelled 2 must be within the 49.9-50.1 range.
3. If the size specification is ∅50 ±0.1 (GG) SCS, this means that the diameter of the circle labelled 6 at a specific cross-section must be within the 49.9-50.1 range.
4. If the size specification is ∅50 ±0.1 (GG) ACS, this means that the diameter of the circle labelled 6 at any cross-section must be within the 49.9-50.1 range.

All 4 scenarios allow for different things to happen with the as-produced cylindrical feature and definitely all 4 require additional geometric controls to be able to say that the feature has been fully defined in terms of its size and form.

 

[greenimi]

pmarc,

Thank you very much for you answer. It is clear now that I don't have a full understanding of the numerous sizes available within the ISO GPS system.
Your points 2, 3 and 4 are understood now.

1. If the size specification is ∅50 ±0.1 only, this means that all two-point sizes labelled 10 must be within the 49.9-50.1 range.

The first point (1) I would like to clarify just a little more: the distance between two opposite points in a cross section perpendicular to the axis of the associated cylinder (the cross section centers should be on the associated cylider axis).
Then, in my understaning, BY DEFAULT, this associated cylinder (number #2 in your sketch) has the association criterion as the total least square algorithm.
So, I concluded that the total least square is the default for the two point size.
Looks like you don't disagree with my above conclusions. Don't you?

(But I fully agree with you that I have some FUNDAMENTAL missunderstandings of the ISO GPS)

 

[pmarc]

greenimi,
The total least-squares cylinder is an associated feature that is the same in all 4 cases that I described. So yes, it is a kind of default.