Spherical Datum

[Gray76]

Hi All,

https://files.engineering.com/getfile.aspx?folder=c66ceef3-070b-44d2-b455-30a6d88222a7&file=BALL.pdf

I'm looking for some help on the component sketch above.
Over the years I've seen many different approaches and would like to finally bottom the issue.
The component is a trunnion valve ball and is supported top & bottom by trunnion/stem pins.
The pockets need to be central about the centre of the sphere and the large through bore must be concentric with the sphere.
Does datum A infer a centre point and three planes? Is there any need for Datum B, or can all the features be related back to Datum A?


I would be very grateful of any help you can give.

Thanks

 

[chez311]

Gray76,

Is this drawn per Y14.5-2009? If so, I'm surprised by your liberal use of concentricity/symmetry. By an initial rough evaluation I would think position would be a better choice in both cases (possibly with orientation refinements) - these controls have very limited use and its typically preferred to utilize some other control. Note that as of Y14.5-2018 concentricity/symmetry are no longer supported and have been eliminated.

You might want to consult Y14.5-2009 fig 7-57, 7-60 through 7-63, 7-65 and 7-66 for some of the allowed variation of concentricity/symmetry in comparison to position - the behavior may not be what you expect or desire for this application.

The datumless parallelism callouts on either dimension of the rectangular slot are also questionable. If the desire is to either control flatness of the surface(s) or the derived median plane, then it should be replaced as such. I don't know if its technically illegal, but I would certainly flag those as problematic as they are not serving their primary function (orientation to a DRF).

Note addtionally that the parallelism to the 2x dimensions of the flat faces to the centerline referencing |B| (bottom right view) are not valid as shown. The non-basic dimensions implies directly toleranced (presumably to some general print tolerance) dimensions which is not recommended for location of features, and not really supported by Y14.5 as shown since it is location to a theoretical centerline. Doubly so since the parallelsim callouts are inline with the dimension which implies control of a center plane, which is not valid since it is not a FOS. If the desire is to control parallelsim of each surface separately, the leaderline should not be inline with the dimension. I would also recommend either changing this to a basic dimension and controlling location with profile or a directly toleranced dimension which spans both surfaces and specified with position.

At first brush, utilization of the central bore as datum feature B looks fine to me. Spherical datum feature A constrains all 3x translational DOF (Y14.5 discusses this as a point where 3x mutually orthogonal planes intersect) and B constrains 2x rotational DOF. This allows translation/orientation constraint to |A|B|.

That being said, with the above changes (concentricity/symmetry -> position) I would additionally hold all but the central bore (which would be of course only to |A|) to |A|B| in position. The 2x bores could be held in 2x position to ensure coaxiality to each other as they would be a pattern. Position to |A|B| would also ensure simultaneous requirements and therefore mutual location/orientation between the rectangular slot and the 2x coaxial bores *since there would be a free rotational DOF. Alternately the width of the rectangular slot could be established as datum feature C or the 2x coaxial diameters could be C to constrain that remaining rotational DOF - depends on what you think best reflects the assembly condition.

Side note - I know the dimensions have been likely removed either for generalization purposes or to protect your design, but it would make it a little easier to reference certain features if they had an approximate dimension attached instead of "999".

*Edited

 

[greenimi]

How do we know if it is ISO or ASME?

 

[chez311]

greenimi,

Hence why I asked in my first line. I was on a roll so I kept going. I think most of my recommendations still hold regardless, but I had a hunch based on the lack of ISO specific modifiers.

 

[Gray76]

Hi,
Thanks for your reply chez311. This is a typical drawing at a company I recently joined.
It is not to any specific standard and many of the existing drawings do not have geometric tolerance.
We certainly need to improve the drawings hence the question.
Can you give me an example of how to constrain the coaxial diameters & rectangular pocket to the sphere as I don't have a copy of Y14.5.

Many Thanks

 

[chez311]

Gray76,

Ah I see now, I think you're putting the cart before the horse. I would first decide along with your peers which standard you wish to adhere to (with some consideration in what make sense with your suppliers/customers), purchase said standard, and ultimately I would recommend investing in some sort of learning program for all those involved with creation and reading of said drawings.

To me, the concepts in these standards don't really make sense to apply "generally" without adhering to a particular standard. If there is no intent to adhere to a standard, I probably wouldn't even bother with any kind of geometrical/GDnT controls as there would be no way to enforce them or interpret exactly what is meant without speaking directly to the designer. In this case age-old tolerancing methods and custom notes would probably work just as well.