A ball and GD&T 2

[Diametrix]

I have a ball from a manufacturer with the following specs: Diameter - 0.500+/-.005" and "sphericity" - .005" Since the term sphericity is somewhat vague I wanted to get away from the notes describing the dimensions and show it with GD&T, however it seems that GD&T fails me here. The way I understand it, I can't use a composite profile tolerance here because there is no locating dimensions. It is almost like I need a basic dimension with tolerances (don't throw rotten tomatoes at me here) and to have a profile tolerance... which naturally is not possible. Any ideas how to show something like that on a drawing without a verbal description?

 

[Garland23]

Profile of a surface is a natural choice for a sphere. However, profile typically requires a basic dimension for the diameter, because it wants to control the form and size of the sphere. It sounds like "sphericity" is just a form idea -- much like circularity but wrapped in a spherical manner.
There is a modifier called dynamic profile, which I believe might be close to what they want, although that's shown in the standard as part of a composite profile tolerance.
Let's see what others have to say, though.

 

[greenimi]

Diametrix,
ISO GPS or ASME?
I think the answer are more relevant if you specify the system.
Just stirring the pot here.....

 

[Diametrix]

Thanks, that's fantastic! That's exactly what would work here. Too bad we are using 2009 standard

 

[Diametrix]

ISO GPS or ASME?

ASME

 

[Wuzhee]

Some resources call circularity applied to a sphere as Sphericity.

I'd say, a simple circularity is enough for form control. Not knowing how your ball looks like (bearing ball? door knob?) the simplest solution might be the best.

 

[Diametrix]

Yeah, I'm looking at a simple ball, like one for a ball bearing. Thank you, that works too!

 

[Garland23]

Wuzhee, I'm not sure if that provides the intent of sphericity.
That's because "each circular element" could be a different size. So the various rings will all be within 0.25, yet put them together and it's not a spherical shape of 0.25 accuracy.

 

[Wuzhee]

ASME don't have a tool for it. I searched for ball bearing tolerances but the ball itself is very rarely available.

We're doing ball valves with only size tolerances. They work pretty well. Rule#1 gives it circularity so I'm pretty sure it's enough to control.

For a bearing ball I'd guess they have circularity|0.001| or similar tolerances. But I'm no bearing manufacturer, better ask FAG\SKF to give a hint.

 

[Diametrix]

Wuzhee, I'm not sure if that provides the intent of sphericity.

I think it does. The key word there is "the common center", meaning that each circular element has a common diameter. That common diameter can vary within the size tolerances but the rest of the points for each circular element must lie within the profile tolerance zones.

 

[Diametrix]

Rule#1 gives it circularity so I'm pretty sure it's enough to control.

The "ball people" usually have grades for balls, which include loose size tolerances and tight sphericity tolerances. Rule #1 would not let you match that.

 

[Garland23]

Diametrix, I agree with you that they pass through a common center. But as you said, each slice can vary in size. So if you put all slices together in a 3-D sphere shape for one tolerance, would the "sphericity" be 0.25 or would it be 0.8?

 

[Diametrix]

So if you put all slices together in a 3-D sphere shape for one tolerance, would the "sphericity" be 0.25 or would it be 0.8?

0.25
Divide the ball into four quarters, like an orange. This will form a common axis (diameter) for each circular element mentioned above. This diameter can be 25.4 max to 24.6 min, but it is going to be the same for all circular elements. Now each circular element will have three things in common with the rest of them: one actual diameter (D) and two tolerance zones. Each tolerance zone diameter is going to be D + .125 and D - 0.125. All the points of each circular element have to fall between these two tolerance zones and not violate max and min diameter constraint(for example a circular element can't be the outline of the bigger tolerance zone in the illustration below)

 

[Garland23]

Now each circular element will have three things in common with the rest of them: one actual diameter (D)

This is where I am having trouble seeing your viewpoint. Why must they all be the same size?

Perhaps we could back up to a simpler example of circularity of a regular cylindrical part (Fig. 8-10 of the 2018 standard). You seem to be saying that each cross-section where I measure circularity must have the same size/diameter. But that is not true, even if their center points all connect perfectly.
So when we transition to a sphere, why is it reguired that every slice where I measure circularity have the same size/diameter?

 

[jassco]

Why do you make this an issue? Circularity feature control frame controls "Sphericity" when it is applied to a spherical feature (dimension with "S" as prefix).

Best regards,

Alex

 

[Garland23]

Alex, it's an issue because people are permitted to legitimately disagree with each other. The misunderstanding is probably on my end, so I'm honestly trying to understand this.

To your point: the size dimension is preceded by S, but the geo tolerance is not preceded by an S. That's the difference, and that's why I'm thinking that the many circular sweeps around the sphere don't have to all conform to the same size.

 

[jassco]

This is because spherical tolerance zone of a spherical surface is wall thickness of a hollow sphere.

Best regards,

Alex

 

[Diametrix]

Why do you make this an issue?

I agree with Garland here, this is how I learn too. The process might seem somewhat adversarial, but unless you try to defend your position quite often you won't get to the point of understanding where you where wrong. Look how much I've learnt from all of you guys about a simple ball!

You seem to be saying that each cross-section where I measure circularity must have the same size/diameter

Not at all. The example from the standard is quite clear - each circular element of a cylindrical part can be any diameter within the specified limits

The difference is that those are "lengthwise slices" and have no common points. When you slice a ball along the same axis you are going to have two common points for each cross section. In other words each cross section is going to have at least one diameter in that cross section that is common for all of them. However, that's where I was wrong. There are two extreme cases here. You can build the tolerance zones here in two ways - one cross section can have the internal circle of the tolerance zone touching the common diameter and another cross section can have the external circle of the tolerance zone touching the common diameter.

So, it appears the actual ball surface profile tolerance is going to be 0.5, which is twice that of the circularity stated on the drawing.

 

[jassco]

You just need to interpret it per its definition in the standard. It is rather simple mathematically.

You have two spherical surfaces that are concentric. You expand the smaller one until it touches the spherical. You also contract the large one until it touches the spherical surface. The thickness or normal distance between the two spherical surfaces is roundness or sphericity.

Best regards,

Alex

 

[Wuzhee]

The "ball people" usually have grades for balls, which include loose size tolerances and tight sphericity tolerances. Rule #1 would not let you match that.

I was specifically talking about our balls. We only have a -0.04 size tolerance on the balls without additional geometric control and they don't leak.
For a ball in a pre-load ball screw, I'm sure a size tolerance only won't work.