converting a spherical polar coordinate tolerance system(!) to GD&T

[Matt137]

Hi all,
Here's a brainteaser for you. I've got a spun metal dome, which has a fancy profile laser-cut out of it. The spherical diameter of the dome is +/- 0.5mm; the profile needs to be cut out +/- 0.2mm. I know that sounds weird (why are you worried about the position of the profile if the dome is out by more?) - it's because there's a sensor at the centre of the dome, so the actual dome diameter is less important.

At the moment this is managed by a series of jigs, and no drawing exists, which makes me itch. I'm trying to get this system to fit into GD&T, but I'm struggling. My first attempt was to (1) put a tolerance on the dome's diameter and roundness, (2) then make the dome into a datum and then (3) put a profile tolerance on the cutout (oh, and the cutout is a really weird shape which is never going to be amenable to individual dimensions). But my drawing checker is objecting to having a spherical surface as a datum. I could put a tolerance of +/-0.2mm on everything, but then we'd probably have to machine it, which would be bonkers expensive and completely unnecessary. So, thoughts? Anyone got any other ideas how I could tolerance this?

 

[AndrewTT]

If your current mfg system if producing usable parts then your GD&T drawing could follow/imply this order of operations. Datum targets or a region of the sphere could be used to establish your DRF (instead of the entire sphere). These would be chosen based upon how you are holding the dome to cut (and/or measure) the irregular shape. You would need to make sure that this GD&T scheme would not allow for non-functional parts (since the scheme is based upon mfg. and measurement instead of function).

Is the dome a full sphere, a half sphere, or less than a half sphere?

 

[3DDave]

Not everything is reducible to a set of symbols developed to handle common characteristics. In particular there's nothing in Y14.5 definitions for Feature Control Frames (if that's what you mean by GD&T) that covers angular/wedge shaped tolerance zones for profile.

If you are able to write a note that describes the intended result, that's good enough. If that's not possible then you will never be able to substitute a symbolic description.

 

[pylfrm]

The spherical diameter of the dome is +/- 0.5mm; the profile needs to be cut out +/- 0.2mm.

+/- 0.5 mm and +/- 0.2 mm relative to what?

How does this part interact with the rest of the assembly or system?

How does the series of jigs work?


pylfrm

 

[Matt137]

Thanks for these responses!
One reason why I'm interested in getting a GD&T scheme running for this is that the current way of testing is highly subjective, and I'm trying to move us to using an CMM for increased speed and repeatability... so I'm trying to communicate our requirements to the CMM suppliers in a way they can quickly grasp. Also, we are using multiple manufacturing routes for making these parts (which are half-spheres with a flange round the rim btw), and we don't really have much of a handle on the process capability of the different methods. So, you know, the usual clusterf*ck.

pylfrm: the jigs are mostly "by eye" comparisons, set up by a scientist, and frankly to my engineer's eye they are not really fit for purpose. The profile control on the laser cut, for example, is a visual check against a (flat) "good" template, so the potential for parallax errors makes it a bit of a farce.

3DDave: I take your point, but if there is some way of doing it via Feature Control Frames then that is in my experience much less ambiguous. It's a tiny bit reassuring to hear there's no clear answer (yay! I'm not an idiot), but also disappointing, obviously.

It may be 'grammatically' incorrect to have a spherical face as a datum, but I still think my current arrangement is the least ambiguous option. To me, anyway - the question is, will other people understand it?

 

[chez311]

Matt137,

As far as I know a spherical datum is technically allowable, see Fig 4-3 in ASME Y14.5 and I would think it would have to subscribe to the same rules as a round feature to classify as a FOS datum (180deg opposed points - which your half sphere satisfies). That being said, for inspection purposes I would agree with AndrewTT's point about datum targets - it would likely simplify your setup a great deal.

 

[Matt137]

Hi chez311,
I totally did not spot AndrewTT's suggestion! Thanks for the nudge Yes, that does sound better, I'm sure I could arrange that. I'll have another look at the standard, but it sounds doable.

Brilliant, thanks all.

 

[drawoh]

Matt137,

Weird, non-orthogonal geometry sounds to me like an application for profile tolerances. How do you plan to inspect this thing?

You can model weird, curved objects by with radii in 3D[ ]CAD. This makes it easier to create conventional drawings. They can also be modeled in 3D by drawing splines. These can be called up as per Fig[ ]1.31 and Fig[ ]1.32 in ASME Y14.5M-1994, or Fig[ ]1.33 and Fig[ ]1.34 in ASME Y14.5-2009. If this is not your design, I would recommend the latter.

Does the car industry produce dimension drawings of curved body parts? You can model your inspection fixtures directly from your 3D[ ]model, and call up profile tolerances.

--
JHG