should datums have FCF's? 2

[hobbs101]

Hi. A long time ago I was told that the correct way to define ABC datums is as the attached drawing. This ensures that the 3 datums are perpendicular to each other. Is this correct? When I look through GD&T books I don't ever see datums defined like this.

 

[M4C7]

If you are at all concerned about the "accuracy" of your datum features, then its a good idea to apply appropriate geometric controls. I suspect the only reason you don't see them more often is that they are omitted for simplicity as a lot of the drawings are left incomplete by intent. For reference, see Y14.5-1994 Para 4.3.3

"Measurements made from a datum reference frame do not take into account any variations of the datum features. Consideration shall be given to controlling the desired accuracy of the datum features by applying appropriate geometric tolerances."

Attached is a figure from the standard that uses a style similar to that of your drawing to define datum features and apply geometric tolerances to them.

 

[KENAT]

hobbs, as usual it comes back to functional requirements.

Applying controls to the datum surfaces will mean the derived datums are closer to the datum features than might otherwise be the case from picking up the 3/2/1 'high points' on the physical surfaces. Doesn't mean it's necessarily wrong not to have the extra controls though - just that the datums may not as closely match the physical part.

I've assumed ASME drawing stds where rule 1 is in place, might be different for ISO where the surface controls may be more of an issue.

(By the way not that you asked for feed back on other aspects but some minor suggestions... Note 1 - you don't really need to give the reason for the specification and ASME Y14.5M-1994 fundamentalists might go so far as to say it's wrong to do so. It would be better to give a quantifiable value for breaking the edges rather than your polite request. While still not ideal as it doesn't really define the minimum material removal acceptable something like "REMOVE ALL BURRS AND SHARP EDGES .0XX MAX RADIUS OR CHAMFER" might be a bit better. [If you want to get really robust there have been other threads on how to do that.] For a nominally symmetrical part which has controls on the limit of symmetry/alignment due to use of position & surface profile controls it is arguably unnecessary to have all the centering dimensions you give. Don't think you need X 4 OFF just '4X' in front of the Ø9.0.)

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[drawoh]

hobbs101,

What you are doing is correct, but so are lots of other ways of doing it.

Read the chapter on datums carefully. There is no need for your datum faces to be perpendicular to each other. Your datum[ ]C actually is whatever point on that face contacts your fixture face.

I have one quibble with your drawing, although I am occasionally guilty of this too. I interpret your datum[ ]C as the left hand face of your part. When you attach it to the dimension like that, you could be using the 140mm width as your datum.

--
JHG

 

[CheckerHater]

Back in the time, when datums were established by fixturing the highest points of datum features, you could go away without attaching FCF's to datums unless it was functionally required.

Nowadays, when parts are controlled using CMM, that suck at finding highest points, attaching FCFs to datums is quietly made mandatory.

Naturally, it has nothing to do with changes in "the process"

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[KENAT]

CH, doesn't ASME still define it as the high points, and most people just aren't using the CMM correctly?

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[JNieman]

If using the CMM correctly means a touch probe would have to scan 100% of the surface to deduce the highest points from which it then creates a datum feature, I don't want to be right.

Touching "enough" points is "enough".


_________________________________________
NX8.0, Solidworks 2014, AutoCAD, Enovia V5

 

[3DDave]

If there is no perpendicularity requirement then the angles between faces are controlled by the default angle tolerance, so placing a control like that prevents a large, difficult to inspect, accepted variation.

You still need the centering dimensions because you are not locating the features to the centers of a datums, but to the faces of the part. More concerning is that there are eight possible orientations for inspecting this part, only one of which may pass inspection. It would be kind to the fabricator and inspector to either change to using a symmetrical set of datums or to add an asymmetrical feature to indicate orientation.

It is also possible to control the outer boundary with a profile tolerance to [A], and change all the remaining FCFs to use only [A], so that the entire part has a single, simultaneous DRF, eliminating the need for the extra centering dimensions.

As for scanning on a CMM to find high points, that's handled very easily if there are inspection angles available. Set them up and pick points on them as datum simulators to locate and orient the DRF. Then place the part in contact with them, just like the part is placed on the CMM surface plate. Voila, no scanning required. The DRF is available to check as many parts as desired with no further need to re-establish it.

 

[KENAT]

No JNiemean, that would be using CMM stupidly.

Like 3DDave says use gauge blocks or whatever you call them locally.

3DDave is also correct about the centering dimensions as the edge is the datum, I overlooked that.

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[TheTick]

The "correct" way to define datums is the way that matches your requirements.