Referencing Centerline Datums

[PRuggiero]

Guys,

Ran across a situation that I am unsure of. Is what I show in the picture legal? Specifically can I call out datum D (plane representing the centerline of that dimension) and then use that same datum in the control frame for a profile of that same feature?

I was asked to do something like this and it doesn't seem right I just can't verbalize why I think its illegal. For one, if a fixture was used to test the part it would be hard to check that profile since the fixture would have a feature that fit into the square hole.

I understand that this part is missing some stuff, I just used it to make an example.

Thanks,
Pete

 

[pmarc]

I'd like to add my 2 cents to this:

1st cent - datum self-referencing
I keep thinking why Jim is trying to covince us that OP's sketch (with D instead of C in profile FCF) makes sense and is legal. And I think I am starting to see a logic behind. Jim, please correct me if I am wrong, but is your interpretation following; datum plane D is derived from cutout width which is in fact a two-opposite-planar-faces feature of size. But the profile callout controls each face independently, so in fact it is not that the FOS is being controlled to itself. Is it what you are trying to say, Jim?

2nd cent - relationship between datum features
I absolutely agree that the relationship between D and C should be somehow specified, and I would most probably go with profile for D since position applied to basic .11 does not look good (although we could argue whether it is legal or not).
Plus, I am again with J-P - this time that Tec-Ease tip is not reflecting the situation we have here and that datum plane D will always go together with datum feature D because nothing else contrains the part's movement in vertical direction during setup for profile callout inspection.

P.S.: I wonder when we come to a conclusion that it would be much easier for everybody if datum D was not here at all and only profile callout wrt C|B|A (or composite profile FCF with lower segment tightening relatioship to B or C) was specified.

 

[pmarc]

There should be A|B|C in the last sentence, sorry.
C|B|A was in the other thread about controlling symmetry

 

[CheckerHater]

It still looks disturbing to me.

Too many things are missing from the picture.

For example relationship between datums [C] and [D].

Should it be explicitly specified? Because otherwise General Tolerances (you know, +-5 degrees) will imply; won't that give the feature too much freedom?

 

[axym]

Here are some thoughts.

In general, I don't like the idea of FCF's in which the considered feature is referenced as a datum feature (or as part of a multiple datum feature). I would say that these applications are generally flawed and are seldom, if ever, functional.

Jim, I know of the runout application you referred to - where the datum features are included in the runout tolerance. There's an example in the standard that shows it - where the datum axis is defined using A-B, and features A and B both have runout tolerances to A-B. I think I've ranted on this one before. I understand that there is nothing impossible going on here, that A and B could be checked back to A-B using a CMM or with some awkward chucking technique. But I don't think that the concept makes sense. A and B just have to have good runout to some axis - it doesn't have to be axis A-B. But there's currently no way to say that in Y14.5.

In the OP's example, I would say that the reference to datum feature D doesn't add any value. Referencing A and B would suffice. The profile zone can float in the X direction parallel to B, and if we can find some location for it that allows the feature to pass then we're good. The profile zone doesn't need to be centered on the feature's oriented AME - this is an unnecessary restriction and is very likely not functionally required.

That said, it is sometimes convenient to add extra restrictions like this when using CMM's. Some software packages do not work well with FCF's that leave degrees of freedom open - they like to have a fully constrained DRF. So it often makes sense to fill in any open DOF's using the considered feature. This gives the CMM software the fully constrained DRF that it needs, without introducing artificial constraints to other features.

But these extra restrictions do not belong on the drawing.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[Belanger]

Forgive my crude sketch (see attached graphic), but I was trying to noodle Jim's idea from last night about two profiles-- one tying down the location of D back to C, and a refinement to then have the slot refer to D.
I still don't think any value is added by mentioning D (see the lower picture on my sketch), but maybe I'm off on a weird tangent from what you intended, Jim. Either way, I'm stumped as to why it's ever logical to have a self-reference of any kind.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[MechNorth]

Pmarc, yes, that's what I'm getting at.
For the need to locate the simulator at its basic location, pls reference datum C in Fig. 4-18 in '09, in particular the last sentence in the note on the top right. Similar in Fig 4-15, -17. This contrasts with the situation for a datum translation modifier where the simulator gets to move to make the best fit.
J-P, I was thinking more of a similarity with cast & machined datums; a feature is realted to the cast drf and to the finish-machined drf with a refinement.
As for whether or not a datum feature can be controlled wrt the datum established by it, please someone explain why not, as opposed to it just not looking or seeming right. As the most extreme case(?), consider a workpiece with nothing but a general surface profile control (allover). The datum features are thus controlled by the general control, are tney not?
Good discussion(s), gents.

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[axym]

I'm with J-P on this one. The self-reference is not illegal in the standard, but I don't see the usefulness.

I racked my brain and thought up a scenario in which the self-reference would actually make a difference (see attached sketch). The surface of the cutout is both tilted and asymmetric, but will fit in the profile zone if optimized in the X direction. If the self-reference is added, the zone must be centered on Datum D (the surface's oriented actual mating envelope). With this extra constraint added, the feature no longer conforms. Darn - I just noticed that I drew the simulator lines on the outside of the cutout and they should be on the inside. But you get the idea - just imagine that it's a boss instead of a cutout ;^).

So the self-reference provides a control that is real and definable, and does make a difference in certain cases. But I'm having a hard time envisioning a functional situation that would require this.

Evan Janeshewski

Axymetrix Quality Engineering Inc.
www.axymetrix.ca

 

[PRuggiero]

axym, in your first case where you only call out A and B in the FCF, the cutout isn't being controlled in the "x" direction (relative to C) at all correct? Meaning the zone could be float from one side of the part to the other.

 

[axym]

PRuggiero,

That's correct. In the first case, the cutout isn't being controlled in the X direction at all. This would make sense if D was intended to define the X location of other features. For example, feature C might have a profile tolerance to ABD.

I fixed the error in my earlier sketch, and the result changed (see updated sketch). With Datum D established from the correct side of the surfaces, the diagram no longer shows the effect I was trying to show ;^(. The cutout still conforms to the profile zone, even with the constraint to Datum D. I'll have to make a new sketch and cook the surface geometry differently.

Evan Janeshewski

Axymetrix Quality Engineering Inc.
www.axymetrix.ca

 

[axym]

Here's a new sketch with different surface geometry. In this example, the self-referencing makes a difference.

Again, I struggle to see an application in which the self-referencing would be functionally necessary.

Evan Janeshewski

Axymetrix Quality Engineering Inc.
www.axymetrix.ca

 

[Belanger]

Good stuff, Evan. Jim, based on his example I guess I'm forced to soften my position somewhat. There are times when a datum derived from a feature might be referenced in a FCF controlling that feature, and indeed have an impact. The actual usefulness of such is still up for grabs, but I get what the sketch is showing.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[MechNorth]

Good sketch, Evan. Tks.
Tks, J-P. I don't recall the details of the applications I've used it for; too many clients back to remember all of their applications. A general surface profile is the easiest example that comes to mind for a practical use.

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[axym]

I'm not sure I understand the Circularity comment either. The Y14.5 definition of the tolerance zone is relatively straightforward, but evaluating Circularity in a repeatable way on an actual feature can be very difficult. This is especially true if the feature is warped or bent.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[axym]

Oops, I somehow posted the last comment into the wrong thread. My fault for trying to do this on a BlackBerry!

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca