Validity 4

[aniiben]

Q1 - referring to a datum feature in a way that it was not initially constrained is a potential problem for understanding the condition. See previous discussion where I make clear that Figure 4-16 (c) is clearly evaluated incorrectly in the standard.

Inspired by a statement made in a previous discussion and also by some of my misunderstandings of the standard’s intent on dealing with datum features of size called at RMB and also at MMB, I would like to ask the members of this forum how would you consider the datum reference frame shown in the embedded picture.

Case 1: A│B│C(M)│
----Datum feature B at RMB (shown in the picture) ---

Is this datum structure or DRF valid? If yes, are there any issues with the violation of datum feature precedence order?

Case 2: A│B(M)│C(M)│
----Datum feature B at MMB (NOT shown in the picture) ---

Same open ended questions as for case#1.

 

[chez311]

But at the end of the simulation process when the part is fully fixtured, the answer to the question which degree of freedom is constrained physically by which datum feature - datum feature simulator interface, is dependent on the exact location of applied force vector or the axis of the applied torque, and the location of the datum reference frame axis system - which was described in this forum as arbitrary and non-critical.

This was sort of along the lines of what I was thinking, however since there was a distinction made between how |A|B|C>| and |A|C|B>| (identically) vs |A|B|C| and |A|C|B| (not identically) constrain [w] I was curious about why.

 

[pylfrm]

1) Why would you say the same logic doesn't hold for |A|B|C| vs |A|C|B|, just because the tertiary datum feature simulator is fixed and cannot expand to its maximum diameter? And are you saying that while maybe not "identical" as in |A|B|C>| vs |A|C|B>| that BOTH B and C play SOME role?

I wouldn't say there's much difference. Both the secondary and tertiary datum features in |A|B|C| or |A|C|B| play a role in constraining [w]. It could be argued that they do not play equal roles because swapping their precedence can change the result, but that's probably just semantics.

2) It seems to me that utilizing the logic you laid out, that for |A|B|C>| vs |A|C|B>| that it would not only constrain [w] identically but also [x,y], right?

Why do you say that? pmarc just explained how the constraint of [x,y] can differ. I'm not really sure what logic you mean.


pylfrm

 

[chez311]

pylfrm,

My apologies, for some reason I had some difficulty wrapping my head around that. I see now what you meant - |A|B|C>| and |A|C|B>| constrain [w] identically because the translation modifiers allow the tertiary datum feature simulator to translate and fully expand in both cases, therefore resulting in the same angular constraint of [w] when they are flipped, even though translation constraint of [x,y] could be different as pmarc showed. In the case of |A|B|C| and |A|C|B| the lack of translation modifiers means that depending on the as-produced variation between the secondary/tertiary datum features the datum feature simulator for the tertiary datum will not necessarily fully expand and could result in different angular constraint of [w] as well as different translation constraint of [x,y]. Thank you for pointing that out.

https://www.eng-tips.com/viewthread.cfm?qid=451664[/URL]]Consider the position tolerance applied to the pattern of four holes in ASME Y14.5-2009 Fig. 4-19. If we assume datum feature A constrains [u,v,z], then datum features B and C actually play equal roles in constraining [w].

Although I admit it can be a useful way to think about things, I disagree with pmarc's assertion that [w] represents rotation about a specific axis. If rotation is constrained about any one axis, it is simultaneously constrained about all others parallel to that.

Would you hold the bolded portion as true across the board in all cases, not just 4-19, such as 4-9?

Also what do you think about the below quote from Sem? Is the discussion of DOF constraint as arbitrary as the discussion of placement of coordinate systems/origin/datum planes? I know it discusses metrics like torque and force - substitute in their geometric analogs such as vectors/rotation/translation if you like.

But at the end of the simulation process when the part is fully fixtured, the answer to the question which degree of freedom is constrained physically by which datum feature - datum feature simulator interface, is dependent on the exact location of applied force vector or the axis of the applied torque, and the location of the datum reference frame axis system - which was described in this forum as arbitrary and non-critical. Take for example the |A|B|C| DRF from fig. 4-9. If the DRF axis system origin can be "on the moon" as it was said, it may as well be located on the center axis of datum feature C. In that case, if a force is applied in the [X] direction through the axis of datum feature C simulator, datum feature and datum feature simulator C (the "clocking" elements) will physically constrain translation in [X]... yes, at the |A|B|C| DRF! It seems like all this can be added under the "unnecessary distraction" category brought up in the other thread.

 

[pylfrm]

Would you hold the bolded portion as true across the board in all cases, not just 4-19, such as 4-9?

Yes.

Also what do you think about the below quote from Sem? Is the discussion of DOF constraint as arbitrary as the discussion of placement of coordinate systems/origin/datum planes? I know it discusses metrics like torque and force - substitute in their geometric analogs such as vectors/rotation/translation if you like.

I think the discussion of forces and torques is out of place for what should be a matter of geometry. I don't think "the location of the datum reference frame axis system" is relevant either.

I'd say DOF constraint is a less arbitrary and generally more useful subject than datums.


For ASME Y14.5-2009 Fig. 4-9, assume [x] is in the direction of the 24 dimension (left/right) and [y] is in the direction of the 40 dimension (up/down).

Looking at it in terms of a sequential process: When only the primary datum feature is considered, [u,v,z] are constrained. When the primary and secondary datum features are considered, perhaps you could say that [x,y] are also constrained, but only along the axis of the secondary datum feature simulator, and are unconstrained everywhere else. When the primary, secondary, and tertiary datum features are considered, [w] is also constrained, and [x,y] become constrained everywhere.

Looking at it in terms of the end result: The primary datum feature alone constrains [u,v,z]. The primary, secondary, and tertiary datum features are all involved in the constraint of [w,x,y].

The customized DRF notation is better aligned with the sequential process view. This may have some disadvantages, but I imagine it generally gets the job done.


pylfrm

 

[Sem_D220]

I don't think "the location of the datum reference frame axis system" is relevant either.

That is true only as long as everyone agrees that w is rotation about any axis parallel to Z, u is rotation about any axis parallel to X, and v is rotation about any axis parallel to Y, just as translation at X direction is caused by translation vector applied anywhere parallel to X, etc. Apparently, there is no such convention.

Looking at it in terms of the end result: The primary datum feature alone constrains [u,v,z]. The primary, secondary, and tertiary datum features are all involved in the constraint of [w,x,y]

I agree. That was also the main point I wanted to convey in my post.

Edit: the only objection - I don't understand how the primary is involved in constraining w, X, Y.

 

[Sem_D220]

Note that by "w" I actually mean rotation about any axis parallel to the B axis (obviously).

And that's the whole problem. As long as you think of "w" this way, B will be always constraining rotation.

But that's not how it should be interpreted - "w" is specifically a rotation about datum axis B and not about any axis parallel to B.

In such cases I like to say that the origin of the CSYS could even be located on the Moon provided that there is a basic location relationship between the Moon and the used datum feature simulators.

pmarc, I was convinced by each of these statements separately when they were posted but trying to understand the meaning of them in conjunction with each other I can't help but find them contradicting. Y14.5 defines w rotation as rotation around Z axis of the datum reference frame, not about any specific individual part-related datum feature. For the part in fig. 4-9 suppose that the "default" non-mandatory location of the datum reference frame shown in the "means this" portion shows axis B coincident with the Z-axis of the DRF. If one is free to make a different choice and put the DRF axis system on the moon, how is datum feature C supposed to constrain w rotation (edit: if constraining rotation around an axis offset and parallel to Z doesn't count as constraining w)?

 

[chez311]

pylfrm,

Thank you for laying it out so plainly, I now believe I have a pretty good understanding of your thought process is in regards to DOF constraint. If I can, I'm going to attempt to answer Sem's question below - please feel free to correct me if I'm on the wrong path or add in clarification if you believe I missed something.

the only objection - I don't understand how the primary is involved in constraining w, X, Y.

If I am following pylfrm's reasoning correctly, until a given DOF is fully constrained everywhere (ie: not just along/around the single axis of a datum feature simulator but also all axes parallel to it) all higher order datum features up to that point can be considered to be involved. In the case of Y14.5-2009 4-9 [u,v,z] are fully constrained along/around all parallel relevant axes by A, so only A is involved in their constraint. [w,x,y] are not fully constrained along/around all parallel relevant axes until C is added, so C as well as all the higher order datum features A and B are involved in their constraint.

Hoping I have not made any errors in the above, let me take a stab at another example which has a slightly different datum feature precedence. Take Y14.5-2009 4-24 but imagine that the 8.0 through hole is specified as datum feature C and let us consider another feature (not pictured) which is controlled by A|B|C. Assume the convention in the view shown is that [x] is in/out of the page, [y] is up/down, [z] is left/right. A only fully constrains [u,v] everywhere while only constraining [x,y] along the axis of the datum feature simulator for A. The inclusion of B fully constrains [z] everywhere. The addition of C fully constrains [w] everywhere as well as adds the additional restriction necessary to constrain [x,y] everywhere.

In this case we could say the following is the end result:
The primary datum feature A alone constrains [u,v]. The primary and secondary datum features A|B are involved in constraint of [z]. The primary, secondary, and tertiary datum features A|B|C are all involved in constraint of [w,x,y].


Hopefully I am not too far off base!

 

[Sem_D220]

chez311, I can not say that I fully grasp the concept you described it, but it is interesting. I suppose it will be up to pylfrm to confirm or deny.
I guess I will just cling to my own perception which is that constraints of degrees of freedom as meant by the standard are such that each type of datum feature can only be involved in constraining movements in directions at which its' datum feature simulator can embody a direct physical barrier: a planar datum feature simulator can only constrain translational degrees of freedom at the direction normal to it, etc. I believe it is so both during the sequential process and at the end result. That's why even after the explanation - I have a hard time to relate to the idea that primary datum feature A in fig. 4-9 can be involved in the constraint the X DOF (given that X is parallel to the theoretical surface) or that datum feature A in the modified fig 4-24 can be involved in the constraint of w given that Z is the axial direction (I'm referring to the suggestion that "The primary, secondary, and tertiary datum features A|B|C are all involved in constraint of [w,x, y].")

 

[pylfrm]

Edit: the only objection - I don't understand how the primary is involved in constraining w, X, Y.

The constraint provided by |B|C| would differ from that provided by |A|B|C| for all six degrees of freedom.

The direction of [x] is defined by the constraint of [v,w]. The direction of [y] is defined by the constraint of [u,w]. The direction of [z] is defined by the constraint of [u,v].

If I am following pylfrm's reasoning correctly, until a given DOF is fully constrained everywhere (ie: not just along/around the single axis of a datum feature simulator but also all axes parallel to it) all higher order datum features up to that point can be considered to be involved.

Although I agree with your post otherwise, I wouldn't say all higher-precedence datum features are involved in all cases. For example, consider the inner ring of a spherical bearing such the following:

Call the spherical surface (dimension dm) datum feature A, and call the width (dimension B) datum feature B. Imagine a position tolerance referencing |A|B>| is applied to the bore (dimension d). When only the primary datum feature is considered, [x,y,z] are constrained, but only at the center point of the datum feature simulator. When the primary and secondary datum features are considered, [u,v] are also constrained, and [z] becomes constrained everywhere. Because of the translation modifier on the secondary datum feature reference, the primary datum feature does not contribute to constraint of [u,v].


pylfrm

 

[Sem_D220]

pylfrm, can I conclude that since datum feature simulator A for the spherical component of the bearing doesn't impose a basic location or orientation relationship on the movable datum feature simulator B, it is not involved in the constraint of the degrees of freedom that start to become constrained only with the involvement of the lower precedence B simulator?

 

[3DDave]

Sem,

Following the rules in Y14.5, that's exactly what you can conclude.

Unpacking the logic of your question:

If the none of the members of the set of constraints that are controlled by A are in the set of constraints controlled by B, then the constraints controlled by B only apply with involvement of B.

You had answered your own question.

 

[Sem_D220]

If the none of the members of the set of constraints that are controlled by A are in the set of constraints controlled by B, then the constraints controlled by B only apply with involvement of B.

That can also be said about fig. 4-9:
The set of constraints controlled by A is u, v, Z.
The set of constraints controlled by B and C: when only B is involved - X, Y, only through the axis of B, once C is involved - X and Y everywhere, w. None of these are controlled by A - at least when only A is involved.
However: "The primary, secondary, and tertiary datum features are all involved in the constraint of [w,x,y]" (pylfrm, about fig. 4-9).

 

[3DDave]

B and C each can control x, y, u, and v; therefore the case is different for 4-9 than the spherical bearing example.

In the spherical bearing case A INTERSECT B is {}, the null set. In the case of 4-9 A INTERSECT B = {u,v}

At the least, all three datum features are required to constrain all six degrees of freedom and eliminating any of them results in at least one degree of freedom being unconstrained.

 

[Sem_D220]

B and C each can control x, y, u, and v

This is why for fig. 4-9 I would say that once A, B and C are all involved - u and v are constrained by A B and C together. However, there are different opinions:

Looking at it in terms of the end result: The primary datum feature alone constrains [u,v,z]

Perhaps you missed the fact that my objection was regarding the involvement of the primary datum feature A in the constraint of w, X, Y:

The primary, secondary, and tertiary datum features are all involved in the constraint of [w,x,y]

If you eliminate the primary, I would say that those degrees of freedom are unaffected.
The only way I can understand this is that the primary has an influence on the constraint of w, X, Y indirectly by constraining the orientation of the datum feature simulators of the lower precedence datum features which in turn constrain those degrees of freedom. Hence my question to pylfrm to assure this.

 

[chez311]

The direction of [x] is defined by the constraint of [v,w]. The direction of [y] is defined by the constraint of [u,w]. The direction of [z] is defined by the constraint of [u,v].

I think this is the key here. Ie: [x] cannot be fully constrained "everywhere" until [v,w] is constrained, same with [y] and [z]. Likewise a datum feature that constrains [v] or [w] is also involved in fully constraining [x] (again, rinse/repeat for [y] and [z]).

It seems to me that we can say something similar, with some restrictions, about the converse. Unless [v] or [w] has been constrained by a higher order datum feature, a datum feature that constrains [x] is also involved in fully constraining [v] or [w] (respectively). Again with the applicable rotational DOF with [y] and [z]. It took me a bit to figure out why I couldn't make a blanket statement about this as a result of the translation modifier - as this applies in the case of 4-19 with the translation modifier (higher order datum feature B which constrains [x,y] is involved in the constraint of [w] when C is applied) but not in your example of the spherical bearing (higher order datum feature A which constrains [x,y,z] is not involved in the constraint of [u,v]).

What I came up with is if you have a datum feature with a translation modifier which can constrain remaining rotational DOF of interest by itself (which is a width FOS* in the case of your bearing) then higher order datum features are not involved (in 4-19 B or C alone cannot constrain [w] so they must both be involved). That being said, and I'm going to preface this by saying that I am not 100% sure about this or that its a control that even makes sense, if we were to take 4-18 and apply a translation modifier to C (allowing translation in [y] ie: up/down) would the same still hold? My instinct says yes, but I have some doubt in my mind that is saying your spherical bearing example is a special case because of the arrangement of the datum features (datum feature simulators would surround the point/axis of rotation vs. offset as in 4-18).

As a final thought, now that I have dug into it - I am questioning some of the earlier logic. I can fully see how in 4-9 and 4-19 primary datum feature A is involved in constraint of [x,y], however I am now not so sure about [w]. I am trying to fit this into my thought process/logic outlined in the first part of this post (relationship between relative translation/rotation DOF constraint) - perhaps I am missing something and you can help me amend my understanding?

*Edit - I mistakenly stated planar feature, instead of a width FOS. If one side of the spherical bearing was (instead of the width) specified as the datum feature B, primary datum feature A would still be involved with constraint of [u,v]. I'm not so sure about other examples with planar features, or whether this statement would apply.

Apologies for the multiple edits.

 

[Sem_D220]

I think this is the key here. Ie: [x] cannot be fully constrained "everywhere" until [v,w] is constrained, same with [y] and [z]

Ok, I think I get it.

pylfrm, chez311,
It is not mandatory to start by constraining just v (or w) in order to finally constrain X "everywhere".

Consider a simple case of 3 perpendicular planar datum features. A(primary), B(secondary), C(tertiary).
XY plane is coincident with A.
B perpendicular to X.
Would you apply the same logic here too and say that A is involved in the constraint of X because it constrains v (which can potentially contribute to the constraint of X "everywhere")? I assume you would not, as you would say that since B alone is enough to constrain X "everywhere", the fact that A constrains v is irrelevant.
Well, for me - cylindrical datum features B and C from fig. 4-9, once both involved - do not differ fundamentally from a planar datum feature B. Together, they do the same job of constraining X "everywhere" independently of A. The fact that it takes a longer process to achieve this (involving cylindrical B and then C instead of just adjusting to planar B) doesn't make any essential difference.

 

[greenimi]

Sem, chez311, pylfrm and all,

Did you consider the can-may-must rule?
"4. Rule of Maximum Utilization (The Can-May-Must Rule) (SmartGD&T): If a Datum Feature can (is able to) and may (is permitted to) constrain a degree of freedom, it must."

https://www.eng-tips.com/viewthread.cfm?qid=443653

see this discussion and Bill's pdf presentation.

3DDave (Aerospace) 1 Sep 18 00:01

Good article.

They suggest the PDF version:

https://www.qualitydigest.com/pdfs/Review of t...

Is this "rule" is alignment with your proposal?

 

[greenimi]

PDF attached

 

[chez311]

It is not mandatory to constrain v (or w) in order to constrain X "everywhere".

I would disagree. Per my previous post I would say it IS mandatory.

Would you apply the same logic here too and say that A is involved in the constraint of X because it constrains v (which can potentially contribute to the constraint of X "everywhere")? I assume you would not, as you would say that since B alone is enough to constrain X "everywhere", the fact that A constrains v is irrelevant.

I would still apply the same logic. B alone CAN constrain [v,w,x] however since it is secondary, primary datum feature A already constrains [u,v,z] - secondary datum feature B only constrains [w,x] so A is involved with constraint of [x] "everywhere".

 

[chez311]

greenimi,

While I am through the course of this discussion trying to gain an understanding of pylfrm's interpretation of DOF constraint (which so far I find very logical), I believe it conforms to Bill's "rule" which is really just a sort of obvious restatement of sequential DOF constraint. The only issue I have with it personally is his views on how it relates to the lower segments of composite FCF's, which is a related topic but I believe outside the bounds of what we're discussing here.