Orientation controls referencing 2 datums 2

[Sem_D220]

What are the opinions on the following schemes?

1. Angularity control referencing 2 datums in the FCF, when the basic angle is specified between the controlled face and the secondary datum. The primary and the secondary datums are perpendicular to each other.

2. Parallelism control referencing 2 datums in the FCF, when the controlled face is nominally parallel to the secondary datum and perpendicular to the primary datum.

3. Perpendicularity control referencing 2 datums in the FCF, when the controlled feature is nominally perpendicular to the secondary datum and at some other angle to the primary datum.

I haven't seen any of these brought as an example in the Y14.5 standard (unless I'm missing one), or in any other sources I was exposed to, but I also don't see how the contents of chapter 6 may reject those schemes. Schemes #1 and #2 are ones I wanted to implement for real cases, but hesitated (eventually I did ). As for #3, I haven't encountered a case requiring this, but I can imagine one. I think I once heard a GD&T professional say that there should always be a basic implied 90° angle (for perpendicularity), or a basic angle of some other value (for angularity) between the controlled feature and the primary datum feature whereas the secondary datum may only constrain DOF / orient the tolerance zone. But, if the DRF should first and foremost reflect the functional interface, there certainly may be cases where a vice-versa scheme is justified. The problem is - there are no figures to point to if such position needs to be supported. In Y14.5, looking for orientation controls that reference more than one datum, I find only figures 6-4, 6-8, and even 6-17, all show an implied right angle relationship or basic angle between the controlled feature and the primary datum reference, never to the secondary.

Has anyone else dealt with this dilemma? Maybe it's only my lack of knowledge / experience, and such schemes are either commonly practiced or clearly not supported? Whatever the case is your input will be very much appreciated.

Edit: I'd like to add that I realize that one solution could be to use profile of a surface for orientation, but from various reasons I prefer to utilize orientation controls and reserve this solution only as a last resort, if needed.

 

[CheckerHater]

Your part can be ugly as hell, but as soon as you manage to fixture it, you can start measuring.
Relating everything to same set of datums will make things easier for CMM, for example.
I can imagine using profile for everything will be seen more and more as CMM, scanners, vision systems are getting cheaper and more widespread.

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

 

[pmarc]

For example, when from functional reasons some feature must be controlled for orientation to a simulated mating face, and that same mating face happens to be not the one which constrains the largest quantity of degrees of freedom in assembly, and therefore not represented by the primary datum plane/datum feature simulator.

Keeping in mind the "functional datum selection paradigm" I further analyze this case:
If the required orientation control happens to be angularity, then according to your previous statements you would consider it legitimate to control a basic angle wrt non-primary datum (emphasizing that the "main" control applied by angularity in this case, whatever "main" might mean in this context, will still be to the primary datum. But, that will not work against the cause, as far as I'm concerned). On the other hand, if the considered control is perpendicularity - you would recommend to change the DRF from a functional one to a non-functional one, or specify a control other than perpendicularity, which fits the basic sorry - nominal relationship of the controlled feature to the primary datum better, and thus "disguise" the described above design intent, all in the name of clarity.

You keep asking about the same things, semiond.

I will just comment on the underlined part of the quote. I am not recommending to change the DRF from a functional one to a non-functional one. If, again using the modified fig. 6-8, your main concern is to control the hole for perpendicularity to datum plane A, then using orientation tolerance wrt |C|A| - regardless if you choose parallelism or perpendicularity symbol - is a non-functional callout. As chez311 seemed to say in one of his previous replies in this thread, the fact that you keep repeating over and over again that what you are attempting to do is driven by design intent does not automatically make it a functional callout.

 

[Sem_D220]

chez311,
I think that the main thing you are missing here is that sometimes measuring something relative to a secondary datum doesn't mean that the measurement or control is "indirect". On the contrary, you are measuring directly relative to what you want to be your reference - which might be a theoretical plane represented by a datum feature simulator that acts similary to a mating surface which is not the primary mounting base, and is expected to support the part on 2 high points only. Measuring any other way will only mean that the control is not verified with full correspondance to function. Making the reference a primary datum in such case will not mean you have a "tighter" control over the required orientation. On the contrary, it may mean that the orientation you measure may not represent the reality of an assembled part well enough.

Edit: pmarc, having just read your last reply, I direct what I wrote above to you too.

 

[Sem_D220]

CH,
I agree about profile.
I first thought that using it instead of an orientation cotrol might be a bit "overkill", (otherwise, what do we have orientation controls for anyway?) But considering the posts here, it appears that all orientation controls are more problematic than I thought.

 

[chez311]

semiond,

At this point I think you're arguing semantics, and actually thats exactly what I meant by indirect. It would be the difference between the indirect control of parallelism (2-points) in Figure 6-4 of the feature to datum B vs. the direct control of parallelism (3-points) in Figure 6-2 of the feature to datum A.

Taking Figure 6-4 what I have been saying all along is that if the indirect control of parallelism shown of the feature to datum B is sufficient then you're done. If its not good enough, then you will have to change the callout to parallelism to datum B. Your suggestion was instead to change the control to parallelism to [A|B] which is what I, and others here, have said is not legal.

 

[pmarc]

semiond,
This is my last try in this thread.

If you are okay that the orientation callout applied to the hole in modified fig. 6-8 is relative to |C|A|, then this means that from functional point of view the hole is controlled for parallelism to datum plane C in the first place, and then for perpendicularity to datum plane A. It does not mean that the primary functional concern is perpendicularity to A. If you want to say that you primary functional concern is perpendicularity to A, the orientation callout applied to the hole should be relative to A primary.

You said: "Making the reference a primary datum in such case will not mean you have a "tighter" control over the required orientation. On the contrary, it may mean that the orientation you measure may not represent the reality of an assembled part well enough."
By saying so you just admitted (whether you are aware of that or not) that the option with A primary is not your main functional concern, meaning that the first thing you are interested in is the option with parallelism to C primary.

 

[Sem_D220]

pmarc,

This is my last try too. For the modified fig. 6-8 we were discussing consider:

Option 1: perpendicularity |C|A|
Option 2: parallelism |C|A|
Option 3: perpendicularity |A|C|

I start by addressing the first two options:
I say that despite of the different "associated" orientations, option 1 and option 2 will result in exactly the same tolerance zone, under the same (coincident) DRF per produced part. From some reason, you claim this tolerance zone can only describe a parallelism control, corresponding with option 2 for the callout. Geometrically - this tolerance zone, whether you admit it or not, will also control perpendicularity to A, because the tolerance zone cylinder must have it's axis perpendicular to datum plane A. Datum plane A for these 2 options is established under certain conditions, so that it represents best the real mating surface of datum feature A at assembly. According to those specific conditions under which it is established, datum A it is considered a "secondary datum".
There can be 3 considerations on whether to choose option 1 or option 2:
a. Geometrical correctness
b. Design intent clarity
c. Legality per the standard

As already explained both geometrical terms "parallelism" per para. 6.3.2 (to C) and "perpendicularity" per para. 6.3.3 (to A) are relevant for the considered tolerance zone. So, "a. Geometrical correctness" can support both options.

Since perpendicularity to A was described as the main concern of the designer, and when I say "to A" I mean that same A which is established in a |C|A| DRF and is expected to stabilize the part on 2 high points, I would say that "b. Design intent clarity" consideration suppprts option 1 (but I don't dismiss your claim on the contrary, because people are used to see a primary datum feature nominally perpendicular to the controlled face at perpendicularity FCFs).

c.Legality per the standard: standardwise, there are no paragraphs that say that the basic orientation relationship associated with the control symbol (0° for parallelism, 90° for perpendicularity) is required to exist nominally between the controlled face and the primary datum. The only drawback is lack of example figures that show an opposite example, apart from fig. 4-7 for angularity, that can be interpreted as the 60° (or whatever angle) control equivalent of perpendicularity to non-primary datum. Since what is obligating are the paragraphs and not the examples shown in figures, there is nothing that can reject option 1 in terms of legality. Option 2 is obviously acceptable too.

So far considerations (a) and (c) supported both options, while option (b) supports option 1 (in my opinion).

Now we consider option 3. perpendicularity |A|C|. This option is fundamentaly different from options 1 and 2.
Parts that could pass a control by options 1 and 2 could be rejected by option 3 and vice versa. That is because the datum reference frame established by this control will not be coincident with the previous one discussed, and consequently, there will be a different tolerance zone established.
Since the DRF established by options 1 and 2 represents functional interface better, option 3 is rejected.

With all that said, as I clearly stated before several times, I do not recommend a scheme similar to option 1 to anyone, because of the same reasons that ignited this lengthly discussion. The sole purpose of this post is to try and explain again the reasoning behind the considered concept (represented by option 1), a reasoning which from some reason you seem to insist to deny, and dismiss that option as a non-functional control. The reasons of your doing so are still unclear to me. If what you say would be true, there was no point to open this thread, and it would do nothing but waste the time of the participants. With the reasoning explained clearly this time (hopefully), I hope you realize that this is certainly not the case here.

 

[Sem_D220]

Taking Figure 6-4 what I have been saying all along is that if the indirect control of parallelism shown of the feature to datum B is sufficient then you're done. If its not good enough, then you will have to change the callout to parallelism to datum B. Your suggestion was instead to change the control to parallelism to [A|B] which is what I, and others here, have said is not legal.

chez311, this is one of the questions I was trying to get an answer on all along: not legal according to which paragraph? Please don't point to the literal description accompanying fig 6-4 (which is only an example of a specific case) , about "datum B invoked to constrain additional rotational degree of freedom of the datum reference frame" it doesn't contradict the fact that as a secondary datum, it sets an orientation of the tolerance zone too.

Edit: one more thing - my arguments here are not over semantics, and the datum precedence order, which you suggest to change - is not a control-dependent semantical matter. "Direct" and "indirect" for me is determined by correspondance to function, not on the associated meaning of the words primary ("most influential?") And secondary ("less influential?") to describe a datum reference. I don't think an orientation control automatically "suggests" a specific DRF according the specified symbol.

 

[pmarc]

From some reason, you claim this tolerance zone can only describe a parallelism control, corresponding with option 2 for the callout.

I am not saying that and never said that. In fact, in one of my previous replies I said that you are allowed to use whatever symbol you want (because the standard does not explicitly forbids you to do this). I am saying that option 2 offers more clarity over option 1.

Since perpendicularity to A was described as the main concern of the designer, and when I say "to A" I mean that same A which is established in a |C|A| DRF and is expected to stabilize the part on 2 high points,

That is the key problem of this entire conversation between you and me. It was not clear to me that by "the main concern of the designer is perpendicularity to A" you meant datum plane A derived from 2 high points. To me the natural implication of "the main concern of the designer is perpendicularity to A" is that A is derived from 3 high points, and that is why I have been calling the orientation callout wrt |C|A| non-functional.

As for option 3, I fully understand the consequences of calling out A primary and the difference between this option and the other two options as well as the reasons for which you might not want to use it... Just wanted to say that C secondary adds no value to it.

Is everything clear now?

 

[Sem_D220]

Yes, it's clear.
I'm glad we managed to clarify all points.

 

[chez311]

pmarc/semiond,

I guess we found one of the sources of disconnect. That changes the conversation a bit.

At the risk of dragging this out - if semiond is stating that perpendicularity to A is of primary interest to the designer, however still acceptable to be controlled at 2 high points (as opposed to 3 which I assumed as well), assuming for a second semiond's scheme of perpendicularity to [C|A] is allowable, and taking the modified Figure 6-8 example (changed basic angle to 90deg and added third datum plane C that is perpendicular to A and B as well as parallel to the feature axis shown) I would like to pose the following question:

Could someone explain to me how perpendicularity to [C|A] in the modified Figure 6-8 would be any different than parallelism to [C|A]? In parallelism to [C|A], the perpendicularity of the feature to datum A would still be controlled indirectly from the same two high points, right?

I know you don't like the term "indirect" in this instance semiond, but its the word I'm using to describe the 2 vs. 3 point reference. And when I'm asking about a "difference" I am not asking about perceived/communicated intent, I'm talking about how the actual part would be fixtured and measured.

 

[Sem_D220]

chez311,
To your question, from the standpoint of DRF and tolerance zone establishment, as well as fixturing the part for inspection, the two are exactly the same.

Allow me to address one issue with your wording, without intention to be nitpicking:

semiond is stating that perpendicularity to A is of primary interest to the designer, however still acceptable to be controlled at 2 high points

What I'm truly stating is:

"perpendicularity to A is of primary interest to the designer, however still and it is only acceptable to be controlled at 2 high points"

I describe a case where control with datum feature simulator A contacting at 2 high points rather than 3 and constraining 1 rotation rather than 2 a more accurate control, not a sloppier control. This is because datum feature simulator A that will be made to behave this way will represent the actual mating surface in real application better than a primary datum feature simulator A. This is the most important point I wanted to convey.

 

[pmarc]

chez311,
There would be no geometrical difference between parallelism and perpendicularity to C|A in the modified fig. 6-8. The only difference is exactly in how the callouts could be perceived. Of course I cannot speak for every person in the world, but I am pretty sure that most of the people would find the option with perpendicularity to C|A very confusing.

 

[chez311]

pmarc,
Thank you for your answer, this is what I believe as well. I don't think we are all going to get full consensus on this point, but your confirmation is good enough for me.

semiond,
You are free to utilize whatever callouts you think describe your part function best. It seems pmarc and I are in agreement that your method, while maybe not expressly prohibited in the standard, adds nothing to the actual definition of the feature except confusion - and I don't believe this is an acceptable trade-off for whatever perceived communication of intent/functionality you believe you are achieving. If you specify a feature in this manner I would expect to hash out this same conversation with designers, inspectors, and many other people who encounter the part drawing with similar debates and disagreements.

Because we probably will not agree on the above I will say this - thank you for bringing this question up, because while we have not arrived at the same conclusions it forced me to think about orientation controls and the definition of associated DRF's in a way I had not previously and as a result I have gained a better knowledge of this topic.

 

[Sem_D220]

pmarc, chez311, I agree that the option with perpendicularity might confuse the drawing users.

Perhaps now that all misunderstandings have been eliminated, it is a better timing to address the dilemma:

Considering the modified fig. 6-8,
Which is the more problematic scenario?

The designer having to specify parallelism wrt primary datum C and secondary A, when what he really cares about and intends to say in the drawing is that perpendicularity to secondary datum A is needed.

Or, the confusing perpendicularity call out wrt |C|A|, which will establish the same tolerance zone as the previous option, translating the intent of designer from English to GD&T language more directly, probably not illegal, but doesn't make a familiar use of the orientation control?

Edit: chez311, I just saw your last post. I'm glad that this thread is beneficial for you too. Obviously It is also beneficial to me, as I gained a better knowledge of what are the common perceptions regarding orientation controls.

 

[Sem_D220]

By my last post, I don't intend to re-ignite the discussion, it is intended more as a prelude to conclusions...

 

[powerhound]

The designer having to specify parallelism wrt primary datum C and secondary A, when what he really cares about and intends to say in the drawing is that perpendicularity to secondary datum A is needed.

This thread is not only really long, there are some extremely long posts within it. As a result, I haven't read every single post so if this has been answered, forgive me for bringing it back up. I still can't imagine a scenario where the true requirement is that a feature be perpendicular to a secondary datum and not the primary.

pmarc, semi has already refused to provide an actual example of this. Since you seem to understand where he's coming from can you make a drawing of something that would actually require this type of callout? This way I might be able to offer more to the discussion. Currently, I'm lost as I just am not grasping the point.

John Acosta, GDTP Senior Level
Manufacturing Engineering Tech

 

[Sem_D220]

powerhound,
I don't remember refusing to bring a real part example on this particular thread. However, unfortunately my previous experience of bringing up specific cases from my field and informing about concrete uses of parts on this forum is not very positive. I learned that it distracts from the GD&T subject and leads to arguments over how parts function instead. I now prefer to only bring very general geometry representations and describe how the part is mounted in assembly in terms of what degrees of freedom are constrained by each considered feature, and the requirement from the controlled feature. If that is good enough, you can address my sketch from 21 Jul 07:47, and consider that the main mounting face against which the part is clamped firmly is datum feature A, the mating surface at datum feature B locates the part at the direction perpendicular to it and constrains another rotation. The controlled face is a clamping surface for another rectangulsr part that once mounted should be parallel to the face that supports datum feature B. Therefore parallelism to secondary datum feature B is applied, and the datum precedence order is chosen in a way that represents real assembly conditions. I hope that this clarifies the intent, and thank you for trying to help.

 

[greenimi]

I learned that it distracts from the GD&T subject and leads to arguments over how parts function instead.

Well, well, I do think that those two go hand in hand.

For me it is an academig good thread and to be used for education.

A stops 2 rotations and 1 translation.
B stops 1 translation and 1 rotation.
TZ should be parallel to B and perpendicular to A. The part should make contact on 3 high points at datum feature A and 2 high points at datum feature B. With that description do you consider the scheme shown in the sketch correct or not?
Side note that might be relevant - for many cases of this type I encountered, parallelism to B is actually more important than perpendicularity to A. However, I do not consider it wise to break it down to two separate orientation segments (parallelism and perpendicularity) with different values, each referencing a single datum. That is because such scheme would not reflect the real application interface of the part. It will create 2 separate tolerance zones within which the face should be contained, and none of those TZs will be required to be oriented as I described above. It is important to orient the part for inspection as it functions in the assembly.

Is the feature shown with parallelism located? If not, shondn' shouldn't be first located and oriented afterwards?
And if additional refinements to B only is needed, why you are reluctant to add the approprite callouts?

 

[powerhound]

I don't remember refusing to bring a real part example on this particular thread.

When 3DDave twice asked you for an example, you replied:

This is the second time in a row you ask this question. And the second time in a row I'm not going to answer on it.

The above is why I said that you are refusing to provide an example.

I learned that it distracts from the GD&T subject and leads to arguments over how parts function instead.

This is actually what's supposed to happen. If you are applying GD&T based on a mysterious notion, rather than how the part functions, you're doing it wrong.

Therefore parallelism to secondary datum feature B is applied

This is incorrect. The parallelism will be to datum B since it is secondary. Datum feature B can be as out of parallel as much as it wants.



John Acosta, GDTP Senior Level
Manufacturing Engineering Tech