Datum as plane tangent to two cylindrical surfaces 2

[tw_0407]

I have a part with two partial cylindrical surfaces and I want to make a datum that is the plane tangent to the two surfaces, as the mating part has a flat surface that is seated up against them.

Part is similar to this example I mocked up:

And I want the datum to be tangent to both surfaces, like so:

Can I just sketch a line between them on a top-down view like in the above picture, and attach a datum to that?

 

[3DDave]

Check my participation across many fora. This is the only one where one respondent to me has been a problem. This is actually the only one where one other person, since booted, was a similar problem.

I suspect it's because there has been only one consistently taking fragments of what I wrote out of context and claiming I meant something I did not.

You can also see that "second or third statement or opinion I express being under attack and accusations" is another misstatement.

In most cases when I see Burunduk has replied first I either don't comment at all or don't comment on what he's said.

How many hundred examples would you all like to see?

https://www.eng-tips.com/viewthread.cfm?qid=518678

is one.

 

[Burunduk]

Another not surprising attempt to discredit me, by 3DDave.

On this thread, it started at 20 Aug 24 17:06, with accusations and hostility, without any trigger other than an opinion that differs from his own, expressed politely and not even addressed to him initially. As usual, even when I responded to the attack moderately and substantively, he kept at developing this into a quarrel. And that pattern has been repeating on and on.

 

[Brian Malone]

3DDave and Burunduk, I think many people on Eng-Tips greatly respect the knowledge both of you have and share. Your bickering does get to be tiresome so I hope you can decide to call a truce without needing to justify anymore why you have thrown barbs at each other. There are probably equal levels of blame without trying to take a tally. Maybe moving forward, you can just state your guidance without adding any commentary on each other? Most of the time I find both of you are top-notch and quite amazing - there have even been times when you have collaborated with simultaneous solutions.

 

[Burunduk]

Brian, I accept your suggestion and will try to implement it.

 

[Burunduk]

If it's ok to get back to discussing tolerancing (hopefully politely, from now on):

"a can of worms"? Have you anything to support that

It would be a can of worms because if it becomes legitimate to attach the 'datum
feature symbol' to a representation of something that is not a 'datum feature' (such as a theoretical datum plane tangent to non-planar datum features), it would open the door for all kinds of ambiguous uses of that symbol.

Some of such uses are explicitly forbidden ("The datum feature symbol shall not be applied to center lines, center planes, or axes.") but still a common drawing mistake in the industry. If the rules of applying the datum feature symbol become less strict, such practices may become even more common.

Other than the explicitly forbidden placements of the symbol, other bad practices could get some legitimacy. Once I've seen a drawing showing a line representing a plane completely detached from the part, and not as part of a coordinate system, with the datum feature symbol applied to it and no robustly specified relationship to any integral features of the part.

If we agree that a datum feature symbol is only intended to assign an integral datum feature, we minimize those issues.

 

[3DDave]

A datum feature symbol needs to indicate real surfaces on the part. My proposal and the diagrams in the standard already show that they can be applied to this case.

What bad practice can happen as long as the requirement is that a real, solid, flat datum feature simulator would contact the part as shown on the drawing?

Not only do the examples show that, several show that one contact with two bumps are sufficient to locate one such plane, that plane is used to set the orientation of a second plane that is contacting one bump.

Since this is the way real parts really function and can really be inspected, the "bad practice" suggestion comes across as unfounded. It certainly would not allow a floating unrelated datum symbol.

If we agree that datums are tied to surfaces on parts, then we minimize those issues. I don't see real surfaces on real parts as "ambiguous." If you can sketch a case where no one could understand how a plane of defined location and orientation could be not understood to have a defined location and orientation and is currently accepted otherwise, it's worth considering.

 

[Burunduk]

3D,
While you bring up some valid points, I can't say I fully agree.
The examples you mention, such as figure 4-33 - the last one you described, detail applications of the datum feature symbol to irregular feature of size. The datum feature symbol is approriately associated with a dimension line of the 3 mm width basic dimension:

I have solid reasons to believe that in this and similar cases the intent is to indicate derivation of a center plane datum from the actual mating envelope of the irregular FOS; the AME would be two parallel planes at maximum separation, one contacting the 2 bumps and the other the one bump on the other side. So, the use is unique for that.

When the intended datum plane to be derived is a plane tangent to the surface, I would say that different conventions apply. For a non-size feature application, there is no dimension line to associate the datum feature symbol to, and we need to consider the 2 main methods to designate a datum feature on the drawing:

1. When a complete surface of the part is to be used for mating with a datum feature simulator, it's labeled by a datum feature symbol, attached to the surface or an extension line from it. This could also apply to 2 coplanar surfaces like in figure 7-52.

2. When the considered datum feature should only contact the datum feature simulator on local elements, datum targets of the appropriate contact type (points, lines, areas) should be used. If the contacts are defined by tangency to a plane, as in the OP's case, this can be indicated on the drawing by a phantom line passing through the datum targets, and possibly a matching note, if considered necessary.

 

[3DDave]

Datum targets are such that a gauge maker can build a rigid fixture with pre-located features, even if some are moveable along rigidly defined directions, to represent the datum targets. You don't elaborate on what datum target type you expect to use, but what targets and what dimensions would be passed to the gauge maker to work with?

For a non-size feature there is already an example of a datum symbol not attached to a dimension that ties two surface tangencies together, but since you restrict all answers to only the exact examples in the book, you limit yourself.

That same limit to reasoning must also apply to the use of datum targets for this exact case since there is no example of targets used on surfaces sharing a plane tangent to both.

 

[Burunduk]

3D,
In the past, you made a comparison between computer programming language and the language of dimensioning and tolerancing. This is a comparison I see value in.

For a programming language to work, one of the most important things is consistency, meaning that every expression has only one meaning.

But if we apply your approach to labeling datum features on a drawing, we will end up with a mix of meanings: using the datum feature symbol would need to be interpreted in some cases as full contact with the datum simulator, and sometimes as partial, localized contact limited to a point, line, or flat element. And partial, localized contact limited to a point, line, or flat element would sometimes be expressed by the datum feature symbol and sometimes by the datum target symbol.

To me, this doesn’t sound like something that could work in a programming language. Are you ok with that anyway?

 

[3DDave]

You haven't made a case, even though you attempted to borrow the metaphor. There is never full contact with the datum simulator, not ever. Datum targets say that only specifically located portions - I have asked for your way to specifically locate the portions and you have not done so. What dimensions and what targets? The target system tells the exact configuration of the inspection equipment. What targets and what dimensions would be passed to the gauge maker to work with?

 

[Burunduk]

There is never full contact with the datum simulator, that is correct, but there is a difference between specific elements designated for contact, and an infinite number of potential contact points on the entire surface of the feature. I assume you got the gist, so it seems going into semantics with that was non-essential.

"What dimensions and what targets?"
As I mentioned: "If the contacts are defined by tangency to a plane, as in the OP's case, this can be indicated on the drawing by a phantom line passing through the datum targets, and possibly a matching note, if considered necessary".

Those would be datum target lines of course, as I've repeatedly mentioned throughout the discussion (and I apologize if I haven't communicated it clearly enough).

The phantom line indicating a tangent plane in combination with datum targets may not be a commonly encountered solution, but at least that solution would be consistent with the choice of the symbol to designate the datum feature, depending on the type of intended connection with the simulator.

 

[3DDave]

Datum target lines have only single points of contact on surfaces they are tangent to.

The desire is to use the full depth of the feature which is why I remain doubtful of your intention to not have the same contact.

There is no example of a target line being tangent to two curves at the same time, invalidating them for the same reason you wish to consider cases that are shown of a plane making this same contact invalid.

If I want a plane to contact two pins, that is shown here:

There is no doubt there are two planes; only one tangent to two cylinders is required so only one extension line is used. This can be seen by even a casual reference to the standard, where a close reading is required to find what you didn't know existed until this thread and does not seem to have been noted by anyone else on the committee. No one using Y14.5 should also be required to know what is in Y14.41 to interpret a drawing.

Also, to note, the special case in the standard figure 7-58 is solely to avoid using a construction such as references to using 2 datums of equal priority to establish a third, unnamed plane with an orientation and location implied by the diagram and not by any formalization elsewhere. This special case does not apply to the OP problem and neither does the example from Y14.41.

 

[Burunduk]

3D, I'm surprised that this isn't clear by now. I'm referring to exactly the same type of contact as you – linear contact in the depth direction (along the cylinders) in the contact areas with the plane.

The difference is only in the way it's defined on the drawing. I suggest, for the sake of language consistency (like in programming languages), to label a datum feature intended to create linear contact with the datum target line symbol, as is done in the more trivial examples.

The information that the measurement fixture designer will receive is that the contact lines are determined by the tangent to the plane. The way to convey this information is through a phantom line, in the view where the plane is a line, and possibly also a note. My goal is to maintain language consistency for all cases where linear contact with the simulator is required.

You suggest using a symbol intended by default for either designating a whole feature for potential contact at datum simulation, or designating a feature of size (regular or irregular).

 

[3DDave]

The line in the examples is not parallel to the axis.

Finding an exact dimension to locate the high points in the vertical directly that will have to be given to the gauge maker in advance of getting the part is the problem, one unaddressed in your suggestion.

A note will not be consistent, varying with each and every creator.

You suggest using a line indication in a direction that it is not intended to apply.

My method is used with a datum symbol applied to a line that is not a dimension line, but you refuse to respond as to how not accepting that use also makes yours unacceptable.

 

[Burunduk]

"Finding an exact dimension to locate the high points in the vertical directly that will have to be given to the gauge maker in advance of getting the part is the problem, one unaddressed in your suggestion."

No need to find an exact dimension to locate the high points. The intent is that a planar simulator is used to contact both cylindrical surfaces. This is the information to convey to the gage maker - and it shouldn't be a difficult task to convey that information.

 

[3DDave]

You were clear that datum target lines would be involved. I see datum target point indicators which aren't in the examples in the standard for datum target lines and that use appears to open an avenue to confusion. In addition, datum targets typically do not get datum labels; they are standalone. No [A], , or [C] appear in Figure 7-58 and no additional labels or notes are used or required to show they do.

A datum symbol is the standard way to indicate this sort of datum, as done in the figure that this example would negate, on a solid line that is aligned with a surface.

 

[Burunduk]

I see datum target point indicators which aren't in the examples in the standard for datum target lines and that use appears to open an avenue to confusion.

I have to disagree.
When a datum target line is shown on an edge view, it is indicated by the the datum target point symbol. A direct view of the surface that shows it as a line is optional, not mandatory. If there is no direct view that shows the same datum target indicated by the datum target point symbol, it is understood to be a datum target line.

In addition, datum targets typically do not get datum labels; they are standalone. No [A], , or [C] appear in Figure 7-58 and no additional labels or notes are used or required to show they do.

In ASME Y14.5-2018, please refer to 7.24.9, at the end of this paragraph, it is stated: "Where datums are established by targets on complex or irregular surfaces, the datum may be identified by a note such as DATUM AXIS A or DATUM PLANE A."
Considering this, I think my way of implementing the clarification as shown is definitely acceptable.

A datum symbol is the standard way to indicate this sort of datum, as done in the figure that this example would negate, on a solid line that is aligned with a surface.

I haven't seen such example in the standard other than the unique case of an irregular FOS as already discussed (and it's different, because of the association with the dimension line), or any indication in the text that such use is acceptable. And I see no justification for it from the perspective of tolerancing language consistency.

 

[3DDave]

When a datum target line is also shown in an edge view it gets that symbol. The location of the datum target line is given as dimensioned to other datum features or other datum targets in that case.

No FOS. No associated dimension line. Just a solid line to indicate planar, in this case, contact the same way solid lines are usually used to indicate a plane is the intended datum.

 

[Burunduk]

That solid line is just an extension line common to two planar, and coplanar, surface outlines. If those surfaces were curved instead, I don't think such representation and use of the datum feature symbol would be well-supported.

Also, application of datum targets does not mean that unnecessary dimensions are necessary. If the two pins are fully dimensioned, no additional dimensioning of the datum targets would be required. The relevant basic relationship (parallelism) is implied.

 

[3DDave]

The use of simultaneous tangency to two or more curved surfaces already is supported and no target lines are required so no dimensions for them are required.

The reasoning for the targets in 7-58 is they are separate parts of a single datum which do not form a plane as opposed to being multiple contacts forming a single datum plane, the condition in this picture:

There are multiple examples establishing a plane of contact on multiple tangencies. There are none establishing datum targets the same way. It makes no sense to interpose an additional step into the process that is reserved for a different purpose.