Positioning hole with edge distance? 1

[Rich68]

I am no GD&T expert, but I can't believe this is correct (See attached). It seems to me that this allows for the hole to float? I've always been under the impression that the feature had to have an absolute position.
If this is not valid, any feedback helping to explain how it is invalid would be greatly appreciated.

Thanks,
Rich

 

[Belanger]

...this allows for the hole to float?

The 4 holes are grouped together under a common callout, which actually gives two position requirements for the holes: the upper tolerance of .030 controls their location on the part (I think that's the answer to your question). We know this because that tolerance references all 3 datums.
The lower tolerance of .005 controls the hole-to-hole location, which is often more important to the function.

That said, I suppose some of the folks on here may wish to discuss datums B and C, but I'm just trying to keep it simple and answer your basic question for now.

 

[pmarc]

Rich68,
You are correct that the position specification, as a whole, is incorect.

In radial direction the tolerance zones are not fully defined with basic dimensions relative to datum axis B. The basic hole-to-edge dimension should be replaced with for example basic "pitch" circle radius.

Datum feature selection, mentioned by J-P, is different part of the story.

 

[Sem_D220]

I think that what the OP is concerened about is the 4X .XXX basic dimention from datum feature B surface to the surface of the edge of the hole.

 

[Rich68]

If the basic edge dimension went to the center of the holes instead of the edge and datum B was the surface of the radius (Not center axis), would it be ok?

 

[chez311]

Agreed with pmarc. As it stands now, regardless of the fact that the hole-to-edge distance is ambiguous (does it intersect with the center of the arc? its not clear) - its an incorrect specification of the hole true position. As it stands the hole axes is technically defined by the directly toleranced diameter, which means the axis true position changes with the size of the hole which is not allowed. As pmarc noted this should instead be fully defined with basic dimensions. This could be done in a variety of ways, including a basic pitch diameter/angular dimensions or linear x/y basic dimensions.

pmarc/JP - I assume in reference to datum selection you guys are referring to the lack of 180 degree opposed points on what seems to be referenced as a FOS datum B?

 

[chez311]

Rich68,

I'm hesitant to say that would be okay, with the caveat that datum feature B would have to be defined with a profile tolerance and its radius would then have to be basic as well. Right now datum feature B seems to be defined as FOS which it is not, since it lacks 180deg opposed points.

 

[Sem_D220]

I would add to what chez311 said above, that if the hole location dimension is kept to the surface of the radius, datum feature B symbol, in that case, should not be aligned with the radius dimension line, but either attached beneath the profile FCF that will define the arc geometry or connected to the surface itself or to an extension curve from the surface, clearly separated from the radius dimension line.

Datum C is also defined in a questionable way. From the way it is aligned with the basic angle dimension line, it suggests that the datum is meant to be a center plane derived from two non-parallel planes contacting two non-parallel surfaces. Is it OK to relate the datum feature symbol to an angle dimension similarly to how it is done with width dimensions defining parallel-surfaces features of size? When we had a GD&T training course at work, the instructor suggested it to me as a valid scheme when I asked him how to define a center plane datum to be derived from a dovetail-type feature. But, I never found anything that supports it in the Y14.5 standard. It would be interesting to get additional opinions about this.

 

[drawoh]

pmarc,

The tolerances are weird, and I cannot see what would make me do it like that. Are they wrong?

I assume the datum[ ]A is the bottom face, and that we are not seeing the whole drawing. The physical inside radius is the secondary datum feature, regardless of size. The two inside holes are centred about an axis defined by angle[ ]XX, datum feature[ ]C. The outside holes are located by their edges from the inside radius, on a radius defined by the inside holes. Inspection of this thing would be a pain. I am glad I am not designing an inspection fixture for it. The imaginary radius the holes are shown located on is not very meaningful beyond showing design intent.

If the drafter is trying to locate holes to clear screws, this is the wrong way. I am being pedantic here.

--
JHG

 

[pmarc]

I think that the true problem in OP's case is proper understanding of how basic dimensions work and what the difference between datum feature B and datum B is.

In case of position tolerance, basic dimensions are used to locate position tolerance zones from datum(s)/datum reference frame, not from datum features. So if we imagine for a moment that datum feature B in OP's example is regular 360 degrees cylinder controlled with +/- toleranced diameter, I guess very few of us (if anyone) would locate the holes from the edge of the datum feature B. Instead of that, there would most likely be a basic diameter dimension applied to the "pitch" circle of the pattern of the holes and together with the implied 0 linear dimension between the center of the "pitch" circle and the datum axis B this would fully define true position of the 4 tolerance zones relative to datum B. I will risk and say that this would be the most natural way to define it.

OP's case is different mainly because datum feature B is not a feature of size, however what is important is the fact that a datum feature simulator B used to establish a datum B is a contoured surface from which a center/axis can be derived. And from that center/axis the basic "pitch" radius dimension should be defined by default.

There is, of course, another problem with datum feature B in OP's sketch, as mentioned by chez311. Since it is not a feature of size AND since it has been referenced RMB in the upper segment of the position feature control frame AND since it doesn't have a location relationship to a higher order datum A (I too assume that datum feature A is the bottom or top face), there is really no way to find a point at which the datum feature simulator B would stop its expansion. This is what makes selection of this directly toleranced arc as a datum feature at least questionable, in my opinion.

 

[Sem_D220]

datum feature simulator B used to establish a datum B is a contoured surface from which a center/axis can be derived

there is really no way to find a point at which the datum feature simulator B would stop its expansion.

If the datum feature simulator can expand without limit, how can a distinct axis be derived from it? As the simulator grows in radius, the axis will dislocate further and further away from datum feature B, won't it? Please explain how these two statements go hand in hand.

 

[pmarc]

Regardless of how big in diameter the simulator is, its axis technically doesn't change and is equally determinable.

The fact that as the simulator grows its axis will dislocate further and further away from the datum feature B exactly proves that this feature referenced RMB is a bad choice (because it also means that the position tolerance zones dislocate together with the axis).

 

[chez311]

pmarc,

I obviously clearly agree that datum feature B is not a FOS, and I mostly understand the practical implications and issues you describe with trying to simulate such a feature at RMB being that its not a FOS - however I'm not sure I follow why the simulator could expand "without limit". Could you explain a little why this is? If as you say "its axis technically does not change and is equally determinable" - even though "size" has limited meaning if it is not a FOS wouldn't this axis still be technically fixed to the simulator itself would only be able to expand (ie: the linear distance between said axis and the surface) within the limits specified?

 

[Sem_D220]

Then, the basic pitch radius dimension scheme can only be valid under the condition that an MMB modifier is used, allowing to use a fixed size datum feature simulator. Correct? If this was already mentioned and I missed it, I apologize.

 

[pmarc]

The simulator could expand without limit because the drawing doesn't define size of MMB and LMB for datum feature B relative to higher order datum A.

If this was defined, for example by changing directly toleranced +/- radius to basic dimension and applying profile tolerance wrt A to the surface of datum feature B, the simulator would be able to expand from LMB size towards MMB size.

This, however, still wouldn't solve the issue of unlimited expansion of the simulator of the datum feature B referenced RMB in the position callout. - but this time the problem would be reduced to the LMB-MMB band.

That is why in this case the only feasible solution (assuming the entire inner arc surface needs to be used as datum feature) is to define datum feature simulator B at a fixed size. This can be MMB but also LMB or even BSC.

 

[Sem_D220]

This, however, still wouldn't solve the issue of unlimited expansion of the simulator of the datum feature B referenced RMB in the position callout. - but this time the problem would be reduced to the LMB-MMB band.

I'm sorry but I lost you here once again. Will the expansion of the simulatir allowed to be unlimited, or will the size variation of the simulator be only between LMB and MMB?

 

[pmarc]

Unlimited within LMB-MMB band making RMB pointless.

 

[Sem_D220]

In fig. 4-29 (edit: illustration a), Y14.5 2009 datum B is called out RMB. Is the scheme shown there pointless?

 

[pmarc]

Since it is not a feature of size AND since it has been referenced RMB in the upper segment of the position feature control frame AND since it doesn't have a location relationship to a higher order datum A (I too assume that datum feature A is the bottom or top face), there is really no way to find a point at which the datum feature simulator B would stop its expansion.

The underlined portion doesn't apply to fig. 4-29a, therefore what is shown in fig. 4-29a is not pointless.

 

[greenimi]

If this was defined, for example by changing directly toleranced +/- radius to basic dimension and applying profile tolerance wrt A to the surface of datum feature B, the simulator [highlight #CE5C00]would be able to expand from LMB size towards MMB size.[/highlight]

Wow, quite of learning from this discussion.

Just a quick question: Does the datum feature simulator geometry originates from LMB? Not saying it is not, but looks like it is the opposite from what expected.