mixing MMC and LMB 1

[ozzkoz]

I'm looking for a little help with determining the appropriate datum feature simulator sizes when the datum modifier does not match the material condition modifier which appears in the FCF of the datum. If you look at Fig 4-16 in ASME Y14.5-2009 I see that the datum feature material boundary for D is 7.3 (option b). This is the maximum feature size (7.1) plus the perp tol (.2). This makes sense to me as the perp tol applies at MMC per the callout for datum D.

What if instead the perp tolerance had applied at LMC and option b still referred to D at MMB? Would the MMB still be 7.3 OR would you need to consider the maximum envelope which the perp tolerance at LMC would allow. I'd think that would be 7.5 (LMC zone would be 6.7 and a 7.1 dia could be line to line with that to give a 3.75 radius * 2 = 7.5).

Can anyone chime in and confirm which line of thinking is correct? The standard appears to assume the datum modifier matches datum feature material modifier. Perhaps it's silly to ever call out the datums and features with mixed modifiers but this is what I have.

Thanks

 

[Belanger]

First off, I think the philosophy of MMC/LMB is such that if a datum reference has the MMB modifier (such as option "b" in the given position tolerance) then it is usually because the datum feature (the pin) is assembling into something in a clearance fashion. Then, when we look at the perpendicularity tolerance on that same pin in the right-hand view, we know that it is still assembling into something. It's function can't change just because we glance from one view of the part to another. So we'd probably want the MMC modifier there; thus, the scenario you ask about seems pretty unlikely.

But I guess we can still pursue the question in a theoretical sense. If the perpendicularity tolerance is changed to the LMC modifier, but MMB modifier remains on the position callout (option "b"), then it means for the position callout that we must still simulate the datum as if it were at its MMB, which would now be a worst case of 7.1 + 0.2 stated perp tol + 0.2 bonus tol = 7.5 mm (which is what you came up with).

That being said, this doesn't mean we can trot out a fixed-size gage, because that 7.5 is not a constant boundary, but a worst case for only a pin of size 7.1. (The MMB for a pin of diameter 7.0 would be 7.3 mm.)

Let's see if other agree with my rambling here... And, depending on how deep this discussion goes, we might also want to look at Figs. 2-16 and toss in some terminology about the boundary conditions (inner/outer locus). Ugh.

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

 

[ozzkoz]

why wouldn't 7.5 be a constant boundary? It would be the min boundary which would accept the part and respect datum precedence. I would think the gage would need to allow for any acceptable part which meets the high order datum. In this case I could have a pin of 7.1 and my gage would have to allow for that.

However like you said this might be solely an academic discussion since it might not make sense. I guess I'd like to know if the standard implies that it is flat out wrong to do this, or if it allows for it and it's a bit of an awkward inspection.

 

[Belanger]

It's not a constant boundary because a pin of 7.0 mm (with the LMC modifier, as you propose) could only tilt 0.3 (0.2 stated plus a bonus of 0.1). To say that a constant boundary of 7.5 always works is to imply that the pin could tilt out to 7.5 at all size values. IOW, your datum simulator could be simulating a part that isn't even to spec.

When an LMC modifier is on an OD like this example, the only constant value is the inner boundary, often called the virtual condition (also see Fig. 2-16). The outer boundary, which is what the MMB concept is seeking for a datum, is not a constant boundary -- notice that Fig. 2-16 has blanks in the last column except for the resultant condition, which is akin to the value of 7.5 in our discussion.

I don't think the standard has a dogmatic stance on the general question you propose; I'd lean toward your last statement of it being an unusual and awkward thing.

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

 

[axym]

Hmm ... this is interesting. I've never gone down this particular path. I agree that the "mixed modifier" case is unusual and awkward, but it would be good to have a way to deal with it.

John-Paul, I agree with you that the outer boundary (resultant condition) of the pin is not constant. But the maximum material boundary, the volume in which there could be material, is 7.5. My gut says that the simulator should therefore be a fixed size of 7.5. I understand that the simulator could accept a datum feature that isn't to spec, but I don't think that matters. The simulator's purpose is not to verify the tolerances on the datum feature - that would be done separately. The simulator just needs to fit over any datum feature that is in spec, and therefore needs to be sized at 7.5.

What do you guys think?

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[Belanger]

Yeah, Evan, that's where I was kind of winging it

My hesitation is that I don't know if the intent is that the MMB symbol be interpreted as the worst-case MMB or the actual MMB for each part made.

IOW, although the datum simulator isn't meant to check the datum itself, how can we know if the datum shift that appears is really allowable?


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

 

[axym]

This is where I try to look past what is "allowable" and try to find out exactly how Y14.5 defines the simulator and its behavior. From this, we can deduce what the datum shift effects would be.

I just read through sections 4.11.5 through 4.11.6.3. It all points to a worst-case MMB, as the collective effects of MMC and geometric tolerances. I didn't find anything that hints at the idea of an actual MMB. This would be very difficult to accomplish physically anyway - the simulator would have to somehow sense the UAME size of the datum feature and then offset out by a calculated amount.

Again, the fixed size 7.5 diameter simulator may not detect an out-of-spec datum feature placed in it. But this can also occur in other contexts. If a part has a datum feature whose size is past its LMC limit, it will still fit on the simulator and allow unrealistic shifts. The part would be flagged as nonconforming in the separate step of verifying that the datum feature conformed to its tolerances.

This leads to the contentious question of what should be reported for a geometric tolerance, when one (or more) of the datum features is out of spec. In other words, is it meaningful to report a value for a Position tolerance if one of the datum features is oversized? Or undersized?

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[pmarc]

I have been thinking what kind of functional application would force me to mix (M) and (L) modifiers within the same feature control frame and to be honest I am not able to find any. I am really interested if anybody sees any situation where it could work.

Some other thoughts:
- Although Y14.5 does not specifically show and describe such kind of mix, I would not say it is clearly forbidden.

- Agree with J-P that outer boundary is not a constant value, however I am rather leaning towards Evan's opinion that presence of (M) modifier causes the datum feature simulator D for option b) to be fixed at 7.5 size. Unfortunately standard is not clarifying the definition of MMB in cases like we have here, therefore my opinion is just based on my personal intuition.

- Y14.5-2009 introduced quite powerful and useful possibility of clearly stating what should be the size of datum feature simulator for a geometrical tolerance if its size is not clear. Even if we find an agreement here on size of datum feature simulator D for option b) we will not be sure that other readers of this particular drawing will get into the same conclusion. Therefore we can use a method defined in 4.11.6.3 and give a precise information about a size of simulator.

- Maybe a little bit off topic, but what really intrigues me is a statement at the beginning of 4.11.6.3: "in cases where[...] other boundary is desired". I am wondering if for instance [dia. 7.6] could be specified instead of (M) for datum D in positional FCF. So the value which is greater than any MMB of datum feature D shown in fig. 4-16. This concept IMO would make sense for some functional reasons, and I'd like to know if you see any problems with that.

 

[Belanger]

Excellent, excellent ... since we can't be too sure what everyone will think of when they see MMB (since it's not constant), then para. 4.11.6.3 supplies the answer. That's what I was reaching for, since I still wouldn't be comfortable with just using the worst-case MMB.

It's crazy how we were discussing that figure on page 61 all this time, yet I didn't bother to read the paragraph that you mentioned, pmarc!

And yes, I suppose that this alternate method allows a datum simulator size that even exceeds the worst-case MMB.

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

 

[axym]

(M) and (L) might be mixed on a feature that has more than one function. One example might be a cast feature that something needs to fit over at one stage of use (making it a datum feature referenced at MMB) and then is finish machined at a later stage (requiring a Position tolerance referenced at LMC). In some cases, a feature of size might be referenced at a material condition that does not directly relate to its mating function. This might be done in order to minimize assembly tolerance accumulation.

To me, the meaning described in 4.11.6 and 4.11.6.1 is clear and the simulator should be sized at 7.5. But I agree that others may not draw the same conclusion, and explicitly specifying the desired boundary size wouldn't be a bad idea.

I don't see a problem with specifying a boundary size that is larger than any MMB. There could be functional reasons for doing that.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[fcsuper]

L is not as common as M, but there are reasons to use either. I don't know that one dimension could comfortably have both applied in the same FCF, but I imagine related dimensions could.

Matt Lorono, CSWP
Product Definition Specialist, DS SolidWorks Corp
Personal sites:
Lorono's SolidWorks Resources & SolidWorks Legion