Datum Shift Explanation and Advantages

[kakalee1]

Hi engineers,

I have a question regarding about specifying datum feature at RMB. What are the advantages of having a datum shift since we are not getting any more or less bonus tolerance? Wouldn't it complicate the inspection process since the part can now move (shift) around?

There is an example from the book that I can't wrap my head around it, and I hope someone here can help me out. On the attachment, for option b and c, I don't understand what the y mean by "to ensure that datum precedence is not violated..." and how they choose the 0.4 and 0.2 for each option.

Any help is appreciated.

 

[greenimi]

3DDave,

Speaking about the shape of the datum feature simulator IN RELATION with the tertiary datum issue here is what has been changed on Y14.43 (the gage standard)
Y14.43-2003 (Dimensioning and Tolerancing Principles for Gages and Fixtures) –used in connection or used to support Y14.5-1994 standard
- Fig. 1, page 10 a diamond pin construction is shown as a tertiary datum feature simulator. The same diamond pin is shown on page 68 Fig. B10. (same 2003 standard)

Y14.43-2011 (Dimensioning and Tolerancing Principles for Gages and Fixtures) –used in connection or used to support Y14.5-2009 standard
- Does not have the diamond pin construction/option shown. Also Fig B-10 (again in 2011) does show a fixed cylindrical pin instead on the diamond pin shown on B10 (2003).

Why was changed (datum feature E simulator from a diamond pin to a fixed cylindrical pin since the part design intent/functionality of the workpiece shown hasn’t changed?

 

[3DDave]

greenimi,

Since the committees don't publish a document on what problems they were trying to solve or who voted for the specific descriptions, there's no way for me to even guess what they were thinking or suggest who to contact. In addition, the ASME generally has a no-opinion policy regarding their standards.

In the thread:

http://www.eng-tips.com/viewthread.cfm?qid=215333

a statement is attributed to Meadows concerning "a long drawn out fight" that led to the (apparently not unanimous) adoption of the translation modifier. If true, then that sounds like a good way to build errors into what is essentially a mathematics book.

The usual fights in mathematics are over discovery credit and notation, not the truth that mathematics represents. Usually, before being widely published as a solution in mathematics, both the problem and potential solution are subjected to a lot of peer review and validations by mathematicians; not by technologists. Not that mathematicians are unwilling to stab one-another in the back, but they do it to control credit and prestige, not to silence critics of their results.

The use of the diamond pin shape was not to represent an idealization of the shape of the hole as a tertiary datum to function as an angular control. It was to alter a round pin to eliminate as much as possible the affect on the angular control that the distance between the holes created. That is, for small variations in distance, the contact using a diamond pin is nearly perpendicular to the surface of the hole. If a round pin is used the contact could be tangential.

In the OP example, if D were diamond shaped the datum simulator for it would be - that's right - a hole 7.3mm in diameter; the same dimension as the width of the slot I previously proposed. Since D is not diamond shaped, then the simulator has to be an obround slot.

 

[axym]

3DDave,

FWIW, I think you're on the right track with the obround simulator. I had to look at this same example in a fair amount of detail a couple of years ago, and came to a very similar conclusion. When I traced out the places where the datum feature could be, and applied the rotational degree of freedom that is open, I got a volume that is obround when viewed from the top. I went even further, and saw that it was like a thick-waisted hourglass shape when viewed from the "radial" direction. The simulator is cylindrical only because there is another section in Y14.5 that states that datum feature simulators shall be the inverse shape of the datum feature, unless otherwise specified.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[pmarc]

3DDave, Evan,

Look at the attached picture and the description below:

http://files.engineering.com/getfil...ef3-bfa3-8ab736117e85&file=fig_4-16_MMB_D.png

The picture shows 4 extreme cases:
- In each case a man is standing exactly at the datum axis B and is looking at all possible locations (as defined in fig. 4-16) of datum feature D actual axis.
- In each case the yellow circle is dia 0.4 positional tolerance zone defined by |pos|dia 0.4(M)|A|B(M)|C(M)| callout. No matter what, the axis of the datum feature D at MMC must fall into that zone.
- Each reddish circle simulates dia 0.2 cylinder within which the axis of datum feature D at MMC will be assuming that this feature has been produced with maximum possible perpendicularity error to A.
- In each case two green solid lines - horizontal and vertical - are 2 out of 3 datum planes of |A|B(M)|C(M)| datum reference frame. These planes are always fixed.
- In each case two reddish dashed lines perpendicular to each other are 2 out of 3 datum planes of |A|B(M)| datum reference frame. One of these planes always passes through the center of the reddish circle, so that when the man is looking along that plane he can only see 0.2 actual perpendendicularity error.

Now, my question to you is: Does the allowable movement of the reddish circle relative to |A|B(M)| really constitute an abround shape or perhaps a circle?

 

[3DDave]

Nice diagram. Thanks.

Obround, as a circle is not required to encompass all the potential pin positions in the |A|B(M)| frame of reference.

 

[pmarc]

It is not obround, Dave.

First you check datum feature D against |perp|dia 0.2(M)|A| & |pos|dia 0.4(M)|A|B(M)|C(M)| callouts. As I tried to show it, these two requirements allow axis of D to be within dia 0.4 cylinder oriented and located relative to |A|B(M)|C(M)| datum reference frame (i.e perpendicularity callout is irrelevant when one tries to find maximum possible volume that encompasses axis of D in |A|B(M)|C(M)| DRF).

Then you get rid of C, and search for MMB of D relative to |A|B(M)| in order to properly establish |A|B(M)|D(M)|. What makes |A|B(M)| & |A|B(M)|C(M)| different? It is only the fact that one rotational degree of freedom is left unconstrained in |A|B(M)|. But the volume within which the axis of D can be does not change - it is still the same dia 0.4 cylinder located at basic 29 from datum axis B. It just that the dia 0.4 cylinder can freely orbit the datum axis B.

Evan,
What do you think?

 

[3DDave]

If I make a part that aligns to A, has a pin that fits B, and a slot that is radial from B and is only 7.3mm wide to fit over D, will that part fit all the possible places D is allowed to go?

In the attached image it is plain the scheme I describe is a workable solution that takes into account the perpendicularity refinement, which the original explanation does not. E is not part of A|B(m)|D(m); it should not influence the outcome.

 

[pmarc]

What exactly do you mean by: "all the possible places D is allowed to go"? Do you mean the yellow circle?

 

[3DDave]

What would the diameter be of the datum simulator for A|D(m)?
Is that diameter based on the position tolerance that includes E?

pmarc, in your diagram, the datum planes do not align between B and D, instead they maintain an alignment with a feature that is not part of A|B(m)|D(m). The observer is not following the datum sequence as is specified in the FCF.

 

[pmarc]

What would the diameter be of the datum simulator for A|D(m)?
Is that diameter based on the position tolerance that includes E?

Are we still talking about fig. 4-16 case c? Where is E on that figure?

My diagram does not show B and D aligned, becasue this is not the point of this diagram. The diagram is to show where in space (in what volume) relative to A|B(M) datum reference frame datum feature D can be. This is needed to find size and shape of MMB of datum feature D relative to higher order precedence datums, that is A|B(M). After that MMB is found, we can construct datum reference frame A|B(M)|D(M)| that is aligned with datum feature simulators B and D.

 

[axym]

Hi All,

Here is what I got for the MMB of datum feature D relative to A|B(M) when I looked at this a couple of years ago. I believe this is like 3DDave's idea, taken a bit further. It's supposed to be the shape that would always fit over all the places that D is allowed to go. The observer can "center" their reference frame in the tangential direction, because the last rotational DOF is not constrained. So we only see positional error in the radial direction. In the tangential direction, we only see the perpendicularity error.

Evan Janeshewski

Axymetrix Quality Engineering Inc.
www.axymetrix.ca

 

[3DDave]

Sorry - lost track of datum names.

Anyway, B & D need to be aligned because D is the orientation control for case 'c'.

The diameter you keep looking for is only a diameter in [A|B(m)|C(m)] When C(m) is removed, then the DRF is allowed to turn. There is no angular limitation. One can freely turn the part so that the pin that D is based on can be anywhere in a 360 position on that part.

If a DRF is set to follow [A|B(m)|D(m)] then the refinement applied to D relative to A needs to be considered. In order to get the answer supplied in the standard, the refinement has to be ignored. It's a guess, but there aren't rules in the standard that specifically say to ignore refinements, meaning the example is wrong.

Again,

What would the diameter be of the datum simulator for A|D(m)?
Is that diameter based on the position tolerance that includes C?

 

[pmarc]

Guys,
You are missing the fact that in A|B(M) datum reference frame the feature D can still be within a cylindrical volume.

Again, look at the part from the top. In A|B(M)|C(M) axes of D can be within a dia 0.4 cylinder that is perpendicular to datum A, located basic 29 from datum B and rotationally aligned to datum C.

In A|B(M) that volume does not change. What is changing is the fact that the datum planes of A|B(M) DRF are now free to rotate about B, but that has no influence on shape and size of that volume. If all axes are allowed to be inside dia 0.4 cylinder and nowwhere else, this cylinder will not magically change its shape and size in A|B(M). I guess, you can think about it that way, because A|B(M) is a subset of A|B(M)C(M).

... Sorry, but I am afraid, I am not able to explain it better.

 

[3DDave]

Frame of reference controls the shape and size of the volume.

In it's own frame of reference the feature only changes size; relative to A, the apparent diameter increases due to perpendicularity allowance; relative to another fixed feature, it increases due to position; but when it is used to locate itself it is only constrained by the intersection of the sets of constraints applied to it.

This is exactly as pictured in the file attached by Evan Janeshewski.

When viewed from B, the maximum apparent width of the solid piece of material that D is formed from can only be seen to be derived from the diameter of D and the tilt of D allowed by the perpendicularity tolerance.

 

[axym]

Hi All,

This is a very interesting discussion. Perhaps one of the more subtle and deep issues I've come across with GD&T.

pmarc,

It is true that D can exist anywhere within a cylindrical volume relative to A|B(M)|C(M). But the effect that Dave and I are seeing is that D cannot exist everywhere in that volume at the same time. Without the rotational constraint to C, the observer at B cannot distinguish the different tangential directions that D can exist at - he/she can keep turning to look straight at D. So the tangential variation of D gets becomes goes away, and the MMB becomes the quasi-obround subset of the cylindrical volume.

I admit that this gets very metaphysical, but there are practical consequences. A simulator made with the cylindrical MMB allows more datum feature shift than it really should - even when the D feature has extreme tangential tilt within its tolerances, there will still be significant datum feature shift. The only time that the D feature will fully constrain clocking is when it is at or near the radial limits.

If the simulator were made with the quasi-obround shape, then there would be no datum feature shift when the D feature has extreme tangential tilt.

3DDave,

Does these descriptions seem correct?

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[3DDave]

Evan,
The description does seem correct.

I did think of an alternative for this case: [A|D(m)|B(m)] It makes more sense if an orientation and location sensitive feature is being made, such as broaching a square hole into the pin. Does adding B(m) cause the simulator for D(m) to change size? I'm thinking the reverse, that the B(m) simulator is now a diamond pin, or a sliding pin aligned with D.

The version in the standard, besides having an incorrect explanation, makes no sense as a real application. Why not use A|B(m)|C(m) for the hole being drilled in D and give it a large position tolerance? It looks tagged on with only a glance at the explanation.

 

[gabimo]

I guess Jim Meadows agrees with pmarc. See the latest newsletter
December 2014

 

[gabimo]

Subject: Calculating the Correct Maximum Material Boundary (MMB)

Hello Jim,

I got a question, if you don't mind. It is about size/shape of maximum material boundary D in case (c) in fig. 4-16 from Y14.5-2009. I am in a middle of a discussion with a person saying that in this case the shape of MMB is not a cylinder of dia. 7.5, but an obround geometry of 7.3x7.5.

His argument is something like this:
"The suggested interpretation does not match the set of allowable places 'D' could get to. If one took all the possible parts that met the requirements applied to D and transformed them to no longer include C, then it is an obround. Picture the view as if standing at the center of B; would D ever appear to be more than its MMC and perpendicularity tolerance wide? Without a reference to C, there is nothing that fixes the direction of the view and so nothing that causes adding a location tolerance to the width."

As you are a member of Y14 committee and most likely are well aware about the logic hidden behind the case (c), could you help me with understanding this?

Thank you very much.

Pawel



Pawel,

The rules in the 2009 Y14.5 standard are very clear that every datum feature of size is simulated at its basic location (whether it is secondary or tertiary) and at its virtual condition. The question is really a gaging question. How would you represent B and D in this situation? The answer is that both would be represented at their basic location from each other and at their applicable virtual condition and by a gage hole or pin that is the same shape as the datum feature.

One could argue that it was not defined that way in the 1994 version of Y4.5, but it is certainly true in the 2009 revision. We had to change some of our gages in the Y14.43 standard (2011 revision) because of this rule change. It was also the reason that a datum feature translation symbol was added to the 2009 standard to allow the equivalent of datum feature D’s simulator to translate along a line toward and away from datum B. But without that translation symbol the gage pin and hole are stationary, basically located from each other and cylindrical in shape.

With the current rules, the 4-16 illustration is correct.

James Meadows
Chairman ASME Y14.43-2011 Dimensioning and Tolerancing Principles for Gages and Fixtures

 

[3DDave]

This is an essential change in interpretation from all previous versions.

As described existing tolerance analysis software will produce incorrect results. This is not a case of a previous indistinct interpretation getting clarified, this is a subversion of the concept of virtual condition.

Thanks gauge committee. 50 years of precedent down the drain.

 

[3DDave]

From Meadow's explanation:
"It was also the reason that a datum feature translation symbol was added to the 2009 standard to allow the equivalent of datum feature D’s simulator to translate along a line toward and away from datum B."

What rule requires that the diameter of the datum simulator will now only be 7.3 and it's travel limited to +/- .2 from the basic position if the translation modifier was used?

In light of this interpretation, what shape is the gage for Fig. 7-29 Bidirectional Positional Tolerancing, Polar Coordinate Method? That is, what is the shape of the maximum material boundary for the hole? Would this gage be a different shape if the hole was then used as a datum? If not, why not?