GD&T Theory - Position/Translation Modifier

[Jacob Cheverie]

I have attached a stripped down print of a part - A cube with two holes in it. Top face is datum A, Center hole is datum B, Side face is datum C.

The question that I have comes down to the theory behind the dimensioning I guess. I have listed three datum reference frames for a Positional tolerance.

1. [A|B(M)|C]
2. [A|B(M)|C|>]
3. [A|C|B(M)]

In #1, datum B constrains translation in the plane of A. The MMB modifier allows datum displacement upon departure from MMC. My question is this - is "shift" allowed in the direction of the basic dimensions indicated while having datum C as tertiary? Would we need to add the translation modifier as is seen in #2? Or would the better approach be #3, where we use C to constrain translation in the direction of the basic and only allow the small hole to be controlled to B in one direction?

Functionally, C will constrain rotation about the axis of A if used as a tertiary datum, so how would that work in case #1 if the part shifts from datum C simulator? Would the datum displacement be ignored while making constant contact with C? That would override the can-may-must rule.

Any thoughts are welcome and appreciated.

 

[axym]

Jacob,

As you already know, this is a can of worms. ;^)

The question boils down to what happens when a higher precedence datum feature doesn't fully constrain one of the degrees of freedom that it is responsible for. Are the lower precedence datum features required to constrain DOF's that were left partially open by the higher precedence datum features, in addition to their assigned DOF's?

In the case of FCF #1, datum feature B is assigned translations x and y but the MMB reference would allow shift between the datum feature and its simulator. So the x and y translations would only be partially constrained. Datum feature C is assigned only the clocking rotation w. Is datum feature C obligated to restrict the translational shift in the x direction, because it can?

I would say that a lower precedence datum feature should not be allowed to constrain a DOF from a higher precedence datum feature. In other words, it needs to allow the shift that would be allowed by the lower precedence datum feature. So in the OP drawing, shift should still be allowed in the direction of the basic dimensions indicated while having datum C as tertiary. The translation modifier would not be required.

This topic is highly debatable, and I expect that others will have differing opinions. You will find that Y14.5 does not address this issue directly, and offers conflicting clues.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[Belanger]

I would say that a lower precedence datum feature should not be allowed to constrain a DOF from a higher precedence datum feature.

Yes, ASME isn't always clear about this. See attached -- ASME's Figure 7-20 (option c) brings up the same question.
That said, I don't think the translation modifier would be the way to go in the OP's example. That modifier was envisioned for a datum feature that itself is given a location tolerance back to the higher-order datums.

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

 

[axym]

John-Paul,

Yes, Fig. 7-20 (c) shows a classic case of a higher precedence datum feature only partly constraining its applicable degrees of freedom. Datum feature A is capable of constraining two rotations and two translations, but it's so loose that it's not even touching the simulator. Datum feature B is only supposed to constrain one translation, but it's shown in full contact with its simulator (constraining two rotations as well as the one translation). Is this what had to happen, or just a scenario that could happen?

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[SeasonLee]

For the scenario #1, Am I right to make datum simulator C movable on my gage design until it makes maximum contact with the datum feature to set the orientation of the DRF.

Season

 

[Jacob Cheverie]

Thanks for your input Evan,
It's a fascinating issue, and I can see why Y14.5 does not fully define what is supposed to happen. It is quite interesting to ponder.

Thanks for the alternative example from Y14.5 John-Paul.

Season,
That was my reasoning for assuming the necessity of a translation modifier - I think John-Paul has a good point in that Y14.5-2018 seems to imply that the translation modifier allows the datum feature to translate within the tolerance zone established by a control associated with datums of higher precedence.

 

[Dean Watts]

In Fig 7-20 (2018) datum feature B cannot be located wrt the primary datum A. I don't see an issue with illustrating the effect of an MMB modifier applied to datum feature A.

In the case that Jacob shows, a lower precedence datum feature C can be located relative to a datum reference frame established by higher precedence datum features [A, B]. Bringing up the possibility of applying the translation modifier for a planar datum feature sent me off to re-read the sections in the 2009 and 2018 standards that cover the translation modifier. This brought about an "OMG" moment, to find that the translation modifier was not restricted to datum features of size. To apply a translation modifier to a planar surface when the default of RMB already creates the equivalent effect, is a problem in my opinion.

My re-reading also led to finding that the word "size" still appears in the requirements for TGCs/datum feature simulators when MMB and LMB are specified. Another OMG moment, since MMB and LMB apply to planar surfaces, not just datum features of size. The wording should not be "adjustable in size, when the datum feature applies RMB", and it should not be "fixed in size, when MMB or LMB is specified". 2018 7.5.2(e) should say "fixed at the specified boundary, when a MMB or LMB modifier is specified." 2018 7.5.2(f) should say "adjustable with regard to the amount of material, when the datum feature applies at RMB." As we know, many things need to be tightened up in the wording in Y14.5.

To address the valid point that Jacob brings up, it seems to me that an added rule that says something like "when a datum feature applies at LMB or MMB then the part may translate or rotate (shift) relative to the datum reference frame due to the effect of the LMB or MMB modifier. The presence, and amount, of this shift depends upon the margin between the datum feature's surface and the LMB or MMB datum feature simulator. Any lower precedence datum feature that is capable of constraining a degree of freedom that may be subject to this shift must also be referenced at LMB or MMB."

Thoughts? Darts to throw?

Edit - Dang! I'll throw a dart at what I proposed above. The wording I proposed would also apply to Fig 7-20, since B in Fig 7-20 "is capable of constraining a degree of freedom that may be subject to this shift". I think the added rule needs to apply to the conflict that Jacob points out, when a lower precedence datum feature can interfere with the shift that a datum feature at LMB or MMB may provide, due to the fact that the lower precedence datum feature can be located wrt the DRF established by the higher precedence datum features.

So, my modified proposal is now "When a datum feature applies at LMB or MMB then the part may translate or rotate (shift) relative to the datum reference frame due to the effect of the LMB or MMB modifier. The presence, and amount, of this shift depends upon the margin between the datum feature's surface and the LMB or MMB datum feature simulator. Any lower precedence datum feature that is location constrainable relative to a datum reference frame established using the higher precedence datum features must also be referenced at LMB or MMB." I think the objective of the new rule should be to prevent the conflict that Jabob points out, and I think that conflict is a concern only if the lower precedence datum feature can be located wrt the higher precedence DRF.

Dean

http://www.validate-3d.com

 

[aniiben]

Original cases
1. [A|B(M)|C]
2. [A|B(M)|C|>]
3. [A|C|B(M)]

Added cases
4. A|B(M)|C(M)>|
5. A|C(M)|B(M)|

Dean,

Do you see your proposed solution working also in cases 4 and 5 above?
In other words, I would guess these scenarious would not be legal, wouldn't they?

 

[3DDave]

It would have been sufficient to explicitly use the (x,y,z,u,v,w) notation, but it certainly seems like the CMM software guys were just wanting to skip having their software do anything that might require simultaneously fitting two features with the established order of precedence and most participants were too disinterested in the tedious setup required for the explicit control.

The use of the moving datum simulator in castings made sense in the context it was used. Borrowing that and dumping it into the '14.5 standard for everything was not well considered.

It was always the option to establish a datum feature by identifying a width, even for a cylindrical feature, to show the precise direction that movement was allowed; unlike the in-FCF symbol, applying the symbol to the feature could show any direction that freedom is allowed.

 

[Dean Watts]

Hi Aniiben,

In case 4, both (M) and the translation modifier are applied to datum feature C. I think that is OK. I also think case 5 is OK. I am proposing that C should be referenced at MMB or LMB if B is referenced at MMB or LMB (and I do think MMB and LMB should be allowed to be mixed in their application for these cases). What I am proposing is only for those cases for which a lower precedence datum feature can be located relative to the DRF established from higher precedence datum features, as in the example that Jacob provided. The cases 4 and 5 you have added are both are OK, I believe.

Since datum feature C in the example provided is a planar surface, I don't think the translation modifier should be applied only if C is referenced at RMB. This is due to the fact that RMB for a planar surface already accomplishes the effect that the translation modifier would provide. I think some wording should be added to Y14.5 to prohibit the application of the translation modifier to a planar datum feature that is referenced at RMB. (M) fixes the amount of material a datum feature simulator has at MMB. For a feature of size, the translation modifier then allows moving it's center from location specified by a basic dimension. For a planar surface I think it would be unlikely that the translation modifier would model the function of datum feature C's mating feature, but the standard should allow it anyway. This is why I say that case 4 should be OK.

The standard should be written to accommodate what is possible and valid, without restrictions based upon how likely the need is for a particular combination of modifiers. Only cases like a planar datum feature with both RMB and the translation modifier applied should be prohibited. With RMB present as a default condition, the translation modifier would change nothing if the datum feature is a planar surface. It would be misleading and confusing to allow it to be applied in this case.

Dean

http://www.validate-3d.com

 

[Jacob Cheverie]

Dean,

Thanks for taking the time to dig into this topic. I completely agree with you that the wording for TGC/datum feature simulators at MMB/LMB should not depend on the term "size" if we are in fact allowing datum features not of size to be referenced at MMB/LMB. I agree with your logic that if a lower precedence datum can be constrained in location w.r.t. a DRF established by datums of higher precedence and will interfere with the potential datum displacement, it must be referenced in some special way. Which way to reference it is unclear to me and I would assume that it would really depend on the application.

In this particular example, if you just assume to reference C at MMB so as to allow for the datum displacement gained by B at MMB, you may shift farther away from the TGC of datum feature C which would open up the rotation about the axis of datum feature B even more. Also, going the other way, what if datum feature C at MMB still didn't allow for all available shift (i.e. the geometric tolerance on C back to [A|B(M)] is much more restrictive than the Position tolerance of the hole. Datum C would conform to it's tolerance by taking advantage of the shift allowed by datum feature B, but referencing datum feature C at MMB would physically interfere with the shift when measuring the Position of the hole.)

 

[Dean Watts]

Hi Jacob,

Yes, I agree that datum feature C referenced at (M) may, or even likely will, still interfere with the shift that may be provided by datum feature B referenced at (M). I think that may be completely fine, especially if this interference models the functional situation accurately.

I think if C is not referenced at MMB or LMB, leaving it at RMB, then it completely conflicts with the shift that B at MMB may provide. I suppose that in some odd cases even |A|B(M)|C|, for your example part
above, with C at RMB leading to the part getting moved to a "one-sided" shift relationship with the MMB datum feature B simulator, could model function. As soon as we say it should be disallowed some creative designer may show us a functional case for which this would provide data that will predict whether the part will function or not.

Now that I'm thinking of this more, maybe no new rule is needed. Maybe it is sufficient to have awareness of the odd situations the material boundary modifiers can create..?

Dean

http://www.validate-3d.com