Rotation 1

[aniiben]

ASME Y14.5-2009
Part assembled on A (flat surface) ( flatness considered--not shown in this sketch for simplicity) then
centered on B (OD) (perpendicularity to A considered--not shown in this sketch for simplicity) and
bolted down with 4 screws (Ø.475 holes) (basic dimensions considered--not shown in this sketch for simplicity)

The other 4 holes (not dimensioned for size in my sketch) are non-functional (aid-process only)

Not sure about rotation.

Can I use C(modified at M) to stop the rotation only? Nothing is stopping the rotation except 4 bolts in the same time.
Is A, B(M) and C(M) valid? Legal? Why? Why not?

 

[chez311]

drawoh - thats a fair point, I guess that lends credence to 3DDave's original suggestion to rely on simultaneous requirements for [A|B(M)] instead of a sloppy datum C.

On the topic of potentially using a pattern of holes for tertiary datum C (either set of 4x holes) is there any opinions (besides semiond - i think its established we agree) whether it is allowed? Already agreed it would be preferable to utilize a tighter tolerance set of holes, I'm just asking about the fundamental concept as I see nothing inherently wrong with it.

 

[aniiben]

When I apply a datum to a Feature Of Size (FOS), I prefer that this be very much more accurate than the tolerances referenced from it. There are specific practical applications where a sloppy FOS can be used at MMB

I do not know if the datum schemes (in general) should take in consideration the tolerances (smaller or biggger than other GDT ones or smaller or bigger than the size tolerances)
The datum scheme should be reflected by how the part goes together with its mating components.

 

[pmarc]

aniiben, semiond, chez311,

I fully agree that the datum/datum features scheme should reflect function of the part, so let me ask you this: if all mentioned features, that is datum feature B and 4 datum features C can locate the part in the assembly, wouldn't it be better from functional standpoint to specify |A|B(M)-C(M)| in the applicable position and profile feature control frames?

 

[drawoh]

chez311,

There is allowed, and there is a good idea. Presented with a four-hole pattern tertiary datum, I can figure out a fixture. Is this a good description of the design requirement? Neither of us knows!

--
JHG

 

[aniiben]

Since datum features B and C are referenced MMB in two position and one profile callouts, what makes you think that when this part is mounted in the gage containing datum feature simulators A, B and C, the datum feature B will be the one that will always stop translation and always leave the rotation-constrainer job to datum feature C

The intent is to have only B stopping the translation DOF and that's why we have Ø.020 tolerance of position at MMC for the 4 holes whcih make datum feature C.

I fully agree that the datum/datum features scheme should reflect function of the part, so let me ask you this: if all mentioned features, that is datum feature B and 4 datum features C can locate the part in the assembly, wouldn't it be better from functional standpoint to specify |A|B(M)-C(M)| in the applicable position and profile feature control frames?

We don't want C to locate. We want C to orient only, hence tertiary, and hence the questions raised.....

 

[semiond]

aniiben,
Something that you said on this thread earlier (18:20) makes me suspect that you don't ascribe enough importance to numeric tolerance values.
If you rely only on datum precedence order in feature control frames to make your part work correctly - you WILL be disappointed. Do make sure that the tolerances you are going to specify for the OD (datum feature B), will allow ONLY the function you intend.

 

[pmarc]

aniiben,
Specifying datum feature tertiary does not automatically mean the feature will constrain orientation only, especially when secondary datum feature of sie is referenced at MMB.

For this particular geometry I see two ways to avoid C functioning as locator at certain conditions:
1. Define B at RMB. But would that reflect how the part works?
or
2. With B referenced at MMB, define datum feature simulators C at fixed size smaller than the virtual condition of datum features C by at least the difference between MMB size of datum feature B and LMC size of datum feature B.

 

[chez311]

pmarc - I'm trying to think, my brain is a bit fried this monday but it seems to me both [A|B(M)|C(M)] and [A|B(M)-C(M)] while potentially communicating different things to the person reading the print, would they not accomplish exactly the same thing since they're both at MMB? What am I missing?

both pmarc and drawoh - it seems to me the more I look at this the more I think relying on simultaneous requirements and not the bolt hole pattern itself for orientation/constraint of the final degree of freedom is probably the better idea.

 

[pmarc]

chez311,
You are right, both [A|B(M)|C(M)] and [A|B(M)-C(M)] accomplish the same thing. And this is how we have finally gotten to something in the Y14.5-2009 standard that is illogical, at least to me. The two should not accomplish the same, especially that if we remove the MMB modifiers, the two options will definitely not mean the same thing.

The problem is in the rule stating that by default the secondary and tertiary datum feature simulators must be basically located from each other. That was not the case in Y14.5M-1994.

 

[semiond]

chez311,
The simultaneous requirement alone might not reflect how the part should function in assembly. If the part is limited in rotation by the 4 screws on the mating part (bad idea for screws to take shear, but that's what the OP described), then the holes pattern should addressed in the DRF.
Otherwise, for example, he could have interference between the extrnal surfaces (controlled through profile) and other parts of the assembly. Simultaneous requirement won't prevent such a problem.

 

[greenimi]

The problem is in the rule stating that by default the secondary and tertiary datum feature simulators must be basically located from each other. That was not the case in Y14.5M-1994

Pmarc,
Are you sugesting that A,B(M),C(M) datum structure would be okay per. 94,but not per 2009?


Semiond,
Interesting point about tolerance values. I learn something new today.

2. With B referenced at MMB, define datum feature simulators C at fixed size smaller than the virtual condition of datum features C by at least the difference between MMB size of datum feature B and LMC size of datum feature B.

I never heard about this scheme....is it something shown with BSC ?

 

[semiond]

With B referenced at MMB, define datum feature simulators C at fixed size smaller than the virtual condition of datum features C by at least the difference between MMB size of datum feature B and LMC size of datum feature B

pmarc, remember that the screw shanks in aniiben's assembly are going to have size and location tolerances, and the datum feature C simulators should take them into account. I wouldn't recommend making those simulators smaller in the fixture. Instead, i would adjust the size and location of datum feature B (not shown in his sketch) to have much smaller maximum clearance from it's mating diameter, than the minimum clearance between the screw holes and the screws. Of course that this concept should also be translated to the tolerances of the part and the sizes of the gage elements.

 

[pmarc]

semiond,
The thing is that no matter what tolerance values will be set up for datum features B and C, there will always be a chance that the part, when mounted on a gage containing simulators B and C designed to their virtual conditions, will be located by simulators C and not simulator B.

That is why, to avoid this, one of my suggestions was to use simulators C that would be smaller in size than the virtual condition of the features C. And actually I think this is how this system/assembly should be designed if the intent for C is just to constrain rotation of the part - there should always be a loose between holes C and a virtual condition boundary that takes into account cummulative effects of location of the threaded holes in the mating part (assuming it is fixed fastener assembly) and the size of the screw shanks.

 

[semiond]

pmarc,
I now think that we approached the problem the same way, only i thought that you were suggesting to make the pin gages smaller than the virtual condition of the screws in the mating part. Apparently i was wrong.

However a question rises - how do you define in a drawing that the gage size for the fixture pins should be smaller than the MMB condition for the holes? If you call out datum C MMB doesn't it guide the inspector to prepare the pin gages at the virtual condition of the holes? This way location by the holes in inspection is still possible, even if we prevented it in the functional assembly by calculating all the tolerance conditions. I don't like the idea of an explicit note specifying this, since the general agenda in GD&T is that the callouts and datums should communicate the design intent clearly enough without additional notes and explanations. The designer is not supposed to write explicit inspection instructions.

 

[pmarc]

How about - instead of referencing C at MMB - putting [∅X.X] after C in the applicable feature control frames, where [∅X.X] defines the desired size of the simulators C?

 

[semiond]

pmarc,
I suppose you mean something similar to what is shown in para. 4.11.6.3 (page 61) in the 2009 standard, only without the M in the FCF.
I like the idea!
I suppose the phrasing that supports it is "In case that the boundary is not clear, or another boundary is desired, ..."
Problem solved. Thanks.

 

[chez311]

You are right, both [A|B(M)|C(M)] and [A|B(M)-C(M)] accomplish the same thing.

The simultaneous requirement alone might not reflect how the part should function in assembly. If the part is limited in rotation by the 4 screws on the mating part (bad idea for screws to take shear, but that's what the OP described), then the holes pattern should addressed in the DRF.
Otherwise, for example, he could have interference between the extrnal surfaces (controlled through profile) and other parts of the assembly. Simultaneous requirement won't prevent such a problem.

semiond - are you sure about that? This whole concept is throwing me through a loop because I think we are assuming that MMB does something for us that it doesn't - the MMB callout does not tighten or loosen the tolerance zones vs. simultaneous requirements. This provides no additional or reduced tolerance to the other features on the part, whether in position, form, or orientation - as far as I can tell in this case there is no difference in acceptance of the other features on the part whether its called out with [C(M)] or relied on simultaneous requirements. Its counter-intuitive but that's the way I'm envisioning it. This of course changes if one or both of the datum features were called out at RMB.

Pmarc's comment above solidified this in my mind. Imagine three different gauges that represented [A|B(M)|C(M)] , [A|B(M)-C(M)] , and [A|B(M)] with simultaneous requirements. Would they actually be any different?

 

[chez311]

pmarc - thats an interesting note about specifying a certain boundary condition. I haven't quite come to a specific conclusion in my mind about what would be the best way to call out this datum structure but your suggestion does seem to mimic the way the assembly would work - whereas the actualy MMC/MMB boundary of the features may locate a part at some extremes of its tolerance zones on a gauge, even at these extremes the part would almost certainly be designed with some clearance to the bolts which fasten it during assembly.

Have you seen this utilized on other bolted assemblies like this or are there any drawbacks to specifying a similar datum structure? I'm interested to know because its not like we're handling anything here that hasn't been encountered a million times in countless bolted assemblies with a centering/locating feature other than the bolt holes themselves - I'm wondering if I'm not overthinking this just a tad.

 

[semiond]

Pmarc's comment above solidified this in my mind. Imagine three different gauges that represented [A|B(M)|C(M)] , [A|B(M)-C(M)] , and [A|B(M)] with simultaneous requirements. Would they actually be any different?

The difference is in which datum feature does what task in constraining the part at assembly.

Here is what the OP stated:

The intent is to have only B stopping the translation DOF ...
We don't want C to locate. We want C to orient only...

The only scheme that describes the OP's design intent, of the 3 you mentioned, is the first one.
The second one can not guarantee the function for all possible tolerance combinations.

The third one completely puts the hole pattern out of the game, unless the pattern itself is toleranced with the same datum call outs and therefore subjected to the simultaneous requirement.
I'm not sure how this will affect the function and i would prefer to simply stick with the features that should locate and orient the part per the design intent as datum features with the relevant modifiers, specify the DRF with the relevant datum precedence order, and do the tolerance analysis to prevent unwanted situations. You read pmarc's suggestions - that's not such a complicated mission. The suggestion from 05:50 about the boundary (gage size) for datum C simplifies things quite a bit.
The advantage is that you fully integrate the mating part's geometry into the design, and eliminate all uncertainties about the function.

 

[drawoh]

aniiben,

How accurate is the diameter of your datum[ ]B feature? If an FOS datum feature is an order of magnitude more accurate than the tolerances referenced from, you do not need to worry about MMB. I think pmarc's concern is that a datum[ ]C fixture will engage your part as it translates in X and Y before the datum[ ]B fixture does. If the feature is accurate or it is called up RMB, it locates reliably in X and Y, and your datum[ ]C feature, whatever it is, constrains in rotation only.

--
JHG