Virtual condition 1

[3DDave]

Suppose there is a basic brick shape - Datum feature A uses the largest face, datum feature B uses the next largest face, and datum feature C uses one of the smallest faces.

In the same face identified as datum feature A is a hole through the part of diameter E, positioned at MMB with [A|B|C] as the DRF with a tolerance of dia. X. Assume the basic dimension to B is b1, and the basic dimension to C is c1.

Now the face opposite of C is identified as datum feature D and is toleranced with a profile tolerance to [A|B|C] with a zone width of Y. Assume the basic dimension from C to D is d1.

The desire is to make a bracket sharply bent of a rectangular piece that mates to datum feature A, will have one edge coincident to datum feature B plane, and hook around the end of the part to mate with datum feature D. A bolt will pass through both parts.

What is the virtual size of that hole in the [A|B|D] DRF?

 

[Burunduk]

Would you mind to clarify how that "caution" remark is relevant, considering that there is no trace of form control attributes in the values used for the calculation of VC per both Y14.5 and Y14.5.1?

"The more informed approach " is only your opinion of what you wish the standard said to support yet another argument.

I suggest you to read the definition of the surface interpretation of position and orientation tolerances at MMC.

 

[greenimi]

Pmarc,

You know I am watching your posts like a hawk (here and on other sites I know you are posting)

I won’t ask you which camp are you in (and I will not try to read between the lines), but I will ask:
What do you think, had Y14.5 decide to interpret position at MMC with only the surface interpretation (canceling resolved geometry interpretation) will this approach solve the conundrum?

I am thinking this based on your post below:

“... The problem #2 is that fig. 5-3 seems to say something different. It clearly shows that the local diameters of the straightness tolerance zone don't depend on the sizes of the circles in each cross-section perpendicular to the UAME axis of the feature, but on the "feature sizes" (whatever these are) obtained in the cross-sections normal to the DML. This isn't mathematically rigorous solution, in my opinion. And this is also one of the reasons why in the new version of the math standard, Y14.5.1, that is going to be released soon, this topic will not be covered. Long story short, I believe this whole problem, although interesting from a mathematical point of view, would have low significance if Y14.5 decided that the only intepretation that is really worth to consider in case of DML straightness tolerance at MMC/LMC is the surface/virtual condition boundary interpretation. “


If that is the case, then probably Independency applied to the feature of size won’t influence the calculation of the VC and one camp won’t have anymore their current (valid) arguments.

Please, when you get a chance could you comment on my statements above

Thank you very much

 

[3DDave]

geeenimi

the problem is this: Per Rule 1 the item is perfectly straight or flat at MMC and with Independency that restriction is eliminated. Only if an item is required to be perfectly straight or flat at MMC is there a non-zero minimum from which to base a virtual condition. As soon as that restriction is removed the feature could be a torus or hollow cylinder or curlicue or spirally twisted and meet the eliminated straightness requirement while also meeting the feature of size at MMC requirement. Since the intersection of all possible permutation of these allowable forms at MMC eliminate entirely any possible volume, there is no MMB and therefore no virtual condition.

What Burunduk continues with is that since the standards don't deal with degenerate geometry that, by default, analysis of realizable geometry must apply to degenerate cases; since they are degenerate, they cannot be analyzed at all.

(typo)

 

[Burunduk]

As long as the feature is controlled for not violating the virtual condition, the extreme cases of curlicue or spirally twisted holes will not be accepted, even if the virtual condition is imposed by a position or orientation control.

Try preventing position control functional gage from rejecting parts based on form error. It is as impossible as making it filter orientation error from location error.

The other approach pushes for not allowing a tolerancing scheme perfectly legitimate and meaningful per the standards that may, for some cases, lead to only good parts being accepted with minimum cost, only because those good parts are not allowed to be produced as twisted as the lack of rule #1 form requirement alone suggests they are allowed to be.

 

[Burunduk]

As for the latest question of axis interpretation versus surface interpretation:
My opinion is that it doesn't essentially matter for the considered issue.
Even with only the surface interpretation, it is still not impossible to look at the virtual condition as a combination boundary generated by a perfect form MMC hole initially at true position and perfect orientation either displaced towards true position radially half the tolerance value in any possible direction, tilted in any direction to reduce the same radial distance to zero, or having some combination of tilting and displacement until contact with the VC.

However, the following is important to bear in mind:

The above certainly doesn't mean that the hole must actually be limited to perfect form at MMC. Thinking otherwise indicates a misunderstanding of the function of the VC as a limiting boundary rather than a resulting envelope. It is indeed possible to think of the MMC diameter which is part of the calculation of VC as the size of the perfect form MMC boundary instead of just the allowed minimum ALS, but it is still not more than a parameter of the VC calculation formula. The VC boundary fits the above-mentioned behavior only if perfect form at MMC is required. Where it doesn't fit that description it's still a functional boundary limiting variations - as it is intended to be. The VC boundary doesn't need perfect form at MMC to work (no prerequisite). Eventually, the calculated boundary imposes an independent control.

 

[3DDave]

Using Independency without a controlling form tolerance is a bad practice. Is there a good reason to argue for using a bad practice that is specifically cautioned against in the standard? Follow the standard, add a form tolerance and there will be a virtual condition. Don't follow the standard, ignore the caution, and there isn't one.

There is a requirement that an explicit straightness or flatness tolerance must be smaller or equal to the related position or orientation tolerance, but nothing covers the case when there isn't any form tolerance at all, particularly when there is no size related modifier - is it MMC, LMC, or RFS? The picked examples for calculating virtual condition are where it is shown to apply to Rule #1 compliant features; they simply don't bother with examples of what to never do.

Anyone is welcome to not only use a known bad practice, but also make up their own rules for cases the standard does not address for their personal use, but it's still wrong to claim they have a basis in the standard.

As to questions of "legal," "compliant," or "legitimate" - these words are not used in the standard and cost isn't a consideration in how drawings are to be interpreted. Saving money is best done by use of explicit and standards-supported methods.

In retrospect the Committee screwed this up. Rather than saying the straightness/flatness must be less than position and orientation, they should have specified that the position and orientation tolerance should be greater. This would make obvious that those tolerances must encompass the given feature form tolerance and it would use the exact same evaluation to MMC as applied when not associated with a position/orientation tolerance.

This would not affect my analysis, but it would certainly stop the incorrect, unsupported arguments for any alternative. Perhaps others would hold out hope the standard would add a contradiction, but I suspect the correct answer is to apply the same requirement to Independency for similar reasoning as to other form controls and say that it cannot be applied on its own when position or orientation are also used on the same feature because its effect is always is larger than the position/orientation tolerances.

As to that last response, it is a convoluted rewording of exactly what I said, even though you have not worked out the conclusion correctly. "Requisite" is another word that does not appear in the '2009 standard, "pre" or otherwise; nothing in the standard is a prerequisite for any other part.

"The above certainly doesn't mean that the hole must actually be limited to perfect form at MMC" That is a strawman argument that no one supported, making the assertion "the following is important to bear in mind:" unnecessary. As an argument technique it is used to suggest, by mixing in one correct statement, that the remaining statements must also be correct.

In this particular thread of argument, the basis of which is what to do when the feature is not required to be of perfect form at MMC, it is peculiar to mention it and the mention seems intended to distract from the flawed interpretation of what to do when the form is specifically defined to have no limit at all.

We already are 100% certain that position or orientation would not control form on their own because the standard specifically requires that the user explicitly specify a tolerance value that is smaller; if it worked otherwise, the form tolerance could be left blank or use any value in those circumstances. Since it doesn't work that way Independency is incompatible with position - there is no virtual condition generated.

 

[Burunduk]

Instead of addressing "argument techniques" in details and pointing out words that don't appear in the standard, you could address the fact that controlling a tolerance of position or orientation at MMC by a functional gage or a simulated boundary is bound to reject an extreme form error that can do any damage to the part function as far as assembly is concerned. This also makes the caution remark irrelevant since "the form completely uncontrolled" is not the case.

A possible application is when the form is not a design concern but assembly is. Since the virtual condition as a variation limiting boundary is the only thing required to ensure assembly function by simulating the worst case of a mating part, there may be cases where requiring perfect form at MMC or any explicit form tolerance does not add any value. That's not to say that the justified cases are a majority or even common. But claiming that skipping a form control in a case like that may never be a cost-effective practice is the same as saying that excess drawing requirements do not influence manufacturing costs.

Everyone is free to make their own rules to ban tolerancing schemes according to what they think is a bad practice, these rules can be documented in internal standards or simply applied as guidelines a designer adopts for his work based on his own judgment. Saying that a "rule" of this type is enforced by compliance to Y14.5 when everything indicates that it is not so is wrong. The added opinion that the committee "screwed this up" is not a good justification. One can criticize the standard and point out the issues without providing misleading answers to validity type questions.

"add a form tolerance and there will be a virtual condition"
This is not true as long as having the size limits stated and a tolerance that applies at MMC/LMC is all that's required for VC establishment and the boundary size calculation.

The condition when there is a stated form tolerance that has requirements applied to it by being used in conjunction with position/orientation at MMC is not comparable with the condition of independency. The reasoning behind the "less than" requirement is preventing portions of a direct form tolerance to be abolished by a boundary that indirectly limits form variations because this combination of requirements "damages" one of the requirements as the stated form tolerance can never be fully utilized. Where there is no direct form control (or implied rule #1 form control) no such issue exists. Independency does not mean that the part must be accepted with any possible form variation.

 

[3DDave]

Funny.

The functional gage for a Independency/Position tolerance hole is no pin at all. Because the virtual condition does not exist neither does the simulator for it.

 

[pmarc]

What do you think, had Y14.5 decide to interpret position at MMC with only the surface interpretation (canceling resolved geometry interpretation) will this approach solve the conundrum? [...]

If that is the case, then probably Independency applied to the feature of size won’t influence the calculation of the VC and one camp won’t have anymore their current (valid) arguments.

I am not sure that cancelling the resolved geometry interpretation for orientation and position tolerances at MMC (and LMC) would help to solve the problem. The second camp might still say that the formulas given to calculate the VC size artificially override the real worst-case boundaries that could be generated by some special (not necessarily very sophisticated) as-produced geometries.

I think what might help, however, would be a development of a geometric tolerance specification directly indicating that a boundary of a pre-defined size is to be protected (without going into considerations whether this boundary is really the smallest/largest ever possible or not).  

 

[Burunduk]

Funny.

The functional gage for a Independency/Position tolerance hole is no pin at all. Because the virtual condition does not exist neither does the simulator for it.

Actually, not funny.

Because you still can't see that a virtual condition is applied by an MMC/LMC control as a variation limiting boundary depending only on the specified size limits for the feature and the stated tolerance in the FCF of that same MMC/LMC control. And you don't let go of the misconception that a VC is a resultant envelope that combines all the different possible variations derived from the other controls or from lack of them.

 

[3DDave]

Don't ever think of what "virtual" means or that your favorite formula results in negative volumes.

 

[Burunduk]

There will be no negative volume if the tolerancing makes sense. That is up to the designer. That's the whole point - conserving a specific isolated volume (for an internal feature) that can't be violated as long as the feature conforms to a tolerance specified with the MMC modifier.

 

[3DDave]

Any time the position tolerance of a hole pattern is greater than the hole diameter the result of that calculation is negative. That's why there is an axial interpretation to position tolerances. The upper segment of a composite tolerance often would result in a negative value; it's only the lower segments that typically have positive values.

 

[Burunduk]

Any time the position tolerance of a hole pattern is greater than the hole diameter the result of that calculation is negative.

This means that the location of the group of holes as a pattern is non-critical. In these cases, it is only the feature to feature relationship which guarantees the conservation of "untouched" volumes in the datum reference frames of the lower segments. It may be allowed when whatever is assembled in these holes may translate/rotate relative the interface features of the datum reference features in the upper segment. A negative value has the same practical meaning as zero.

 

[3DDave]

If it was not critical it would not have a tolerance.

So there is no volume reserved and no virtual condition in spite of your insistence that this was always positive for a position tolerance and a hole.

And independency has exactly the same effect - it makes the virtual condition equal zero.

Do what you want. It's not like you'll ever put this example on a drawing and explain to your QC/CMM guy how it's just the same as if you did not. That's the beauty of your arguments here - no need for you to ever apply it to your products.

 

[pylfrm]

3DDave,

Imagine a hole is specified as follows:

⌀5.0±0.2 THRU
[box]⏤[/box][box]⌀0.1Ⓜ[/box]
[box]⌖[/box][box]⌀0.5Ⓜ[/box][box]A[/box][box]B[/box][box]C[/box]

Interpreting this per ASME Y14.5-2009, would you say the position tolerance establishes a cylindrical boundary of diameter (5.0 - 0.2 - 0.1 - 0.5) = 4.2 which the surface of the hole is not allowed to violate?


pylfrm

 

[Burunduk]

If it was not critical it would not have a tolerance.

You may find reading "fundamental rules" (para. 1.4 in the 09' standard) useful.

I didn't insist that the calculated value must always be positive - I said "if the tolerancing makes sense" and should have clarified that it was in a context of the general case where assembly feasibility is considered in the datum reference frame at which the VC is calculated.

Despite missing that nuance, I did say it is up to the designer.
If in a given datum reference frame it is OK with the designer that when considering the possible displacements of the pattern of holes as a group, every point on the surface periphery of each hole can translate towards the true position axis of that hole a distance greater than the MMC radius, then a negative virtual condition calculated number is fine.

It is only similar to the independency in conjunction with the MMC position case if in the independency case too, the calculated value that results from the hole size and position tolerance is negative.

Also, I did provide a description of a case where the scheme of independency + virtual condition may be useful. Note that the issue was brought up by greenimi following a real case that was discussed: "...but after some heated discussions with a friend (initial inquiry is coming from his place of work)". greenimi asked about validity.

 

[3DDave]

I have. Perhaps you should point out what you think is the problem with them.

You continue to back fill what you should have said and what you really meant.

You have said that the VC represents a clearance for a mating part. Are there any negative value mating parts in your designs or will you back fill this as having yet again meant something else?

May be useful? At no point did you provide an example; at best you postulated without proof that one might exist. If you claim something exists you need to provide an example in order for others to believe you. Right now, I cannot believe you.

"I do realize this part has a tendency to deform and I don't require perfect form at MMC for the hole, but I do require that an empty space boundary of a size that allows a fastener to pass through is maintained at the basic location and orientation of the hole" describes a part subject to free state variation, not independency.

However, per your interpretation there is no difference if it is used or not. If the Independency allowed straightness is magically limited by the position tolerance (no direct proof in the standard,) then that is exactly the same as not using Independency at all; which means that using it alone violates those fundamental rules you should know about and it should not appear on a drawing in that capacity.

Looking forward to the next deflection onto some other aspect than what I've repeatedly demonstrated to be correct.

 

[greenimi]

Also, I did provide a description of a case where the scheme of independency + virtual condition may be useful. Note that the issue was brought up by greenimi following a real case that was discussed: "...but after some heated discussions with a friend (initial inquiry is coming from his place of work)". greenimi asked about validity.

I still maintain that is my intent, but I did not realize (lack of my proper education in this area) the level of complexity of my question and I 've fallen in Evan's "swamp".

Imagine a hole is specified as follows:

⌀5.0±0.2 THRU
⏤⌀0.1Ⓜ
⌖⌀0.5ⓂABC

Interpreting this per ASME Y14.5-2009, would you say the position tolerance establishes a cylindrical boundary of diameter (5.0 - 0.2 - 0.1 - 0.5) = 4.2 which the surface of the hole is not allowed to violate?

Interesting approach. I am not sure if that would solve the issue (in the same way pmarc's proposed -from a different discussion- solution could not be applied here, as pmarc mentioned above).

Or maybe does it solve?

 

[chez311]

As far as I am concerned, virtual condition is what the feature appears to be like within the allowed limits such that the intersection, for internal volumes, and union, for external volumes, of all possible variations is considered as a single volume. In other words, the virtual condition acts as if the feature was all the variations combined and it does so within the defined datum reference frame, if existent, and it is likely unique within that drf.

If it's the case that that combined feature is non-existent then there is no virtual condition to be had.

If the hole feature ends up in any manner of curlicue shapes such that the intersection of them is not a hole, then that not-a-hole cannot have a virtual condition. Being free of Rule #1 allows such curlicues in the absence of any other form controls.

3DDave,

Even if we are to interpret the combination of Independency and MMC position with only the resolved geometry interpretation, it doesn't seem to me that MMC position allows such curlicue shapes or other unrealizable geometry except in the cases of extremely large position tolerances in which case I would agree that absent rule #1 an additional form control is almost certainly required.

The allowable tolerance zone is:
b = t_0/2 + (size_UAME/2 - size_MMC/2)

and actual value is:
b = D/2

Where t_0 is the tolerance zone in the FCF and D is the "axis containing zone" fixed in location/orientation to the DRF. Freeing the feature from rule #1 would allow a size_UAME < size_MMC and therefore a value for b < (t_0/2) however I don't believe a value for b less than zero makes any sense - I would consider this along the lines of "unrealizable geometry". It seems to me the lowest value of b should be 0 as I don't believe a negative tolerance zone or actual value for the resolved geometry is interpretable and I would say should be rejected. Perhaps this is backdriving an indirect control of form, however it seems to me this is the only way to interpret the control per the resolved geometry interpretation.