DML straightness material condition modifiers and relation to position/orientation at MMC 1

[Burunduk]

When the straightness tolerance at MMC is used in conjunction with an orientation or position tolerance at MMC, the specified straightness tolerance value shall not be greater than the specified orientation or position tolerance value and does not contribute to the IB or OB of the position or orientation tolerance. The collective effect of the MMC size and form tolerance produces a VC, OB, or IB resulting from the form tolerance but does not affect the IB or OB created by any orientation or position tolerances on the feature.

Why is the above rule of not specifying larger tolerance for straightness than for position/orientation applies only to DML straightness @MMC & Position/orientation @MMC, and does not include DML straightness @RFS & Position/orientation @MMC?

I think that if DML straightness @RFS & Position/orientation @MMC is applied the size of the unrelated AME may be greater than MMC size, however we wouldn't want it to become greater than the VC for Position/orientation @MMC, since we want the surface of the cylinder to be limited by the VC gage, without any unrealizable form allowance - same logic as for not wanting the VC for DML straightness @MMC to become larger than the VC for Position/orientation @MMC that will limit the surface anyway. Is this not the guiding logic?

 

[pmarc]

3DDave, Burunduk,

Not sure where the 'position' came from in your replies, but thank you anyway.

 

[3DDave]

pmarc - doesn't matter if it is position or perpendicularity or angularity or parallelism. Not sure why you would not have specified to make it clear.

You can hit multiple spots on the GDT Battleship game with this one, explaining why the straightness cases of less-than, equal-to, and greater-than make sense.

 

[pmarc]

You can hit multiple spots on the GDT Battleship game with this one, explaining why the straightness cases of less-than, equal-to, and greater-than make sense.

Feel free to do it for me so that those interested in participating in the game know how to do it properly. Sort of example in the instruction.

 

[3DDave]

pmarc - already did. There was a link to a suitable example in the instructions.

The chart in the game should be what the committee has been working to since the 1980s. Better late than never.

 

[Burunduk]

Not sure where the 'position' came from in your replies

Right, that should have been perpendicularity. But my answer is the same - the comparison of the VC sizes indicates what I think is considered the problem with such schemes. My interpretation regarding this rule was always that it is intended to prevent unusable allowance. Sort of like you're not supposed to assign a form tolerance larger than the size tolerance (other than DML straightness/ DMP flatness), so that the entire form tolerance can be utilized. Am I missing something?

 

[pmarc]

Burunduk,
I think the problem you described wouldn't exist if the rule in the standard was that the size of the orientation or location constrained VC of a feature controlled with an orientation or location tolerance @ MMC was:

For external features:
VCconstr = VCunconstr + orientation/location tolerance value @ MMC

For internal features:
VCconstr = VCunconstr - orientation/location tolerance value @ MMC

Where:
VCunconstr is equal to:
-- MMC size if no DML straightness or DMP flatness specified
-- MMC size + DML straightness or DMP flatness value, if specified, for external features.
-- MMC size - DML straightness or DMP flatness value, if specified, for internal features.

The formulas for VCunconstrained would apply regardless if the DML straightness or the DMP flatness were defined at MMC or RFS.

 

[Burunduk]

pmarc,
This would make sense in a way, but it could create a new problem:
It would force the enlargement of the VC once DML straightness/ DMP flatness is assigned & rule #1 is overridden, whereas the VC protects a functional boundary.

I think the current state of affairs is generally not that bad.
As it is right now, at least the way I see it, whenever DML straightness or DMP flatness is specified, there is an "effective size" that includes the actual size, plus (for an external feature) the out-of straightness/flatness. Essentially, that "effective size" is the unrelated AME. That value plus the orientation/position error must not be greater than the orientation/position VC for a passing external feature. Under that condition, whenever u-AME > MMC & the entire position/orientation tolerance is used, the pin/tab's size limits can't be fully utilized and the actual local size can't be the MMC size along the entire feature (or it will make the feature be rejected for violating the position/orientation VC). But, the designer may not care. The manufacturer may either utilize the entire limits of size range and keep it within the boundary of perfect form @MMC or violate that boundary and utilize the full form allowance, but under any circumstances the functional VC corresponding to the chosen MMC size plus the chosen position/orientation tolerance can't be violated. This actually allows some flexibility in different types of variation that may be allowed without compensating the function of the part.
Thoughts?

 

[pmarc]

Burunduk,
I am not sure what you mean by "whereas the VC protects a functional boundary".

In my earlier pin example, the functional, orientation related, VC boundary for the pin is dia. 10.2. If I am OK with allowing the actual pin to have all local sizes at 10.1 and the unrelated AME size greater than dia. 10.1 as long as the orientation related VC boundary does not get violated, then why should the standard be stopping me from doing that?

Note: As a matter of fact, since I am initially proposing the use of perpendicularity of zero at MMC in my example, I think the same end result could be accomplished by replacing the DML straightness of 0.1 @ MMC with the independency modifier and changing the perpendicularity tolerance value from 0 to 0.1. On one hand, this would allow me not to get in conflict with the standard, but on the other hand would, in my opinion, highlight the artificiality of the existing rule.

 

[Burunduk]

If I am OK with allowing the actual pin to have all local sizes at 10.1 and the unrelated AME size greater than dia. 10.1 as long as the orientation related VC boundary does not get violated, then why should the standard be stopping me from doing that?

But the orientation related VC is dia. 10.1 (based on 10.1 MMC & 0@MMC perpendicularity), so how can it contain a pin of unrelated AME size greater than 10.1? Looks like even though the functional limit is dia. 10.2 as you say, a functional gage for perpendicularity will reject anything that comes out larger than a 10.1 orientation-constrained boundary, either due to the combination of size+ form+orientation, or size+form alone with or without violation of perfect form @MMC. The rule in the standard, as far as I understand, basically says that you should be able to utilize the entire specified size+form allowance without having to reject the part by an orientation/position VC.

 

[pmarc]

But the orientation related VC is dia. 10.1 (based on 10.1 MMC & 0@MMC perpendicularity)

It is dia. 10.1 because the standard simply says that:
VC_rel = MMC size + specified perpendicularity tol.

The whole point of my "food for thought" part of the thread was to provoke a conversation about the formula and hear arguments against the following modification of it:
VC_rel = VC_unrel + specified perpendicularity tol.

 

[3DDave]

pmarc - a general rule that should be used is "Accept what you tolerate." Previous discussions about why a non-circular Virtual Condition can result from using diametral tolerances show that "what it says" trumps "what it actually does" for some users.

In software terms you are applying the orientation tolerance to the combination of size and form tolerance rather than, as the standard says, that the form tolerance is entirely dependent as a refinement of the orientation tolerance.

So if one wants to tolerate a larger form tolerance then that contribution should be accepted in determining the virtual condition for the orientation/position tolerance.

 

[pmarc]

3DDave,
Yes, I agree with that.

In my mind, that's one of the biggest problems in the quest for solid mathematization of the rules given in Y14.5. In some places, the standard defines its own math that simply doesn't go along with the underlying principles it established.

 

[3DDave]

pmarc - It's why I press for a simulation based description rather than just the patchwork of rules and pictures. If one can describe how something works to the ultimate idiot - a computer - then it's a description anyone can use, even if only via that simulation. The infamous GIDEP alert that showed CMM makers were not following the standard resulted in improving the CMM software and -poof- CMM operators everywhere didn't need to know Y14.5.1 to do their job.

Google for "M Berta, Z Humienny" to see this applied to ISO.

 

[pmarc]

I would be really surprised if the software worked the way you would want it to work.

 

[3DDave]

pmarc - here's the thing. No matter what one imagines software will do, it does what it does. So if one is intending to add the first 10 numbers in a list and they don't understand the process the software will add the first 9 or the first 11, or just crash. But after examination of that, which software developers tend to do, at some point an understanding dawns.

This seems to rarely happen with those who read the standard, think they can do what they want, and apply their imperfect understanding to analysis. It doesn't give them the wrong answer or crash - it gives them the answer they expect and that is dangerous. That same process, unrepeatable, is done over and over as the drawing moves through the organization. People then argue over who has the correct understanding, they are bullied to accept the wrong one and, sometimes, they get it right, but there isn't an easy way to determine which is actually happening.

In my projected tolerance example anyone can examine the construction, see the equations, note that the actual width of the tolerance zone doesn't match the assigned width because of sine error, but it does match what the standard says should happen.

Or, as the collapsed bridge recently in the Accidents section - they can wait until it goes wrong and realize they were clueless.

If people won't try anything then nothing will work, but the present training isn't working too well either. Beyond lining up holes the geometry becomes too complicated and without a mental simulation based on experience most never seem to get anything more than the trivial copy/paste examples.

 

[Burunduk]

It is dia. 10.1 because the standard simply says that:
VC_rel = MMC size + specified perpendicularity tol.

The whole point of my "food for thought" part of the thread was to provoke a conversation about the formula and hear arguments against the following modification of it:
VC_rel = VC_unrel + specified perpendicularity tol.

In my opinion that may add a lot of confusion because if the formula is changed this way, 0@MMC for perpendicularity of the pin in your example would no longer mean zero tolerance when the unrelated AME reaches the diameter of the MMC limit of size, but rather zero tolerance at the state when the worst case boundary for size + out-of straightness is realized. And that phenomenon will only take place when DML/DMP form tolerance related exceptions to rule #1 take place, otherwise it's business as usual. Possibly to clear up the newly created confusion, the committee will want to add some more terminology - how about "Zero tolerance at MAME (Maximum Actual Mating Envelope)"? It will be a good step for catching up with ISO on the quantity of concepts and abbreviations. Also, I suspect 3DDave will like another opportunity to point out that the committee makes an update to gain more content for new training materials.

 

[pmarc]

Burunduk,
This would only require a change in the definition of MMC

 

[Burunduk]

pmarc,
"Only" a change in the definition of MMC doesn't sound like a small thing.
If the definition of MMC was changed what would be the effects on the Envelope Principle (Rule #1) and all the related things like the exceptions to that rule, form tolerance values other than for median line/plane (their relationship to what is currently the MMC), determination of material boundary sizes for true geometric counterparts referenced @MMB...?
Changing the MMC definition means starting a chain reaction by altering a fundamental concept in ASME product definition.
Additionally, if you are getting at redefining MMC to include the derived median line/plane form variation, then I'd say that this would go against the core idea of the "Maximum Material Condition". That pin from your example with the size limits of 9.9-10.1 & DML straightness controlled within 0.1@MMC, if it's made banana-bowed to be inscribed in an unrelated envelope of 10.2 that happens to equal to the upper limit of size plus the DML straightness tolerance, doesn't necessarily contain more material than another pin that occupies a diameter of 10.1 and may be perfectly straight. In fact, it surely may contain less material.

 

[pmarc]

Burunduk,
Yes, it wouldn't be a small thing. I quickly proposed this because I sensed sarcasm in the second part of your reply. My apologies for that.

As a side note, notice that the way geometric callouts at MMC (or LMC) are commonly interpreted today doesn't really follow the current MMC (or LMC) definition as well. A zero @MMC perpendicularity tolerance on a dia. 10 +/-0.1 hole doesn't mean zero when the part containing the hole has "maximum amount of material". It means zero when the UAME size of the hole is 9.9, which does not necessarily mean the part containing the hole must have maximum amount of material. So from that perspective the current MMC definition adds little value.

 

[Burunduk]

pmarc,
I should have understood that when you said "This would only require a change in the definition of MMC" you were sarcastic too. Sorry for missing that.

Your side note is correct, but there is an exception to that: DML straightness or DMP flatness specified MMC (or LMC).
In these cases, each cross-section is evaluated for departure from MMC (or LMC) in the true sense of "material" (amount) condition, and gets a local bonus for form deviation according to the local size.
Axis or center plane controls do work differently, though.