Orientation & Size Limits

[JLang17]

This may be the simplest question ever, but are perpendicularity and/or angularity restricted by size limits?

I'm looking through Section 6 of Y14.5-2009, but it only mentions parallelism as being restricted to limits of size. If parallelism is, then why not perpendicularity?

A question would be, if I have an "L" shape, can the horizontal surface angle "down" beyond the size limits of the vertical surface?

 

[looslib]

NO
The size limits apply at all times.

"Wildfires are dangerous, hard to control, and economically catastrophic."

Ben Loosli

 

[powerhound]

Actually the answer is not as simple as "no". If you only have an L shape drawn with no GD&T or datums specified BUT you have invoked the ASME standard as the standard to interpret the drawing to, then what applies is the implied 90 degree rule. The part can be at the top of the tolerance PLUS it can be out of 90 degrees by whatever the tolerance block shows as a default tolerance for angles. There is no frame of reference by which to measure it's envelope or orientation without datums.

The reason perpendicularity is not mentioned is because perpendicular elements do not constitute features of size and thus are not restricted to the limits of size rules.

If you are specifying perpendicularity then you will have to determine a datum to specify it to. The perpendicularity must be a refinement of the implied 90 degrees. It will also be independent of the vertical dimension and should have no bearing on the vertical dimension. They should be considered 2 separate things. To be clear, perpendicular elements are not features of size.

Powerhound, GDTP T-0419
Production Manager
Inventor 2009
Mastercam X3
Smartcam 11.1
SSG, U.S. Army
Taji, Iraq OIF II

 

[JLang17]

Using the "L" shape as an example:
The vertical surface is Datum A and dimensioned 3.5 +/- .02
The horizontal surface is called out perpendicular to A with a value of .01

So if the vertical is manufactured at 3.7, the horizontal is allowed to bend "down" beyond the 3.7 dimension within it's perpendicular value?

 

[desertfox]

Hi JLang17

Your vertical surface can vary between between 3.52 and 3.48 according to the tolerances given in your post however the horizontal surface must be at right angles to within 0.01 and you cannot use the tolerance band given for the vertical height and add to the perpendicular tolerance which I think is what your attempting to do.

desertfox

 

[JLang17]

Yes I meant 3.52, not 3.7, I better get some more coffee...

 

[axym]

The intent is that an orientation control must be a refinement of any other controls that indirectly control the orientation of the feature.

In the case of a slot, a size (width) tolerance would indirectly limit the parallelism error of one side of the slot with respect to the other. If the size tolerance was +/- .005, the worst possible parallelism would be .010. A parallelism control applied to one side of the slot with respect to the other must refine i.e. have a tolerance value smaller than .010.

There are no cases in which a size tolerance indirectly controls perpendicularity or angularity.

JLang17, I couldn't find a reference in Y14.5-2009 where it says that parallelism is restricted by size limits. Where are you looking?

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[desertfox]

Hi JLang17

I think if your bracket was say 3.48 high then your perpendicularity could make the bracket at some point 3.49 however if the bracket measured 3.47 then it would fail on the size limits does this answer your question.

desertfox

 

[JLang17]

axym,

I was looking at Fig.6-2. My question arose when I looked at Fig.6-3 and there was no dimension or size limit involved.

I always thought a width dimension, or height, applied to the entire surface the dimension is extended from. In the case of the "L" this would mean the far end of the horizontal surface could not surpass the limits of the vertical dimension, thus perpendicularity would be restricted.

 

[ewh]

That was my take on it also... An angle with a height dimension and a width dimension would have perpendicularity indirectly controlled by the dimension tolerances. They define an envelope within which the part must fit.
So, how is it that these dimensions {i]do not[/i] indirectly control perpendicularity?

"Good to know you got shoes to wear when you find the floor." - [small]Robert Hunter[/small]

 

[axym]

I didn't initially see how sizes on an L-shaped part enter into this, but after reading ewh's post I think I see what's happening. There is a common misconception that size tolerances on two features will indirectly control the relative orientation of the features. This isn't true - at least not in Y14.5. Here's an excerpt from Section 2.7.3 in '94 (2.7.4 in '09):

Relationship Between Individual Features

The limits of size do not control the orientation or location relationship between individual features. Features shown perpendicular, coaxial, or symmetrical to each other must be toleranced for location or orientation to avoid incomplete drawing requirements. These controls may be specified by ...

So on the L-shaped part, the width and height tolerances don't control the angle between the vertical surfaces and the horizontal surfaces at all. The part could be made at a 45 degree angle and still pass the size tolerances. The angle would have to be controlled by something else - often there is a default angle tolerance of +/- half a degree or something like that. If that is not adequate, a perpendicularity tolerance can be specified as in Figure 6-3 of '09 or Figure 6-34 of '94.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[desertfox]

Hi axym

But even if the faces were out by 45 degrees on the L shape they would still have to be within the height range of 3.5 +/-0.02 otherwise they would fail inspection.

desertfox

 

[JLang17]

If we have a rectangle with +/- toleranced dimensions, then the rectangle can end up being a trapezoid or parallelogram due to the tolerance zones. Orientation tolerances can tighten these zones, but can they work the other way and loosen them? i.e. an orientation tolerance greater than the dimension tolerance.

 

[axym]

JLang17,

If we have a rectangle with +/- toleranced dimensions, the rectangle can end up being trapezoid-ed (to a limited extent) due to the size tolerance zones. The opposing sides can converge (i.e. be have parallelsm error) but only to the extent allowed by the size tolerances. A parallelism tolerance would have to refine the indirect parallelism control provided by the size tolerance, and therefore would need to have a value less than the size tolerance.

But the part can be a parallelogram-ed to an unlimited extent, because the size tolerances don't control the "corner angles". Not directly, not indirectly - there is no control whatsoever. So a perpendicularity tolerance isn't a refinement of the size tolerance. It provides perpendicularity control where there otherwise would be none. So the perpendicularity tolerance value could be greater than the size tolerance.

I think we need a diagram.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[JLang17]

See attached jpeg.

Does the 3.56 dimension not matter?

 

[JLang17]

Woops! Datum A is suppose to be the vertical surface.

 

[ewh]

Ah, but wouldn't an angular general tolerance be the default which would control it ([¶]1.4j of the '94 standard)?

"Good to know you got shoes to wear when you find the floor." - [small]Robert Hunter[/small]

 

[ewh]

oops... I meant to say [¶]1.4(i).

"Good to know you got shoes to wear when you find the floor." - [small]Robert Hunter[/small]

 

[JLang17]

Consider the .04 the implied 90 tolerance (or it's equivelent in decimal terms), then we still have the issue of whether the size tolerance matters.

The problem I'm seeing is - when do we consider the dimension to include the entire surface it extends from, and when do we not?

 

[ewh]

I agree with looslib; you always have to take the dimensional limits into account first. The 3.56 measurement on you part should cause it to fail inspection. If the part meets the dimensional limits, then you apply the indirect limits. If it doesn't meet the dimensional limits, you don't apply the indirect limits to try to get a good part.
If the part had a very short leg and generous tolerances, it could pass the dimensional limits, but fail the indirect ones (angular tolerance).

"Good to know you got shoes to wear when you find the floor." - [small]Robert Hunter[/small]