Parallelism

[Kedu]

For the part shown use ASME Y14.5-2009 and determine what would be the maximum envelope for this part?

Will the parallelism of 0.003 (black option) versus 0.010 (red option) make any difference in calculation the maximum envelope?
(datum feature A is the same for both options = the bottom surface controlled with flatness within 0.005)

 

[drawoh]

Kedu,

How meaningful is a parallelism of .003 when it is referenced from a face that is flat to .005?

--
JHG

 

[drawoh]

Kedu,

Your 15+.040/-.020 is taken from your datum[ ]A face. If Rule[ ]#1 does not apply, and your datum[ ]A face is maximally concave, your maximum height could be 15.045.

This is an excellent case for a profile tolerance applied to the top surface.

I cannot see any design logic behind your asymmetric tolerance.

--
JHG

 

[drawoh]

Kedu,

How do you verify parallelism of .003 if your .005 flatness turns out to be a convex curve?

--
JHG

 

[Krumlauf]

drawoh,
I concur on the 15.045 regardless if the parallel tolerance is .003 or .010, but the top surface is only wrt |A| so just three highest points on each surface, no?

 

[greenimi]

May I ask how come the parallelism has no influence, whatsoever, over the maximum envelope of this part, as requested by the OP?

If rule#1 is not enforced, the part can be a banana shape, waisting or barreling effect of the surface could occur. If that's the case, and as far as I understand the geometry of this part, only the flatness callout would effect the envelope and not the parallelism (regardless of the value of the parallelism)

Am I correct?

 

[drawoh]

Krumlauf,

The test for the top surface is a dial indicator over the whole face. Three points are for the datum face only. If surface[ ]A is concave, then there are three points sticking down to contact the reference surface, and you have constrained three degrees of freedom. If surface[ ]A is convex, it constrains one degree of freedom, only.

A crude practical assumption is that the datum feature is flat, or at least very much more accurate than the tolerances being called up from it. If your datum face is not reliably flat, you need to make it flat, or work out datum[ ]targets.

The rule of thumb with tooling is that it must be ten times as accurate as the thing being fixtured or measured. If you want parallelism of .003 with respect to datum feature[ ]A, face[ ]A must be flat to .0003.


--
JHG

 

[greenimi]

If you want parallelism of .003 with respect to datum feature A, face A must be flat to .0003.

I do not think that is true.

 

[Krumlauf]

[Using Figure notations from ASME Y14.5-2009 as References]
As shown, Rule #1 must apply as flatness is only applied to the lower surface. If the flatness tolerance were applied to the feature of size (shown directly under 15+.040/-.020 [Figure 5-8]), then Rule #1 is overridden, but would still allow a size of 15.040 and .005 for flatness [Figure 5-7]. Whether the parallel tolerance shown is .003 or .060 (max allowed in this example), it only applies to the orientation of the upper surface relative to the lower surface (must have a datum reference) and is only measured by the distance between its 3 highest points to 3 lowest points [Figure 6-2].

 

[Krumlauf]

drawoh, both surfaces would require a dial indicator. Surface A requires a dial indicator over the whole face (relative to a controlled surface hopefully flat to within .0003) to verify flatness requirement is met before placing it on a controlled flat surface for further measurements.

 

[axym]

Kedu,

I would say that the parallelism tolerance would not affect the maximum envelope. I agree with drawoh that it would be 15.045 whether the parallelism is .003 or .010.

Regarding the tolerance amounts, the parallelism tolerance can be smaller than the flatness tolerance on the datum feature. So upper face can meet a parallelism tolerance of 0.003 even if the lower face is only flat within 0.005. This is because the datum plane is established from the plane that the datum feature sits on (surface plate, for example).

I agree that it is a good practice to make the form of datum features as good as economically possible. If the form of the datum feature is poor, then there is the potential for instability and rocking, and local gaps. Specifying a form tolerance (flatness) limits this. But there is no rule saying that this must be done - the flatness tolerance on the datum feature could be 10 mm and we can still specify a parallelism of 0.003 mm (or 0.0003 mm) on the upper surface.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[greenimi]

As shown, Rule #1 must apply as flatness is only applied to the lower surface

The note in black under the datum feature A symbol will cancel rule#1

If the flatness tolerance were applied to the feature of size (shown directly under 15+.040/-.020 [Figure 5-8]), then Rule #1 is overridden, but would still allow a size of 15.040 and .005 for flatness

Well the flatness is not aligned with the size dimension so no derived median plane flatness (DMPF) is applicable here. Rule#1 is canceled by the "black note" and not by the DMPF.
That is a moot point.

it only applies to the orientation of the upper surface relative to the lower surface (must have a datum reference) and is only measured by the distance between its 3 highest points to 3 lowest points

I do not think that is true ...the orientation (meaning here the parallelism) means the entire surface (all the points of the surface) and not only 3 highest (or 3 lowest) points, must be between a tolerance zone of .003 or .010. Unless my understanding of your above sentence is inadequate, I guess your statement is wrong.

 

[axym]

greenimi,

I agree with all three points from your last post.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[drawoh]

...

Regarding the tolerance amounts, the parallelism tolerance can be smaller than the flatness tolerance on the datum feature. So upper face can meet a parallelism tolerance of 0.003 even if the lower face is only flat within 0.005. This is because the datum plane is established from the plane that the datum feature sits on (surface plate, for example).

Here is the problem.

If the bottom face is concave, it sits on the reference surface and it provides a datum. If the bottom face is convex, the part rocks when the dial indicator is run over the top face. We need the bottom flatness to be very much better than the top parallelism.

Note how the thickness tolerance works. I am looking at ASME Y14.5-2009 Figures[ ]2.7 and[ ]2.8.

--
JHG

 

[greenimi]

If the candidate datum set condition is applicable by default (and IT IS applicable if you are using the default condition within ASME Y14.5-2009) then as long as long you can find A POSITION (one rocking place) where you can make all the points of the upper surface within the tolerance zone (of .003 or .010) then the parallelism condition is achieved and the surface in question is qualified as being in the tolerance---regardless of other rocking places---

In order words you are ALLOWED to optimize.

And the point of optimization is to find a place that works and not to demonstrate that there are plethora of places that do not work.

 

[axym]

drawoh,

I agree that the convex condition you showed in the lower sketch is a problem. But greenimi is right - the parallelism would conform by the rules of Y14.5-2009 and Y14.5.1M-1994. This is why it's a good practice to specify a much tighter flatness tolerance for the datum feature as you suggest, even though it's not mandatory.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[Krumlauf]

So sorry greenimi, I missed the perfect form at MMC note!! Agreed, Rule #1 is cancelled.
Regarding the 3 highest/lowest points comment, yes, I should have clarified by stating all points of the surface should be contained within two parallel planes each constructed by either the 3 highest or lowest points of the surface.

 

[pylfrm]

I would say that the parallelism tolerance would not affect the maximum envelope. I agree with drawoh that it would be 15.045 whether the parallelism is .003 or .010.

Could you (or anyone else who agrees with the quoted statements) please describe the shape of a part that meets the requirements of flatness, size, and .003 parallelism, but has UAME size of 15.045?


Kedu,

I am curious: Why do you ask this question? If the UAME size is relevant, why wouldn't it be controlled more directly?


pylfrm

 

[drawoh]

axym,

Try out this one.

Look closely at the OP's tolerances. This is a feasible outcome given the sloppy tolerances on the thickness. You can force it to work, but is this a good idea?

--
JHG

 

[chez311]

Could you (or anyone else who agrees with the quoted statements) please describe the shape of a part that meets the requirements of flatness, size, and .003 parallelism, but has UAME size of 15.045?

Something like the below, unless I'm missing something (deviation is exaggerated on the figure due to scale)? Seems to me the form control (flatness) on the bottom surface limits the maximum envelope/UAME size since rule #1 is overridden. Flatness of 0.1 on the bottom surface would allow a maximum envelope/UAME size of 15.14, it doesn't seem to me this is at all impacted by the parallelism control on the top surface.