Measurement between 2 planes ?

[Rwelch9]

Hi all,

If you have 2 planes features which create a mid plane Datum B. If the print asks for the planes to be for example 50mm width ± 0.02.

Where do you take the measurement for the 50mm width ?

These planes are perpendicular to my Datum A.

So I thought I would have two simulators perpendicular to Datum A, which would contract onto both planes at the extremes. This then would allow me to check the 50mm width and create the mid plane Datum B.

There is no parallelism call out attached.

Thanks

Ross

 

[chez311]

Verification of a +/- directly toleranced width is not impacted by use of a feature as a datum feature. The feature must also satisfy any geometric tolerances applied, but they do not change how the features width is measured.

The procedure you describe sounds like it would give you the RAME size (simulators of perfect form contracted about the feature oriented to A). A similar procedure, but not oriented/located to any datum features would give you the UAME size. Neither of these are "size" per ASME - 2 point measurements will give you an approximation but the swept spheres interpretation per Y14.5.1 is technically the final word.

 

[greenimi]

You need your mid plane only if you want to qualify perpendicularity (if one available) of 50mm to datum plane A.
The size 50±0.02 has not much to do with the mid plane simulation.
Could you post a picture, if I am missing your question?

What standard are you using? ASME ? ISO? Which?
If ASME then 2 points measurements and rule#1 should be check for.

 

[Rwelch9]

[img

https://res.cloudinary.com/engineering-com/image/upload/v1599235780/tips/Plane_Measurment_fiurpp.jpg

Chez 311 + greenimi

ASME is the standard we are working to.

My main question is where exactly do your report the 50mm width from. Would it be the extremes high points of both planes ? created by the simulators which would perpendicular to Datum A ? I would then take the 50mm measurement off the simulators ?

Thanks

Ross

 

[chez311]

created by the simulators which would perpendicular to Datum A ?

No. As I said previously that would be your RAME size.

Two point measurements can be used (think caliper or micrometer) however this is technically only an approximation of conformance per the math standard (Y14.5.1) which introduces the concept of swept spheres (feature must fit within a volume created by spheres of MMC/LMC size swept along a valid spine - this spine has perfect form at MMC if MMC/RFS used and rule#1 in effect or at LMC if specified LMC). As pointed out many times on this forum this swept spheres interpretation is typically ignored and 2 point measurements are used. Or often the UAME size is measured, which for a feature subject to rule#1 will at least tell you whether you conform to the MMC limit at least, and for a feature with reasonably good form may give you some amount of confidence in the LMC limit as well.

 

[Tarator]

For a regular feature of size (rFoS), you will have to measure the following:

1) Actual LMC size (2-point measurement),
2) Actual MMC size (2-point measurement,
3) Unrelated Actual Mating Envelope (UAME) size as per Rule #1.

So in your report, you should have these 3 values. For your part with relatively tight size tolerance, these values will most likely be very close to each other.

 

[Burunduk]

As others pointed out, you don't need a datum-constrained simulator to evaluate size limits conformance.
What you should be interested in are Actual Local Sizes and Rule#1 conformance.
I agree with Tarator's suggestion about 3 measured values, but since the largest Actual Local Size of the feature can never be greater than its Unrelated Actual Mating Envelope (it's geometrically impossible), you can consider only evaluating the size of the UAME and checking if it's smaller than the MMC value (thus ensuring both that Rule#1 is not violated and the maximum size limit) and the smallest ALS for the minimum size limit. So that's two values instead of three and conformance is still guaranteed. If you want to follow standardized rules, check thread1103-471922 where a draft of the new standard on measurement reporting was posted. Do tell what you find out on size reporting if you check it (I haven't checked myself yet)

 

[chez311]

since the largest Actual Local Size of the feature can never be greater than its Unrelated Actual Mating Envelope (it's geometrically impossible)

Actually I wouldn't be so quick to say that. Largest actual value of size per the swept spheres interpretation yes. Largest "actual local size" not necessarily. For a feature with a reasonable size tolerance and very small amount of form error the maximum discrepancy would likely be pretty small but strange things could happen if the tolerance limits are opened up. For a simplistic explanation as to why, just consider that "actual local size" whether evaluated by a 2 point or circular sections* method is a 2D evaluation - it need only be taken perpendicular to a "local size spine" - a 2D line (two point) or circle (circular section) of a size>UAME size is possible if the spine normal at that point is not also normal to the UAME. This is not possible for a sphere, as you cannot fit a sphere of size>UAME size.

*I know OP is about a width shaped feature. The circular sections is another method presented by the standard obviously for circular/cylindrical features.

 

[Tarator]

Generally speaking, it is POSSIBLE for the actual MMC size to be NOT between the actual LMC size and the UAME size, especially if you have relatively large size tolerance and large deformations (due to lack of clear definition of actual local sizes and the methods of their measurement).

The benefit of reporting the 3 items is to give information about the actual part and how they behave, especially over a long period of time. It will give you information about the process, etc.

 

[Rwelch9]

Hi all,

Thanks for all the help and advice.

My issue was i actually had components with a two point measurement either with a micrometer or my CMM was just under bottom limit. Taking planes on the CMM, average ( Least squared plane ) at each side the part was in spec and was well within spec at the highest or extreme points of the plane again using a two point measurement using both the micrometer and the CMM.

I gathered the full plane features at any two point measurement had to be within tolerance. I just wasn't sure if being datum features this changed the approach in an way. Especially as parallelism was not called out either. To further this generally all I have to report is 1 size, I wasn't sure what is the actual measurement between two planes were as the are obviously many sizes you can report depending where you measure.

Thanks

Ross

 

[Burunduk]

chez311, Tarator,
I stand corrected regarding "geometrically impossible", although I can only imagine a maximum actual local size larger than the UAME for perhaps a thin plate with a relatively large size tolerance and thus allowed a large form variation that can cause a condition like that. Seems unlikely for 50+/-0.02 size feature with just 0.04 form variation available. Still 3 values per Tarator's suggestion is probably the more correct suggestion.

Rwelch, if you measured Actual Local Sizes under the minimum limit then the part is considered non-conforming, but you may want to attempt to see if the swept spheres concept (as described by chez311 on 4 Sep 20 16:34) produces more favorable results. As mentioned this definition is considered the more obligating for a drawing made per Y14.5, although it's a different concept from the Actual Local Size which is covered by Y14.5 itself. Also, if you intend to report just one value, then I suppose it needs to be the "Actual Value" related to the same concept (swept spheres), which is the diameter of the largest sphere you can sweep along a spine within the feature. Still, the LMC and MMC limits need to be respected too as well as Rule #1 so at least a pass/fail report is needed on these requirements. Does your CMM software support the swept spheres calculation? See the mathematical definition ASME Y14.5.1M-1994 for more details.

 

[aniiben]

Actually I wouldn't be so quick to say that. Largest actual value of size per the swept spheres interpretation yes. Largest "actual local size" not necessarily. For a feature with a reasonable size tolerance and very small amount of form error the maximum discrepancy would likely be pretty small but strange things could happen if the tolerance limits are opened up. For a simplistic explanation as to why, just consider that "actual local size" whether evaluated by a 2 point or circular sections* method is a 2D evaluation - it need only be taken perpendicular to a "local size spine" - a 2D line (two point) or circle (circular section) of a size>UAME size is possible if the spine normal at that point is not also normal to the UAME. This is not possible for a sphere, as you cannot fit a sphere of size>UAME size.

Chez,

Do you have a sketch handy to show the differences between 2 points and sphere concept?