Formulate Diameter and Roundness

[UchidaDS]

For example:
The drawing callout of the part has a diameter of 10.000 +/-.005"
Using CMM to measure this part:

Diameter 10.003"
Roundness is .011".

Does anybody know how these number formulate?

 

[CheckerHater]

To the best of my understanding, variation of circular shape cannot be greater than size tolerance.

Is that "formulaic" enough?

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[UchidaDS]

Oh, I understand the roundness should be within size tolerance.
I want to know how does the roundness "calculated" or obtained.
In addition, is that mean the diameter is the average result and the actual size become 10.003 +/-.0055?

 

[Belanger]

Roundness is calculated using one of four algorithms: minimum radial separation, maximum inscribed circle, minimum circumscribed circle, or least-squares circle. Each may yield a different result!
Are you using ISO or ASME? That could make a difference about which one is the default algorithm.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[UchidaDS]

ASME.
But is my interpretation correct?
-->that mean the diameter is the average result and the actual size become 10.003 +/-.0055?

 

[Belanger]

For the roundness check: ASME Y14.5 rules say that "minimum radial separation" is the default algorithm. For more specific details, check out the standard ANSI B89.3.1.

I don't know what the intent is on that report of diameter. If that's the only number given then I'd question it. Maybe they mean that 10.003 is the maximum diameter -- trying to show that it's still within spec. But it's worth asking the people who provided that report, since it's not in a standard format that I know of.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[CheckerHater]

We haven't figure out the most important question: ISO or ASME?

Given that "roundness" is ISO term and ASME now has "circularity"

And who knows which standard the CMM follows? Sometimes it's own...

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[Belanger]

He said ASME. (see 5th post)

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[CheckerHater]

There still question of CMM.

The size could be some sort of average, like least squares size. Then "roundness" tells us how far the part extends from that average.

I guess the only way is to ask CMM people exactly "how these numbers formulate?"

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[UchidaDS]

As I know it use LSQ-->Least square method.
So the number of the diameter is average?

 

[Belanger]

We don't know. It should have been recorded as a maximum/minimum. But knowing that the roundness error is .011, I think there is a problem with that part. (Roundness values are to be within the size tolerance, unless the print specifically says AVG (average) next to the size designation.)

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[drawoh]

UchidaDS,

I am trying to remember my college metrology here. There are shapes that measure out to accurate diameters, that are way out of round. Try sketching out a three-lobed shape.

Is it possible your diameter was measured at one point, only?

--
JHG

 

[UchidaDS]

I understand the three lobes.

The diameter was measured 12 points, uniformly distributed, using CMM.
I hope I interpreted correct: It seems for me that the diameter 10.003 is an average value. Roundness represents this result's bilateral tolerance.
I think I badly worded this subject...