Circular runout 2

[gabimo]

Is circular runout controlling:
1.) Straightness of the cylinder surface?
2.) What about Derived Median Line (DML)?

Same questions for total runout:
Is total runout controlling:
3.) Straightness of the cylinder surface?
4.) What about Derived Median Line (DML)?

In other words, if a cylinder surface (surface feature A) has a circular or total runout shown (in relationship to a datum axis defined by other cylindrical surface, feature B) and a straightness callouts or a derived medial line callouts for the surface feature A are also shown, does the straightness or DML should be smaller than the circular/ total runout?

 

[pmarc]

mkcski,
If on ISO drawing a feature of size does not have any form conrol applied (either directly through tolerance frames, or indirectly - for example by invoking a standard with general geometrical tolerances), then the form of that feature of size is not controlled at all. Is this what you were asking for?

 

[pylfrm]

derived median line: an imperfect (abstract) line formed by the center points of all cross sections of the feature. These cross sections are normal (perpendicular) to the axis of the unrelated actual mating envelope.

How is the center point of a cross section determined?

pylfrm

 

[greenimi]

"To derive the median line, one must (at least in theory) take and "infinite" number of cross-sections, find the centers of each of those cross sections, and connect the centers together. The result would be an imperfect "spline" (like a noodle). So, you could have a part with large surface straightness and perfect DML straightness. Picture "barrel-shaped" pin. So, surface straightness controls DML straightness, but DML straightness does not control surface straightness. As long as the out-of-straightness of the surface is symmetrical, the DML stays straight."

"DML straightness (done correctly) can be a time-consuming thing to measure, but can be done with a CMM. Often we will just "estimate" (and what measurement isn't just an estimate anyway) the DML straightness by measuring the minimum circumscribed cylinder diameter (can be done with a comparator or a CMM) and take the difference between that value and the local diameter measurements. Often, DML straightness is specified with an MMC modifier, so then it could simply be gaged (or checked with a dial indicator and some gage blocks."

---per Mark Foster on linkedin website

 

[greenimi]

And a question for pmarc,

Why D1, B2, I1 and G2 are not considered redundant (or in conflict / incompatible)?

Just me trying to understand.

 

[mkcski]

pylfrm:

Regarding finding the center point:

Y14.5 is a product definition standard and consequently does not define manufacturing or inspection methods. Given too that there is no discussion about how to determine the center at each cross section, I could go in several directions to find the center points and "connect the dots". I could suggest: 1) the axis of the smallest circumscribed circle; 2) a complicated surface analysis to "remove" form error (like for Concentricity), 3) a best material condition FIM using a dial indicator and a select number of opposing locations, 4) best-fit circle using CMM algorithms ....take your pick.

Certified Sr. GD&T Professional

 

[3DDave]

Determination of the center should be covered in Y14.5.1

 

[pmarc]

How is the center point of a cross section determined?

This is a good question. And I wouldn't agree with mkcski here. Although it is true that Y14.5 is a product definition standard, certain things should/must be defined in it. If there is a definition - Derived Median Line in this case - that says that certain center points must be found in order to determine/find the actual DML, the definition should clearly specify (either directly or indirectly by giving a reference to another document) how to establish the center points. This is not what is currently happening in Y14.5-2009. What is the reason that the committee was able to give clear definition of, for example, Feature Axis (which is of course extremely useful in determining actual position or orientation errors), but was not able to offer a hint on how to find the center points constituting the DML? This I would really like to know.

The interesting thing is that in Y14.5M-1994 the definition of the DML had a statement: "The cross section center points are determined as per ANSI B89.3.1". This is the standard for Measurement Out-of-Roundness amd in that document four different methods of establishing the center point have been defined, as given in para. 2.8:

2.8 Centers for Out-of-Roundness Measurement
The centers of the measured polar profile which may be used to determine the out-of-roundness value when specified are those related to one of the following alternative methods of out-of-roundness assessment:

2.8.1 Minimum Radial Separation (MRS). Tis center is that for which the radial difference between two concentric circles which just contain the measured polar profile is a minimum¹.

2.8.2 Least Squares Center (LSQ). This center is that of a circle from which the sum of the squares of the radial ordinates of the measured polar profile has a minimum value.

2.8.3 Maximum Inscribed Circle (MIC). This center is that of a largest circle that can be inscribed within the measured polar profile².

2.8.4 Minimum Circumscribed Circle (MCC). This center is that of the smallest circle which will just contain the measured profile³.

¹This is also known as the center for minimum Total Indicator Reading (TIR). The British Standards Institution publication 3730:1964 refers to it as Minimum Zone Center (MZC).

²This is also known as the plug gage center and is generally used for internal diameters.

³This is also known as the ring gage center and is generally used for external diameters.[highlight #F57900][/highlight]

I am not sure why exacly pylfrm asked the question about the center points, but with all of what I just said, I don't think that different methods of establishing the center points (potentially leading to different actual DMLs) could change anything in the answer to greenimi's question.

And a question for pmarc,

Why D1, B2, I1 and G2 are not considered redundant (or in conflict / incompatible)?

Why are you only asking me? If I am seeing correctly, mkcski gave you exactly the same answer.

But seriously, the reason these four cases are ok is because it is possible, although quite unlikely to happen, to have a cylinder produced with maximum allowable DML straightness error, say .010, that will also give the actual circular/total runout reading of .010. This can happen when the axis of the Unrelated Actual Mating Envelope (UAME) of such cylinder is perfectly coaxial with the datum axis (the cross-sectional form of the cylinder would also have to satisfy certain requirements).

 

[3DDave]

Apparently B89.3.1 has it's own set of FCF modifiers. I haven't got a copy, but Jim Meadows mentions it in his book:

https://books.google.com/books?id=M...KHY_gBCYQ6AEIQzAH#v=onepage&q=B89.3.1&f=false

Measurement of Geometric Tolerances in Manufacturing By James D. Meadows

ISBN-13: 9780824701635
ISBN-10: 0824701631
Publisher: CRC Press
Publish Date: June 1998
Page Count: 455

 

[gabimo]

Pmarc,
Quick follow-up:

You said (and please correct me if I am wrong):
The circular runout is controlling Derived Median Line if the circular runout tolerance is smaller than cylinder size tolerance. Correct?

If the circular runout tolerance is NOT smaller (and is bigger) that cylinder size tolerance than consequently is NOT controlling the DML. The quick question is: then what is controlling the DML is this later case.
Probably, you will say…sic…the DML callout. Right? If not, please correct me…. I can stand corrected….

So, the DML can be any amount? (any functional amount to be more precise)?

In this case (when a DML is specified) the surface form is left uncontrolled? I would say no. Since rule#1 is no longer applicable, then the circular runout will control the form surface and even if circular runout is bigger than the size tolerance is still controlling the DML. (regardless of the cylinder tolerance size or the circular runout tolerance)

I know……………. , it does not make any sense my explanation.

 

[greenimi]

You see pmarc….. I am not the only one who is addressing the questions only to you.

And I know why……….. and you know why too…

Because your knowledge, experience and ability to explain in the layman terms / too all level of education. I developed such of level of trust in your opinion that is like “if pmarc said so, that must be true”. No more questions or doubt.

 

[pmarc]

gabimot,
Just to make sure that we are on the same page, I indeed said that: "The circular runout is controlling Derived Median Line if the circular runout tolerance is smaller than cylinder size tolerance", but this was the answer given for a scenario assuming that no direct DML straightness callout was shown on the drawing.

If, however, the DLM straightness callout is directly specified in addition to the circular runout tolerance, then that statement is no longer always true. As long as the DML straightness tolerance value is smaller than the circular runout tolerance value, the circular runout callout will not be able to directly limit the straightness error of the DML - see greeinimi's examples B2, C2, D2, E2.

So if you are asking what controls the DML if the circular runout tolerance is bigger than the size tolerance AND no DML straightness callout has been explicitly specified, my answer is - it is the size tolerance (assuming Rule #1 is in charge).


greenimi,
I can only say thank you once again.

 

[mkcski]

To all who engaged:

I really appreciate this in-depth discussion into this little used DML concept. I have a much better appreciation of the this "gray" area.

I totally agree with greenimi in his kudos to pmarc. Pmarc is the "man". If this "group" ever gets together for a meeting (not sure if this allowed), we need to give him some sort of SPECIAL recognition for sharing his wealth of information in such a concise and clear manner

Certified Sr. GD&T Professional

 

[pylfrm]

I am not sure why exacly pylfrm asked the question about the center points, but with all of what I just said, I don't think that different methods of establishing the center points (potentially leading to different actual DMLs) could change anything in the answer to greenimi's question.

I asked to point out a shortcoming of the standard, to hopefully gain some insight (thanks for the ANSI B89.3.1 reference), and because I came up with a different answer for some of greenimi's scenarios. If the DML is determined using minimum circumscribed circles, then I think it is possible to meet a size tolerance (with envelope requirement) of 0.005 and a total runout tolerance of 0.005, but have straightness error of greater than 0.009. This requires non-circular cross sections.

pylfrm

 

[pmarc]

... and because I came up with a different answer for some of greenimi's scenarios. If the DML is determined using minimum circumscribed circles, then I think it is possible to meet a size tolerance (with envelope requirement) of 0.005 and a total runout tolerance of 0.005, but have straightness error of greater than 0.009. This requires non-circular cross sections.

I am afraid I don't understand something from the above:

1. Why did you say "(with envelope requirement)"? In greenimi's scenarios there is no envelope requirement as far as I see - it has been overriidden by the DML straightness callout.

2. Even if we assume for a moment that the envelope requirement is in charge (we would then also have to assume that there is no explicit DML straightness callout applied on a drawing), could you describe in more details how in your opinion it is possible to have straightness error greater than total size tolerance, and how it is possible to have straightness error greater than total runout tolerance?

 

[pylfrm]

pmarc,

To clarify slightly, I am interpreting "DML Redundant" to mean that the straightness tolerance c is greater than the maximum straightness error possible in the absence of the tolerance. I am assuming the envelope principle applies to the size tolerance if the straightness tolerance is removed.

I probably should have said "(with or without envelope requirement)".


For more details on the statement in my previous post, see attached image. Due to software limitations, dimensions are in units 1000 times smaller than greenimi's, and circles look rather polygonal.

The cross section (shaded in dark blue) is bounded by two arcs. The first is the size tolerance envelope boundary, has diameter 100.000, and is centered on the datum axis. The second is the minimum circumscribed circle of the cross section, has diameter 99.544, and is centered 4.772 away from the datum axis. Imagine every cross section of the feature is this shape, but the direction of decenter of the second arc varies through 180 degrees or more over the length of the feature.

I believe this feature meets a size tolerance of 5.000 and a total runout tolerance of 5.000, but has DML straightness of diameter 9.544.


pylfrm

 

[pmarc]

pylfrm,

Thank you for clarification.

Unless I am missing something in your explanation and the picture, I would disagree that the feature shown (with the cross-sectional decenter varying along the axis in a way you described) has actual DML straightness error of dia. 9.544.

In my opinion the DML straightness error is 0. It is simply because in each cross-section the center point that contributes to the overall shape of the DML should be derived from the center of the minimum circumscribed circle (which is always of diameter 100.000, and not 99.544) that does not change its location between the cross-sections.

One way, as far as I imagine, you could get the DML straightness error of 9.544 for this particular geometry would be if you started offsetting cross-sections relative to each other, so that, for example, the center point of the cross-section at one end of the feature would be offset from the center point of the cross-section at the other end by 9.544. But then this geometry would not be able to meet the total runout tolerance of 5.000.

 

[pylfrm]

pmarc,

Why do you say the minimum circumscribed circle is of diameter 100.000? Which portion of the cross section do you believe would fall outside the circle of diameter 99.544?


pylfrm

 

[pmarc]

pylfrm,

I stand corrected. I had to sketch your example in my CAD software to see that you are right. The minimum circumscribed cylinder diameter is indeed 99.544, and the DML can behave exactly in the way you described it. Thank you for showing this - really an eye-opener to me.

 

[pylfrm]

pmarc,

Glad to hear.

The main thing I take away from this, and other similar discussions on this forum, is that there is often not much value in trying to interpret the control provided by one type of tolerance in terms of another. Things quickly become more trouble than they're worth.


pylfrm

 

[pmarc]

That is true, but still these are good theoretical exercises. They definitely help to keep a sharp mind and sometimes (like in this case) expose to totally new aspects of the problem.