Circular runout control; understanding potential surface effects

[dtmbiz]

Would anyone like to comment on the following?


Would the following be accurate in understanding the potential surface effects for a cylindrical feature as a result of only a circular runout control?


1. A circular runout tolerance does not need to be less than the dimensional size limits.

True?


2. Surface elements can have “steps” along the cylindrical feature axis as a result of the feature’s actual size, which may have varying circular element diameters, along with those circular element center points being displaced within the allowed circular runout tolerance.

For example, a 1.0” dia. ±.02” external cylinder with a circular run out control of .02”, could be produced within size limits and the .02”runout control, yet still allow “surface steps” as much as .04”.

True?




3. Whereas, a total runout control of .02” for the feature would only allow maximum surface element deviation of .02” regardless of the feature’s size limits.

True?

 

[Belanger]

I don't think the premise you give in the first question is true. I would say that both runouts always control circularity, unless:
the size tolerance on the diameter is tighter
it is applied to a feature perpendicular to the datum (such as an end face)
if you try to do something goofy like apply runout to a non-circumferential circle, such as tips of gear teeth

But on a cone, I think circular runout would control circularity. The dial indicator might be on an angle, but the rotation about the datum still creates a plane that is perpendicular to the axis of rotation. (Since we're talking about a theoretical ring, it has no width.)

I have to think more about your 2nd question. But taking the words at face value, I would agree that the dial indicator has to be readjusted when it goes to a part of the cone which has a different angle. Ugh ... that might be a can of worms depending on how many rings you want to check on that cone.

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

 

[CheckerHater]

To pmarc:

I think you have to set your indicator to 45˚, or whatever the blueprint sais.
Just to be consistent with general rule - you build your fixture to basic dimensions within gage-makers tolerances.

 

[axym]

CH,

I agree that checking the Circular Runout to the same centers that the part was turned/ground on will generally be a very good approximation of Circularity.

In theory, the characteristics and tolerances should be based on the final part requirements and not influenced by how the part is made. In practice, overall costs can often be lowered by specifying a process-related requirement that is more restrictive but is easier to verify. This is one of the eternal struggles of GD&T application, and different people have different philosophies on it.

pmarc,

I don't have Y14.5-2009 in front of me, but I can briefly comment on your questions. You're opening two different cans of worms here!

1. I believe you are right. For a conical feature, the cross-sectional elements that Circular Runout would control would be different from those that Circularity would control. This difference is generally insignificant for gently tapered cones, and would be much more significant for very steeply tapered cones. To me, the meaning of Circularity becomes dubious for steeply tapered cones.

2. This is a nasty issue, and the answer (if there is one) has far-reaching effects on other tolerances that control line elements (Straightness, Circularity, Perpendicularity of line elements, etc.). It boils down to whether the indicator (and thus the tolerance zone for that element) must remain normal to the as-designed toleranced feature, or whether it can be oriented to the as-produced toleranced feature. There is still debate over this, but the general opinion I've seen is that the indicator must remain normal to the as-designed toleranced feature. So whatever the basic angle of the cone is (was it 45?), the indicator must remain normal to that cone. There are practical difficulties that result from this, and I'm sure you will have further questions and comments.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[pmarc]

Evan, CH,
Just for clarity:
The angle of a cone in fig. 9-2 is not basic - it is directly toleranced 45˚+/-2˚.

 

[Belanger]

Right, pmarc. If the intent were to hold the dial indicator at the desired angle, then we'd be required to show it as a basic angle. Thus, the can of worms...

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

 

[axym]

pmarc,

I was afraid you would say that. The issue of how to treat directly toleranced angles has not been resolved either. Its another clash between the simplistic, non-rigorous plus/minus world and the complex, rigorous GD&T world.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[pmarc]

Guys,
I had a hidden reason in raising these two issues. In my opinion the root cause of ambiguity here is an unclear Y14.5's explanation of how both circular and total runouts must be verified.

Let's take another example with total runout applied to a cylindrical surface as shown on fig 9-3. I think we all agree this is very typical case shown in almost every GD&T book. The problem (or I should say simplification) of this figure is that it presents a cylinder with perfectly straight surface line elements that are parallel to datum axis A. In such case an indicator can really slide along the surface staying perpendicular to datum axis and normal to the probed surface at the same time.

But what if the actual line elements of surface were like sine. According to my understanding of the requirement stated in last sentence of para. 9.4.2, the indicator would always have to be normal to the toleranced surface, so automatically could not always be at right angle to datum axis. It would not be a case for this particular example if the term "toleranced surface" was undestood as as-designed feature, but it is not written anywhere so is really confusing - at least to me.

And even if the term "toleranced surface" was indeed undestood as as-designed feature, what would happen if we took the example with the cone from fig. 9-2 but instead of specifying 45+/-2 we placed limit tolerance 47/43. How would one know what was as-designed angle and therefore at which angle to set the indicator for runout measurement?

My point is that it would be much more consistent if the requirement for verifying runout of cylindrical surfaces was to always place indicator normal to a datum axis not to a toleranced surface. Then it would go perfectly in line with statements like: "Where applied to surfaces, constructed around a datum axis, total runout may be used to control cumulative variations such as circularity, straightness, coaxiality, angularity, taper, and profile of a surface" (para. 9.4.2.1)

Extending this approach to planar surfaces normal to a datum axis it could be said that indicator must always be parallel to a datum axis, not perpendicular to measured surface. This again would play nicely together with statements like: "When applied to surfaces at right angles to a datum axis, total runout controls cummulative variations of perpendicularity (to detect wobble) and flatness (to detect concavity or convexity)." (para. 9.4.2.2)

I think the approach could work for conical features as well with some additions though. I would not like to talk about it in details now, because I am really interested in your opinions about runout verification in general.

BTW: Happy Thanksgiving Day to All!

 

[Belanger]

I guess I completely agree with all this. Perhaps the standard should say that the indicator is to be normal to the datum axis, thus avoiding issues of surface variation, especially a sine wave.

It would be easy to change the wording in the standard for a cylinder ("normal to the datum axis") or a flat shoulder or end face ("parallel to the datum axis"). But as you say, the problem is for cones: rather than 45 ± 2º, what if Fig. 9-2 had the angle as 44 +3/-1 ? Should the angle of the indicator be adjusted simply because someone rearranged the nominal value of the angle?

On to some turkey!

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

 

[TheTick]

Imagine a corkscrew: constant diameter, crazy runout.

 

[axym]

pmarc,

I see your point, that the explanation might be more clear in terms of the indicator's relationship to the datum axis. This is true for runout applied to cylinders and planes, because the relationship between the indicator and the datum axis is simple and constant. But when we get to cones and curved surfaces of revolution (which we must, even though these applications are much less common), the indicator's relationship to the datum axis is not so simple. The real requirement is that the indicator must be normal to the as-designed geometry. This applies equally well to cylinders, planar surfaces, cones, and curved surfaces of revolution. I prefer explanations that are based on the general case, as opposed to special cases with unique properties. This approach usually makes the complicated things easier but also makes the simple things harder ;^).

Having said that, I would agree with you that the Y14.5 explanation needs work. It needs to clarify that the "toleranced surface" is the as-designed surface and not the as-produced surface. It also needs to define what the as-designed surface of a cone is, when the angle is directly toleranced with plus/minus or limit dimensions. This is a debate in itself, because it is an attempt to impose rigor on decades-old loosely defined practices.

Traditionally, the runout tolerances have been applied to single cylindrical features and single planar features. Applying the runout tolerances to more complicated features, or groups of features, stirs up all sorts of issues that are generally glossed over in the simpler applications.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[dtmbiz]

The standard does say that the indicator is to be oriented “normal to the true geometric shape”.

I believe this should be interpreted this as the “true geometric counterpart”; which IMO takes away the confusion over the +/- size tolerance and the variations of a surface questions. I understand this to mean that the indicator is not oriented to the actual surface(s) or surface elments but rather to an axis determined by the true geometric counterpart.

After the indicator is oriented it will actually touch the surface feature for measurement.

It has been mentioned that basic dims are a better way to dimension cylindrical features when applying a runout control; regardless if the feature has been dimensioned with basic dims or +/- size tolerance (even for a cone) determining the indicator orientation based on the true geometric counter part of the produced feature eliminates that confusion for me.

 

[Belanger]

dtmbiz ... the standard says that the indicator is to be normal to "the toleranced surface" (para. 9.4.1 and 9.4.2). I don't see any mention of being normal to the "true geometric shape" (although it would be nice!).

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

 

[MechNorth]

As indicated above, runout controls are nice for simple cylindrical surfaces about an axis and for planar faces perpendicular to an axis; beyond that, profile (line and surface) controls should be used. The recurring problem is that some (too many / most?) people are not adequately trained in profile controls and therefore can't extend their uses to non-documented cases (i.e. within the standard). It is frustrating because it's like having a 6-speed transmission in a performance vehicle, but never leaving fourth gear. What's the point of having the performance vehicle?

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[dtmbiz]

John-Paul,

This is where I find "true geometric shape" in the standard(s);

"At any measuring position, each circular element of
these surfaces must be within the specified runout
tolerance (0.02 full indicator movement) when the
part is rotated 360’ about the datum axis with the
indicator fixed in a position normal to the true
geometric shape."

2009 Std fig 9-2 and 9-3 page 181
means this....

1994 std fig 6-47 fig 6-48 page 190
means this...

 

[pmarc]

dtmbiz,
Nicely noticed. I think the issue is that proper paragraphs in the text do not say that (9.4.1 & 9.4.2).

This however does not solve the problem in case of directly toleranced angles for conical features.

 

[Belanger]

Thanks, dtmbiz. I did a search in the PDF and those words never popped up because they are embedded in the graphic.

But that shows an inconsistency in the language used for runout. We're all OK with the true geometric counterpart stuff until we get to a cone, though...

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

 

[MechNorth]

...again...just use Profile and avoid those nasty plus-minuses ...

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[dtmbiz]

It would seem to me that a cone would have the indicator positioned to the angle of the cone’s true geometric shape. The indicator angle for a plus/minus 2 deg as shown in the example would be the 45 deg in the example shown. The plus/minus 2 deg only controls the size for one of the cone’s diameters.

If the cone is dimensioned with 2 diameters, then those diameters control the size and the true geometric shape driving the angle of the indicator is trig’d out using the 2 diameter’s values before applying the tolerances. (e.g. a cone with a base diameter of 1.0” -/+.02” and a top diameter of .50” -/+.02”, with height of 1.0” +.02”; use the 1.0” base dia, the .50” top dia, and the 1.0” hght to trig the indicator angle.)


How does this not follow the standard in reference to the true geometric shape?

 

[pmarc]

dtmbiz,

Let's use fig. 9-2 again.
What if the angle of cone was not 45°+/-2° but 46° +1°/-3°? This would create the same limits (function has not been changed) but the angle of the indicator would theoretically be different. So does it mean you would have to change a direction of measurements depending on the way how the angle was expressed? How could one know a direction of runout verification if the angle was expressed e.g. 43°-47°.

You were using symmetrical tolerances in your example and then one can more or less assume at what angle the indicator has to be set up. But that does not have to be always a case and that is why we are saying that for cones the procedure of runout verification is inconsistent.

 

[dtmbiz]

pmarc,

Regardless of the angular tolerance whether bilateral, unilateral, or equilateral distribution of the tolerance; looking at this as the true geometric shape, the angle is 46 deg in your latest example. The angular plus/minus tolerance is the limiting size of the cone. If I took those angles and replaced them with high / low limits expressed by equivalent diametrical values (values that are more or less equivalent to the +1°/-3°) the true geometric angle would still be 46°.

A cylinder seems to be easy for most to accept regarding the true geometric shape. We know that the angle of the cylindrical surface is parallel to the cylindrical axis. That is theoretical exact geometry.


In the case of a cone there needs to be the angled surface. It is the theoretical exact surface which includes the angle drawn or modeled. It would be the true geometric shape that the angle indicator is positioned to for runout inspection.

The true geometric counterpart references the actual mating envelope. In the case of an external feature the actual mating envelope is...

... A similar perfect feature counterpart of smallest size that can be circumscribed
about the feature so that it just contacts the surface at the highest points.

If the produced cylindrical feature has a taper as a result of tolerances, the actual mating envelope is still a “...similar perfect feature....” still a cylinder and not adjusted to be a cone. The actual mating envelope contacts the high points of the actual size cylinder.

A cone’s surface is not parallel to its axis and therefore the actual mating envelope would include the angle of the surface in the true geometric shape.

Too me, if this is not an adequate interpretation as to how to position the indicator for checking runout on a cone; then what angle would you use for a surface built at a right angle to an axis. It is implied by the standards fundamental rules as 90 deg. If the right angle surface is produced within tolerance yet has a 5 deg taper to it; then what angle is the indicator positioned to for a runout control inspection?