MMC on Datums for Runout 1

[swertel]

I am in need of a clarification to a runout callout.

Datum -A- is a flat index surface. (The mating part threads up against this surface.)
Datum -B- is a stepped shoulder diameter slightly larger than the external thread major diameter.

I have a runout on a conical surface that is
|RO|.002|A|B|

What I'm wondering is if I can use MMC for Datum -B- in the runout feature. By doing so, I can design my gage to only the max diameter of the shoulder (Datum -B- ) and not require a collet in order to find the diameter RFS. My theory in doing so is that I end up with more of a functional gage because the mating part indexes against -A- anyway. Tolerance stack-up between the shoulder diameter (-B- ) and the mating part may result in a loose fit, so only the pitch diameter of the threads actually contacts. And thus, the runout of my assembly is really going to be based on the flatness of Datum -A- and the mating surface so gaging to something similar is a reasonable functional check.

--Scott

http://wertel.eng.pro

 

[MechNorth]

Dave,
You are correct on the first point; then again, I never said any runout could be at MMC (control or datum), so I'm not sure why you're raising it again.

It does, however, confound me when people automatically assume that profile of a surface will cost more to verify than other controls. Profile of a surface establishes a boundary within which the feature must be contained. Is that not the same function as an attribute gage? I understand that profile is not yet as widely understood as some of the other controls, however its versatility and general ease of verification are indisputable to those who understand and use it. I use surface profile extensively when all that I am worried about is a boundary rather than separate and distinct size and location or orientation controls. In every case, once the fabricator understood what a surface profile was, and how to verify it, the costs of the part dropped and the inspection was facilitated. As a side bonus, if you hook up your indicator for electronic data acquisition, you would also get data for SPC.

I am curious how you would verify the size of the cone? At one end? Both ends? With a +/- size tolerance, you only get size controls at locations where the size is specifically detailed; nothing in between. What about the form (conicity for lack of a real term)? Rule #1 doesn't apply here because it's not a feature of size, so you can't use Taylor's Principle. What is controlling the barrelling / waisting along the length? Circular runout will control the straightness of the axis, but not tell you anything about the surface deviations along the conical feature.

Before the OP uses any GD&T suggestions from any source, they need to understand the design intent and the real impact of each type of control. If the runout needs to be tighter, then put it as a refinement of the profile control.

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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[dingy2]

Jim:

If one converted a circular runout tolerance to a profile of a surface with the same value, the requirements are more stringent and the costs increase. Tighter tolerances cost more $$. This has nothing to do with the difficulty in confirming the profile tolerances or whether the part is automotive or not.

If the cone size/angle tolerance and the circular runout tolerance were combined, then I could see that there would not be a cost increase but I don't think that happens in real life.

The function and relationship of the cone determines whether a profile of a surface is more appropriate than circular runout.

Dave D.

http://www.qmsi.ca

 

[MechNorth]

Dave, please re-read my original post ... "Use profile of a surface where the limits are the equivalent inner boundary and outer boundary from your original combination of size & runout." I was pretty clear that you don't just substitute one control for the other without doing some math. Some designers may not understand how to do the math in some cases, but again, I can assure you that such conversions are done, and I've seen them done properly enough times to know that industry is doing it.

Your assertion that size & circular runout are adequate still doesn't address the questions I posed to you about the size controls between the cross sections with toleranced sizes, and the form effects along the length. These aspects need to be controlled.

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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[axym]

Here's an alternative strategy that avoids the problems of circular runout and surface profile:

Specify Symmetry to control the orientation and location of the cone. The datum features would be the two theoretical planes that are associated with the datum axis. This would provide control in two perpendicular directions. Because the conical surface is tilted, the opposed pair midpoints would have to be calculated normal to the basic angle of the cone. I propose that this "radial symmetry of a cone" characteristic be named Cone-centricity. Then refine the form and angular size of the cone by applying an Angularity tolerance to each linear surface element, with respect to the opposite surface element.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[MechNorth]

Sorry, Evan. I misunderstood your intent, so I've red-flagged my previous post to remove it.

My concerns are;
1) Symmetry and concentricity require the averaging of directly opposed points to find the derived median plane or derived median line, which is then checked against the location wrt the datum axis or datum plane.
2) Both of these controls (concentricity and symmetry) and their datums apply RFS (5.12.1, 5.14); MMC and LMC aren't permitted as the controls are based on the actual surface rather than the MMC virtual surface, which would give you zero positional error upon inspection.
3) Symmetry applies to nominally planar surfaces and concentricity applies to nominally circular elements.
- How will you determine which pair of planes to use?
- How will you establish the planes?
- A tangency modifier on the tolerance would establish a plane at max'm limit of the surface, however you would have an infinite number of opposed tangent planes around the circumference, so it would be time consuming to verify.
4) After establishing these line-element control zones for location, you propose to "refine the form and angular size of the cone by applying an Angularity tolerance to each linear surface element, with respect to the opposite surface element.
- Angularity within Y14.5 is a linear zone established at a basic angle from the datum, but if I understand this correctly, you are now defining each opposite line element as the datum, similar to putting "individually" on a datum feature callout.
- Looking back at #3, you'd have an infinite number of such opposed line element datums.
- As your datum has now changed, you'd need a new inspection setup at each rotational location.
- Form doesn't have any datum reference, so you'd have another setup for form, which appears to be the same as straightness along the surface if I understand you correctly. If that's the case, then you have to float the workpiece until you get the minimum straightness reading along the cone's length at each radial location.
5) I'm not sure how this new control would simplify the situation.

I'd still stick with surface profile, maybe even a composite surface profile if refinements of orientation & form are needed.

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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[axym]

Jim,

Ok, I'd better put a stop to this right now. I don't think that either one of is sure whether or not the other is kidding anymore. Other readers are probably confused as well.

My cone-centricity/angularity strategy was intended completely as a joke. I had tried to make it so ridiculous that it wouldn't be taken seriously. But I suppose that it didn't really sound much crazier than some of my other ideas that were intended as serious.

If you were trying to be nice and provide an earnest critique without inviscerating me, then I'm sorry. I guess I'm not too good at conveying sarcasm and feigned indignation in a text form, and the intent wasn't clear. If your comments were in jest, and the joke's on me, then I deserve it and that's fine.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[MechNorth]

Nah, we're good. I figured you were joking, but with your response I thought ... oh, he's actually serious?!

As an instructor, I get to hear lots of "but that doesn't work for us, so why don't you nice standards people just do it this way instead, and have this mean that instead ... but only in this case 'cause we like the rest of the standard pretty much." I had a student comment recently that it wasn't nice when I laughed openly at someone's suggestion for a universal customization to suit them. It wasn't really a laugh, it was more of a chuckle because everyone thinks their issues are unique to their company and products. I guess I should'a let them in on the joke too. C'est la vie! Live & learn.

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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com