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

 

[dingy2]

Circular and total runout reference the datum diameter in RFS according to ASME Y14.5M-94. One cannot use MMC here.

Sorry but no attibute gauge here.

If the runout was converted to positional in a MMC mode, then one could reference the datum diameter in MMC and now it would conducive to an attribute gauge.

Dave D.

http://www.qmsi.ca

 

[powerhound]

If you are trying to control the conical surface, then circular runout at RFS is in order. As Dave said, you can't do what you're trying to do with a runout specification. If you simply must use an MMC modifier on datum B then you should specify profile of a surface on the conical surface and then reference datums A and B.

I don't think position is the answer here but maybe it is. Can you shed some light on your thought process Dave?

Powerhound, GDTP T-0419
Production Manager
Inventor 2009
Mastercam X3
Smartcam 11.1
SSG, U.S. Army
Taji, Iraq OIF II

 

[swertel]

We have actually considered chaging it to a positional tolerance, because then it would control circularity as well as straightness. (Don't read too much into those in terms of the definition of them via Y14.5, I can't think of another word that adequately describes the concept.) But the problem with changing the tolerance at this point is that it is a completely different philosophy of inspection for these parts, and the change to the tolerance stack-up for assembly will also have to be ammended, flowing into all the tooling and gaging for those components as well.

Profile on a surface may work very well. I'll have to ponder that a bit and consider how to inspect it. We usually use CMM or Optical Comparitor for profile inspection. Both of those aren't possible. We need hard gaging.

Thanks for the clarification on the runout callout. It is the same as I interpreted. Just hoping I was wrong.

--Scott

http://wertel.eng.pro

 

[dingy2]

Powerhound:

I am rethinking my suggestion to go to a positional tolerance.

In a conical situation, we have a height dimension to consider and this is never a factor in a cylindrical pin or hole. If the cone was in relatively true position but the height was too high, possibly the checking fixture would not descend all the way down. It may appear that the cone is out of position and, possibly, it may not be the case.

Positional is not a good replacement here. Unfortunately, I would stick with circular runout and then measure.

Dave D.

http://www.qmsi.ca

 

[MechNorth]

Use profile of a surface where the limits are the equivalent inner boundary and outer boundary from your original combination of size & runout. Inspection is done exactly the same as if it was a runout control, but now you zero your indicator at the basic radius, and watch that your indictor doesn't exceed +/- half of the profile tolerance. One setup and you're done for size, position and runout ... not bad, eh?

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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[swertel]

That is fantastic. You deserve a start for that. Now if only management realized how much money you just saved them!

--Scott

http://wertel.eng.pro

 

[MechNorth]

Tks. Do I get 15% commission too?

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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[axym]

Jim,

I'm confused. The considered feature is conical, not cylindrical - did you not notice that?

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[MechNorth]

Tks Evan, I had missed it. It's essentially the same answer though. The only geometric control that locates a surface is profile of a surface. By establishing the datum axis and rotating the workpiece about the datum axis, you still have an absolute location where any point on the surface is supposed to be, and a tolerance zone normal to the conical surface in which the surface must be. Vision-type systems are great for verifying cones.

While it is legal to use +/- tols to size a cone, I'd like to see how anyone measures accurately at opposing points. Without referencing the datum, there is no way to ensure that the opposed points are directly opposed (i.e. a circular cross section instead of an elliptical one). Again, a surface profile provides the size & runout controls in one setup. You zero the indicator (offset from the datum axis at the basic radius at one end) and now instead of moving the indicator longitudinaly along the axis, you move the indicator longitudinaly at half the basic included angle of the cone.

It's been a rainy, chilly day here, so I decided to make some pretty graphics too. Note: Marvin is from SolidWorks 3DContentCentral.

http://www.profileservices.ca/files/tidbits/thd1103_239358_1.pdf

http://www.profileservices.ca/files/tidbits/thd1103_239358_2.pdf

http://www.profileservices.ca/files/tidbits/thd1103_239358_3.pdf

In these graphics, Marvin will move the indicator normal to the basic (nominal) surface, as represented by the blue arrow. The hazy brown cone represents the nominal (basic) geometry, and the green workpiece surface is the actual workpiece.

Please, no comments on Marvin's poor inspection methodologies ... he's new at it.



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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[ringster]

Jim,

Are these graphics intended to relate to the Original Post By Swertel?

If they are I cannot quite make the association. It would seem that the conical surface should be positioned vertically and rotated about the base rather than the axis.
It is not readilly apparent as to what is the primary and secondary datum feature in the setup. Could it be somehow clarified?

 

[dingy2]

Jim:

I would just love to watch someone try profile of a surface on a cone using a circular runout setup. The cone must be shown in basic dimensions along with either the angle or the taper end. Where would one place the indicator initially? How would one sweep the surface on a basic angle?

Circular runout is still best in the situation described.

Dave D.

http://www.qmsi.ca

 

[axym]

Ringster,

You're right, Marvin's setup deviates from what was described in the original post. The part appears to be chucked at both ends, but should be oriented to one end face first and then centered using one chuck. I don't think it needs to be positioned vertically though, are you thinking that the part might sag due to gravity and affect the reading? This shouldn't be a big issue on Mars.

It also appears that Marvin isn't holding the indicator perfectly normal to the surface and will get some cosine error. In fact, in the first diagram it's apparent that the stylus isn't even touching the part at all! And what's with the second indicator? Is he going to be inspecting concentricity next? Also notice that both indicators show exactly the same reading. An odd coincidence.

The whole thing is fishy if you ask me. It almost seems made up.

p.s. I'd like to hear Marvin say "Oh goody. My Mitutoyo Q-36 high-resolution dial indicator"


Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[dingy2]

Evan:

You are correct about Marvins measuring method. I do like Jim's animated drawing although it isn't really correct.

Dave D.

http://www.qmsi.ca

 

[MechNorth]

No, the graphic doesn't reflect the datum scenario given in the OP. The larger issue, I thought, was the question of using profile to control a cone. As I have some existing models that I use for training & documentation, I decided to use them to illustrate how to verify a conical surface spec'd with a profile control.

The drawing would, indeed, show basic dimensions for the end diameters, the length (unless a FOS, in which case +/- tol could be used), and the included angle.

Verification methodology depends on what you have at hand.
1) runout stand inclined at half the basic included angle by using a sine bar or angle plate; this puts the basic profile to the horizontal (or vertical), and the indicator can then run in a straight line wrt the surface plate.

2) runout stand as shown in graphic, and indicator on an inclined guide rail (picture two vertical stand posts, and two cross rails which remain parallel to each other while skewed to the vertical posts; I haven't seen this one in use, but I have seen the setup). I have also seen a single cross-rail, but don't like it as the indicator can pivot about the cross rail axis, thereby losing its angularity relationship wrt the basic profile geometry.

3) manual optical comparator, with the datum axis aligned horizontally or vertically; you offset the object from the center by the basic radial distance at either end, rotate the crosshairs & grid on the screen to the basic angle, then you can travers +X/-X or +Y/-Y until you get the limit of the surface at that section, or you can use an overlay with the nominal, min & max boundaries and see if the conical surface violates that boundary as you rotate it about the axis

4) an automated profile vision system rotates the workpiece about its datum axis at high rpm, and a light curtain & camera traverse longitudinally along the axis to establish the surface which is then compared electronically to the CAD geometries directly or to a manually entered basic geometry and offset boundaries.

Options #1 & 2 use basic inspection apparatus that should be available in pretty much any shop. This reinforces how easy it can be to verify a surface profile with basic equipment.
Option #3 requires an optical comparator, but that's pretty common in most shops that I've seen. Again, no real challenge with this methodology.
Option #4 requires a comparatively inexpensive piece of equipment which, in a production environment, saves far more on labor costs than its purchase price.
A CMM in scan or point mode can be substituted for an indicator if preferred.

Conical tapers are the bread & butter of the tooling and molding industries. These and other methods are used every day for conical tapers.
As for profile of a surface vs size & circular R/O;
- Profile checks size, form, orientation & location in one setup and inspection process
- size is to be checked by Taylor's Principle (Full form at MMC, 2 pt check at LMC)
- how do you know you got the actual circular cross-section rather than a skewed cross section?
- how will you check form at each point?
- circular runout is a check at each cross-section perpendicular to the datum axis
- what happens between these sections? waviness along the length of the part (as my modeled workpiece) will not be detected by circular runout
- what if you want a large size tolerance, but need good form along the length of the workpiece? Size & Circular Runout can't do this; you'd have to add a straightness control also, but that would only control single longitudinal elements, not the net effect. Would you then add a circularity control as well? Of course not because Form controls are nasty to verify. The alternative is a composite surface profile, and the inspection apparatus would typically be a runout setup.


Now, Marvin is a difficult creature to work with, and he's still tentative about his job, so he's just moving into position. As in reality, there will always be a bit of a cosine error as he can never truly get the deflection probe to a perfectly normal orientation to the basic surface.

Marvin's second indicator is part of his Q2009 Inspection Modulator Probe, complete with aiming reticle & hair-trigger download of digital data. Marvin would not, however, ever use his capabilities for the evil task of verifying concentricity. ;~}

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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[swertel]

Nobody likes to verify concentricity. I don't even think Duck Dodgers could pull off the task; or his Eager Young Space Cadet.

I'm actually home with a sick child today. I'll see when I get into the office tomorrow if I can dumb-down my original print to something non-proprietary and share a pdf of it.

--Scott

http://wertel.eng.pro

 

[axym]

For Duck Dodgers, just pronouncing the word "concentricity" would be enough of a challenge.

Evan better, imagine Elmer Fudd talking about the three different orientation tolerances.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[dingy2]

Jim:

I reviewed your methods and found each one, if they can be performed, much more difficult than circular runout.

1 Usually a divider head or chuck is horizontally mounted on a table and placing a sine plate (bar) under it is not possible. Maybe if one had a small portable chuck, but I have not seen one in my experience.

2 Using an inclined rail is possible if one used a metal lathe but they are not used as measuring instruments. If they were, each would have to be calibrated using the 10% tolerance rule.

3 This is possible but we are only measuring down the incline or taper rather than around the circumference. Should one rotate the part and perform this again? This could be really messy.

One could use a computer controlled CMM and that is probably the way I would suggest if a profile of a tolerance is really required but is it. The questioner here wanted to use an attribute gauge rather than confirm the feature with a circular runout.

Dave D.

http://www.qmsi.ca

 

[MechNorth]

Dave, in each of those cases you'd be doing essentially the same as a circular runout, except that you are also verifying the size at the same time.

Your perspective seems fixed on large automotive parts. Small & moderate-sized parts are often measured in the ways I have mentioned (there are likely other setups that I've not seen). Small & medium sized indexing fixture components are readily commercially available, and can use collets, mandrels or chucks. If the workpiece is large enough, hydraulically lifting one of the ends of the chucking fixture can easily achieve the inclination needed.

Guidance rails are commonly set up parallel to the surface plate, with the indicator moved along its length. It's not a trivial setup perhaps, but it is not uncommon. Granted, the error should (not must, should) be less than 10%, as with most setups.

As for #3, you seem to be looking at it strictly as a surface straightness indication. If you leave the indicator in a single longitudinal position as you rotate the part (as you would for a circular runout), you are then verifying the surface with respect to the basic profile; as long as the indicator has been zeroed before starting, the deviation at any position along the surface must be within the surface profile zone. You take readings progressively along length of the feature as the workpiece is being rotated.

By using any of the setups that I've described, in conjunction with a surface profile control, you effectively have an attribute go/no-go gage.

The term CMM encompasses contact and non-contact systems. The majority of time I see them used is in automated mode rather than manually operated; the typical exception is one-offs or first-offs, where a program hasn't been established yet. There are vision systems marketed for doing profiles of a rotated body, and I have used them to verify profile controls on a conical surface of great precision.



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

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[dingy2]

Jim:

Here is what I see.

The originator of this question wanted to make a functional gauge to check the runout on a tapered surface and wanted to know if the datum could use the modifier MMC. Well, it can't. Both circular and total runout are always in RFS.

You suggested a profile of a surface and is much more restrictive and could have a cost impact on the product. Is it easier to confirm than runout?? No, it isn't. Can it be done? Yes, it can.

Even in the new ASME Y14.5-2009 it states "it is even more important that the design more precisely state the functional requirements". Do we really need profile of a surface? Just imagine the originator of this thread suggesting we place profile of a surface on this taper and it has a cost impact of even $0.50 a part. One would need some strong background and I doubt there is one.

If I were the person making a decision on this design change, I would ask the person suggesting the change to tell me what the function of the feature and why circular runout cannot meet its functional needs. Why do we need profile of a surface? I hope all Designers think like this but I don't think it is happening out there.



Dave D.

http://www.qmsi.ca