Spinoff2: Using orientation to control position 1

[fsincox]

The whole idea of using orientation to control location has also come up in my past work as the “GD&T guy”. A very well respected senior inspector proposed the use of parallelism to control the location of shaft keyways. His idea was that you would line up the shaft diameter (the primary datum) and then roll the keyway, itself (the secondary datum) to center and proceed to measure the resulting parallelism of the opposed sides of the keyway. His rational was that the resulting parallelism of the sides of the keyway in this representation of the installed state was important to the key not working its way out under heavy loading. I have to admit I did not feel it was technically illegal, what do you think?

 

[CheckerHater]

Frank,
You are right!
The purpose of my example with 2-hole block was exactly that: To show that Orientation alone doesn't have enough power to compete with Position. Datum [A] is good enough for Perpendicularity, but to locate with precision you need complete framework; and I am afraid only in ISO Perpendicularity pins down position and only by mistake.
And bad news: there is no definition of "location" in Y14.5-2009

 

[fsincox]

CH,
Thanks, I suspected that, so you never know.
IMHO, The standard is wrong on concentricity, symmetry, runout, the adoption of ISO datums without conforming to ISO rules, common zone vs. continuous feature, anything else you want to add?
As you point out we all seem to agree the ISO was wrong in your cited instance. Of course the standards have the convenience of just defining themselves right. I was going to bring this discussion up in the discussion on the ISO issue, but I knew I needed my own thread.
FYI, I laid out the rough example pmarc had sketched earlier. I did not know if he had actually literally calculated those (2) bad part examples he showed. I get about a .060 position error translates to about .010 surface parallelism (we are missing a dimension for the depth of the keyway).
Frank

 

[fsincox]

I have thought about it some more last night and had an epiphany of sorts. I am sorry, I cannot give up right to call this fact of containment a form of: “location” in a non-beginner crowd. Orientation does not control “absolute location” that is the kind or location you are all thinking of and concerned with. It does absolutely control the “relative location” of the other points of the feature so defined that is the kind of control that this relies on.
Frank

 

[axym]

Frank,

CH is right, there is no specific definition of "location" in Y14.5 We must pull its meaning from the context(s) it is used in. Here is the breakdown that I would use:

1. Location of a feature relative to datum(s)
-we get this by applying any of the Tolerances of Location (Position, Concentricity, Symmetry) to a feature or features
-I would say that we also get this with both types of Profile, Circular Runout, and Total Runout
-location is controlled because the tolerance zone is not allowed to translate relative to the datum(s)
-the relationship between the considered feature and the datum(s) must be basic described with explicit or implicit basic dimensions/angles)
-I usually refer to this as "absolute location"

2. Location of a feature relative to other features
-we get this by applying the Position or Profile characteristic to a pattern or group of features
-location is controlled because the tolerance zones are not allowed to translate relative to each other
-the relationship between the features must be basic (described with explicit or implicit basic dimensions/angles)
-I usually refer to this as "relative location"

The use of the terms "absolute location" and "relative location" here is my own, and others may suggest alternative terms. Also, directly toleranced dimensions can also locate features but I have not dealt with this case.

Frank, you're extending the idea of relative location to apply to the location of points on a feature relative to other points on a feature. I would say that this is a valid concept, but the Y14.5 standard does not go there. Using the term "relative location" to apply to within-feature relationships would confuse most GD&T users.

Can we agree to acknowledge that containment is technically a type of within-feature relative location, but reserve the term "location" to apply to features as described above?

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[fsincox]

Evan,
I hope you know I want to, but, don't all points have to be contained in a tighter orientation or form zone under ASME rules as you fundamentally understand it? I really am misunderstanding ASME rule fundamentally if that is case! How can we all say that an orientation tolerance will control form if it is not true? Have we stumbled on the ASME's version of independency?
Thanks for getting me here!
Frank

 

[Belanger]

Frank,
Orientation tolerances are applied to the axis/centerplane of the "actual mating envelope," so that's why we can say they don't necessarily control form. Think of a bent cylinder.

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

http://www.gdtseminars.com

 

[fsincox]

JP,
Thanks, Thats right we now hace AME's! I was thinking and discussing surfaces here, all points on a surface are still forced to be contained in a zone, are they not?
Frank

 

[Belanger]

Yes, on a surface the form would be controlled. Sorry if I threw in a curveball about the AME on a feature of size; I haven't been following the thread closely and wasn't sure if we were talking about surfaces or features of size.

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

http://www.gdtseminars.com

 

[fsincox]

JP,
Thanks, That was good for me, because, I need to flesh out this whole difference in an AME and an actual FOS better. I am getting to the point where I would like to understand the differences in imparted control.
Concentricity still controls the actual center of a FOS, while the control imparted by the AME is tough to really conceptualize, is it not, or do you find it easy?
Frank

 

[axym]

Frank,

Yes, all points need to be contained in an orientation or form zone. So when applied to a planar surface, an orientation zone will control form. I don't think anybody's going to argue that.

We (well, at least I) just don't want to apply the word "location" to the containment of points within any kind of tolerance zone. The concept of containment is a fundamental paradigm in GD&T - every geometric characteristic requires that something be contained in a certain tolerance zone. That something might be the entire surface, a derived point/line/plane, a surface element, or whatever. The tolerance zone can have various types of constraints applied to it. But there is always a tolerance zone, and something that must be contained within it.

It seems to me that the underlying, if unstated, concept of location in Y14.5 has to do with controlling translation of the tolerance zone:
-All geometric characteristics that control location constrain the translation of the tolerance zone in some way.
-All geometric characteristics that do not control location allow the tolerance zone to freely translate.

How does that sound?

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[Belanger]

"All geometric characteristics that control location constrain the translation of the tolerance zone in some way."
Boy, Evan -- why didn't you mention that in my other thread? I used a dozen different ways to claim as much, but people still want to allow weasel room for position to control only orientation.

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

http://www.gdtseminars.com

 

[axym]

J-P,

Sorry, I only came up with that description this morning as a result of pondering Frank's "containment" thing. But the idea had been floating around somewhere in my head for a while - contained but not located, if you will. ;^)

Don't worry, I'm going to apply it to your other thread. I think that the translation concept strengthens my case, that Position controls location even when applied to a single perpendicular hole. But I might have to pull out some pretty abstract stuff to argue it. Stay tuned.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[fsincox]

As I mentioned what is important, is the result, not how I want to describe it. The establishment by framework definition places the keyway on center. The containment provided by the parallelism controls the sides surfaces to zones that are parallel to the shaft axis yet still allowed to float inside the greater size tolerance zone.
I mentioned above that a .010 zone controls position roughly .060 total position zone (.030 off center) in this case.
I am certainly curious if you believe this callout, pmarc kindly provided for me, is a logical extension of principles and legal and I never told you what I was doing with it.
Frank

 

[axym]

Frank,

I see it now, you and your inspector are correct. It took 4 or 5 days to get over my preconceived opinion, but I finally see it. I followed your logic about the framework and I agree with it now.

I humbly retract my earlier statements about pmarc's diagram. The two imperfect as-produced part geometries shown would not meet the print. The actual Parallelism error seems to be about 2 or 3, far exceeding the 0.05 tolerances.

It does appear that the Parallelism tolerances somehow indirectly limit the amount Position error that is possible. I don't know if this qualifies as location control, but it sure mimics it pretty well. This brings into question what type of geometric error pmarc's diagrams represent - is it location error, orientation error, or somehow both?

Thanks for sticking to your guns on this and not giving up on us. I think that this is going to lead to an improved understanding, at least for me.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[fsincox]

Evan,
I thank YOU for taking time to stick with me! You guys are about my only resource for this kind of stuff and I defiantly do not wish to ruin it. People come to me for answers on this stuff and if it is not in the book, I have to “make the call”. The drawings we make are out there for all to see, we are not like doctors we do not bury our mistakes (just an old cliche)
This was 20 years ago, now we might use the angularity symbol (under 2009) or maybe composite profile, profile was outlawed by a management dictate at the time.
Frank

 

[DeanD3W]

Frank,
I owe you an apology, and if we ever meet in person, a dinner, lunch, or a beer... I sent an email to Evan directly today and his reply with "think of the simulator for datum feature B" hit me over the head firmly, so my feeble brain could finally see your point. You are absolutely correct and I was quite wrong to mindlessly cite the common "orientation tolerance do not locate a feature"... Except when they do, as you have very patiently continued to point out.

The datum feature simulator for B must be perfectly centered on a radial line from datum A (or, if it is a physical datum feature simulator, centered within tooling tolerance to that radial line)... Using the 2009 terminology of "Regardless of Material Boundary" (RMB) the simulator for B will grow to fill the slot and as it does, any location error of the slot will cause the part to rotate about datum A until the simulator for B cannot grow any more... This rotation needed to accommodate the location error of the slot may cause either or both of the sides of the slot to tilt out of their respective parallelism tolerance zone. Because of this effect the parallelism tolerances on the side of the slot have indirectly controlled the slot's location relative to the datum reference frame [A,B]. I know what I've just stated is no more than what you've been saying, but I like to explain it using slightly different terminology.

So, I get the "closed mind of the month" award on this one. Thank you very much for politely and patiently persisting with your point. It is a very good one.

Dean

http://www.d3w-engineering.com

 

[Belanger]

I haven't really been following this thread, but based on the last two posts I said, "I gotta see what this is all about!"

Frank put it well when he said that location is controlled "through the back door." I can see what he means, using pmarc's sketch: first we grab onto the OD to hold the shaft, but then we rotate it until the one side of the slot meets the datum B simulator. That's the key: only one point of the slot's wall may hit the simulator, so now the rest of the slot will look angled, thus failing parallelism.

So having datum B on there does mean that we'll eventually discover if the slot's location is off center. (I'm just saying all this to explain it to myself, really.)

Jim S. had a similar analogy with profile a while back, and it took time for me to fully buy it. I claimed that a profile tolerance really shouldn't reference a datum that is part of the surface being profiled. But the light bulb went on, and there may be a time when such a "self-datum" actually changes the meaning slightly.

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

http://www.gdtseminars.com

 

[pmarc]

I promised to myself that I will not take part of this thread any longer, but now, when Evan, Dean and J-P have collectively agreed with Frank, I feel like have to jump back in and understand where did this sudden change come from.

I must admit, I have some difficulties in visualizing some of your statements, like:
Evan's: "The two imperfect as-produced part geometries shown would not meet the print. The actual Parallelism error seems to be about 2 or 3, far exceeding the 0.05 tolerances."

Dean's: "The datum feature simulator for B must be perfectly centered on a radial line from datum A (or, if it is a physical datum feature simulator, centered within tooling tolerance to that radial line)..."

J-P's: "That's the key: only one point of the slot's wall may hit the simulator, so now the rest of the slot will look angled, thus failing parallelism."

It seems that the most crucial for understanding the case is to visualize how - based on my sketch - datum feature simulators A & B (or actually their mutual relationship) look like. Can someone of you enlighten me, please? A sketch would be perfect.

 

[Belanger]

Well, attached is a VERY crude sketch, but it is how I see the situation.

First, in your mind rotate the entire picture so that the gage plate representing datum B is at 12 o'clock. Notice that this gage plate must be in line with datum axis A. Now, since our actual slot was off center, the shaft is rotated clockwise until we make contact with the gage plate. My hand scrawled note will explain it from there.

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

http://www.gdtseminars.com

 

[axym]

pmarc,

JP's sketch was a good start, here's one that goes a bit further. I haven't shown the datum feature simulator for A - there's just a small dot to represent the datum axis. But the sketch should illustrate the statements that were difficult to visualize.

The most important detail here is that datum feature simulator B is directly in line with datum axis A. This isn't a trivial detail and is worthy of discussion. I'll start this in another post.

Evan Janeshewski

Axymetrix Quality Engineering Inc.
www.axymetrix.ca