Radius as a datum 3

[ImnotfromMars]

Hello,

Can a radius be a datum?

 

[Belanger]

Pmarc -- if an irregularly contoured surface is identified as a datum feature, I would still say that the true datum is a plane.

See Figure 4-28 and notice that the datum is a plane, although offset from the part by the 44.4 dim. (Or is this a case where you would say that the standard doesn't use the word "datum" in its explanation of that graphic?)

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

http://www.gdtseminars.com

 

[fsincox]

I still go back to, if it functions that way, I do not care if it is easy to find, it is. Then the limitation is ours, self made, and the standard needs to evolve to handle it.
Frank

 

[pmarc]

OK. I am convinced.

 

[axym]

Hi All,

Interesting discussion. Very interesting.

I agree that examples such as this expose the limitations in Y14.5's current datum reference frame theory. The plane/line/point datum concept makes things certain things easier to understand for very simple cases, but breaks down very quickly for more complex cases.

Y14.5-2009 provides a table in Figure 4-3 showing simulators, datums, and DOF constraint for various types of datum feature geometry. But these only apply to primary datum features. There is no explanation of how datums work for lower precedence datum features, where certain DOF's have already been constrained. Hence the confusion over Figures 4-29 and 4-30, in which a secondary datum axis or datum plane constrains only one rotational degree of freedom. How does this work exactly? We're left to wonder how an axis that is capable of constraining 4 DOF's only constrains one.

I think that pmarc is on the right track by questioning things, and exploring what the datum would be (if any) if the surface was irregular. I would say that if an irregular surface was referenced as a secondary datum feature, the datum would be a plane/line/point combination as defined in Figure 4-3. But exactly how that datum constrains only certain degrees of freedom, I'm not sure. The only way I've been able to make sense of it is to forget about the datums, and focus solely on the datum feature simulators. I agree with pmarc that when proposing this type of idea, one should prepare for attack ;^). It clashes with the core Y14.5 principle that a datum reference frame is a three-plane coordinate system derived from datums.

The way I like to look at it is that the reference frame is the "space" that the datum feature simulators are defined in. The coordinate system can be placed anywhere in this reference frame - it is arbitrary. The datums are arbitrary as well - they are only a conceptual convenience. This is why there are so many examples in Y14.5 in which the datums are difficult to figure out, or seem arbitrary. Figure 4-26 with the datum axis derived from a hole pattern is one of the worst. Don't get me started on that one!



Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[pmarc]

Thanks for support, Evan. I couldn't describe it better.

 

[fsincox]

Evan,
The hole pattern seems to be a very functional type of datum. What are we trying to do, define functional requirements, or make things easy to check? We can't always have both. I hear it all the time, the management always says they want both: quality and schedule, it is pretty obvious to me which drives and which suffers.
Frank

 

[fsincox]

I am interested in your POV because to me 4-26 looks like the ultimate functional requirement (without knowing the actual requirement, but projecting one), why would we not want that?
Is it not what we have been striving for?

 

[fsincox]

When everything is made by Stereolythography (sp)2 (or some such process) and machining is a rare occurrence will anyone care much where the edges are anymore?

 

[CheckerHater]

The hole pattern seems to be a very functional type of datum.

Frank,
I was always wondering how car industry was doing it for last 100 years.
To make sure you have good wheel, you have to check runout of the rim in relation to pattern of 4 or 5 conical holes used to bolt the wheel to the car.
Has anyone ever seen a blueprint?

 

[fsincox]

CH, I wish I had copied some of those auto company prints from back then (when I worked in an industry that supplied them machine tools).
Tapered cones is something always debated here, but has long been the foundation of the machine cutting tool industry. It was always an acceptable datum to me, I did not need ASME's OK to see that.

 

[axym]

Frank,

Let me explain further.

I don't have an issue with a hole pattern being specified as a datum *feature*. I agree that many parts mate this way and therefore referencing the holes together makes sense and captures the functional requirement. No problem there.

My problem is with the *datum* that Figure 4-26 shows as being extracted from the pattern of holes. The datum is described as an axis, which does not make sense.

Going by the feature types in Figure 4-3, the pattern of 4 parallel cylindrical holes falls into the category of Linear Extruded Shape (f). Thus the datum should be an axis and center plane. The axis alone would only control the X and Y degrees of freedom and would be insufficient to control w (clocking). I can't begin to count the number of questions I have heard over the years, on the meaning of the "datum axis" in Figure 4-26 (and the similar figure in Y14.5M-1994).

Even if we use the correct datum type in Figure 4-26, I still question the usefulness of the datum. Section 4.12.3 describes the pattern of datum feature simulators based on virtual condition boundaries, which is great. Then it states that the origin of the datum reference frame, which is based on the datums, is established either at the center of the pattern or at any other location defined with basic dimensions. In other words, the datum feature simulators are the important entities and the datums and DRF are just reference entities tacked on for convenience.

The wheel with 4 or 5 conical holes fitting simultaneously is a different can of worms.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[CheckerHater]

The wheel with 4 or 5 conical holes fitting simultaneously is a different can of worms.

This is what I find the most disturbing.
People were making wheels and bolting them to the cars/carts before GD&T was invented.
And yet in 21st century part that is produced literally by millions for several decades is still considered nuisance, can of worms, etc. Isn’t that a shame that we can figure the way how to make and inspect part, we just cannot figure out how to draw in on the piece of paper?
Is it just me thinking that today’s GD&T is suffering from severe disconnect from reality?
This is definitely subject for separate thread, but could it be the reason for today’s slow rate of GD&T adoption?
Sorry, this isn’t Friday yet; I must be getting grumpy ahead of time.

 

[pmarc]

And what is really the problem with GD&T in case of a car wheel?

 

[axym]

The difficulty that I see with the wheel is that the 5 conical holes center on the 5 fasteners simultaneously. In order for this to happen, something has to give (deform or bend) and act as a non-rigid part. This goes beyond the default Y14.5 assumption of perfect rigidity. So it can be difficult to establish a DRF on the 5-hole pattern in a way that will duplicate the final assembly condition, without actually mounting it physically. Even then, the final DRF will depend on the exact geometry of the mating part or gage. There is some unpredictability there, and some sort of averaging or approximation is often used.

So it's not that we shouldn't apply GD&T here, it's just a challenging application for DRF establishment.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[pmarc]

I am probably missing something, but why would one want to establish DRF from a pattern of 4 or 5 or 6 holes and not from center hole in the rim? As far as I understand this is the centering feature.

 

[CheckerHater]

Not on all wheels, especially steel ones.

 

[pmarc]

http://tires.about.com/od/understanding_wheels/a/hub-centric-vs-lug-centric-wheels.htm

 

[CheckerHater]

So the article said that some wheels center on hub and some not.

How is it different from my statement? And how does it help with GD&T issues?

I suspect I will never be able to convince you this way. Apparently you and I were changing tires on different cars, so we have different feelings about it. There are several designs out there. Have you ever seen wheels that are first roughly located by pins mounted to the hub and then bolted down with tapered screws?

Here is another automotive application: truck half-axle being mounted and centered using tapered holes/fasteners. Feel free to deny.

 

[pmarc]

All I was trying to say is that for hub-centric wheels I see no major problems with proper GD&T application.
And I was not challenging your statement in any way.

 

[CheckerHater]

Sorry if my tone was a bit harsh. I am not trying to start a war, just to convince that there is legitimate case. Also, in my understanding FUNCTIONALLY it is conical holes/fasteners that do all the heavy work, holding the weight of the entire vehicle. Just a thought.