Simulating a datum center plane 2

[semiond]

I need help in understanding this issue regarding simulating a center plane from a datum feature related to an external width dimension: let's say a width dimention is defined as a primary datum feature. When using a vise-like physical datum feature simulator with two almost parallel faces that close on the part, unless the tangent planes on both sides of the datum feature are perfectly parallel (and in the real world they're not), one of the vise faces will act similar to a primary datum plane - touching on 3 high points, and the opposite face will touch on only one point, similary to a tetriary datum plane. Now, depending on which side of the datum feature will make the more stable contact with the simulator, we might get a different separation width between the vise faces, and therefore the simulated datum plane will also be different. For example, if the measurement set up has the vise faces oriented horizontally, the side of the datum feature facing down will orient the part in the fixture, and if you flip the part upside down for a repeated measurement you might get different results on whatever control called out that datum. Now, I understand that there is only one "actual mating envelope" to the datum feature per ASME and only one of the sides facing down will produce the "minimum separation" condition per fig. 4-13, But that means that you have to mount the part twice in the fixtute and re-check your results, and I somehow doubt that this is the recommended practice... on the other hand, if the vise is oriented vertically, we will have no control over which side we stabilize better in the simulator - which is even worse. Everyone's insight will be much appreciated... Thank you!

 

[greenimi]

All the applicable credit for this sketch shall go to pmarc.

Is 1 mm the maximum value of the parallelism that could be used/placed in FCF to make physically sense ? (Assuming y14.5 )

I cannot imagine in pmarc's configuration shown here that this value being too big due to the configuration of the shown part. Am I correct?

 

[axym]

CH,

As usual, I agree with pmarc. Controlling the parallelism of each of the 2 sides of the width, with the width as the datum feature, is not shown in the standard but there is nothing fundamentally wrong with it. greenimi's latest sketch sums it up pretty well.

If I understand it correctly, the A-B scheme you are suggesting is similar to what is shown in figures 9-4 and 9-6 of Y14.5-2009, where the runout tolerances on features C and D reference C-D. I would say that this is self-referencing to the same degree. I don't see pmarc's A scheme as any more self-referencing as this. The simulators and resulting datum would be exactly the same, wouldn't they?

I would even take it a bit further. I would like Y14.5 to allow parallelism to be specified for a group of surfaces without any datum features. I think that this would capture the real requirement on the OP's part. The two surfaces of the width need to be mutually parallel, but they don't necessarily need to be parallel to their own combined centerplane. They just need to both be parallel to the same arbitrary centerplane.

 

[greenimi]

greenimi's latest sketch sums it up pretty well.

No, no, please do not give me credit when sure I do not deserve it. Pmarc does. He is the one who created the sketch. I just use it and embedded into the thread for convenience and to serve my own agenda and my own set of questions.

 

[CheckerHater]

Thank you axym,

I am sure you will have no problem to explain me the difference between cases A and B:

To me in both cases:
1. datum is a centroid derived from feature of size
2. feature to be controlled is the part of the same feature of size
3. this is why both cases must be avoided, and why case A is not supported in the standard

When it comes to case C (did I understand your proposal correctly?) I miss it too. I agree it would be nice.
Unfortunately it's an old news.
It was present in old ISO and / or national standard(s) but was removed by "purists"

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[semiond]

CheckerHater,
What you describe in fig. B is essentialy the same thing as cylindricity (think of the inspection method).
It has a symbol of it's own hence there is no sense defining it through total runout.
Cylindricity is a form control but if you think about it total runout does make sure that the controlled feature is of cylinfrical form.
As for fig A, well ... i'll try to apply the same logic: If you added another parallelism callout on the bottom face i guess it could be considered something like "partial cuboidness" (???). Since "partial cuboidness" doesn't have a symbol in the standard, i guess it would be OK to define it through parallelism controls to the center plane.

 

[CheckerHater]

Semiond,
I like your ideas, but believe me, this is not how cylindricity is measured.
Also, partial and complete cuboidness may be measured using Profile.
Maybe profile is the solution?

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[semiond]

CheckerHater,
How is cylindricity measured? From theory, i thought it was like that. Never seen it myself though.
Regarding surface profile, i thought about it, but then the width dimension must be basic, and in that case, I'm not completely sure if it is cosiderd a FOS (for center plane derivation). There is a discussion regarding this on another thread here and opinions are split.

 

[greenimi]

Let me make shame of myself and trying to explain the differences (at least one of them) from CH’s sketch between cases A and B:

Case A: the parallelism of the top feature does not control the location of the surface to the datum plane (centerplane) . I do not see this as self-referencing).

Case B: total runout does control the axial location (coaxiality aspect of the location) of the cylinder’s surface to its datum axis. I do see this scheme B as self-referencing.

 

[CheckerHater]

Semiond,
For the beginning, cylindricity has no axis. No datum.
Then, I copy-pasted question from your post "How is cylindricity measured?" and... Google is your friend:

https://www.google.com/search?q=How...&biw=1504&bih=866&safe=active&ssui=on#imgrc=_

For more details, ISO 6318, ASME B89.3.1

greenimi,
Runout does not control size just like parallelism does not control size.
If you insist that "total runout does control the axial location of the cylinder’s surface to its datum axis" in any form (I don't know what "coaxiality aspect of the location" is), you indeed make shame of yourself. (Sorry, I didn't mean to be rude)

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[greenimi]

No worries. I come here to learn. I do not get offended very easy.

Probably, in my post I had poor choice of words. What I meant is that in you case B, total runout will not control the surface location in relationship to the datum axis, but will control the axis location and that’s why I am seeing as self-referencing.
Therefore, one of the location’s aspects in case B is controlled, versus in case A where no location’s aspect is/are controlled.
Don't worry if I am wrong I will stand corrected. No problem.

As far as “Runout does not control size just like parallelism does not control size". I agree.

 

[semiond]

CheckerHater,
Look at the google pictures you linked to -This is exactly what i meant. The cylinder is rotated around it's own axis and a probe perpendicular to it is checking the variation along the length of the surface.
This is almost-exactly like total runout. The only difference - instead of checking the feature in reference to another feature's axis, it is checked in reference to it's own axis. And, this is -exactly what you described in fig. B.

"Cylindricity has no axis" - that depends on the context. If "no axis" means you must not put a datum in the FCF - that's correct. But if the context is the inspection process, it does involve rotating a part around a rotation axis. Yes it's the axis of the same feature that is controlled, so what? Saying "this is not how cylindricity measured" is totally misleading.

 

[CheckerHater]

No problem greenimi,
One should be particularly careful around Runout
Because it's compound control it may control roundness, eccentricity, parallelism, taper, BUT ALL AT THE SAME TIME!
So it can be used and abused in many different ways. I personally avoid declaring "runout controls this or that". It does (sorta), but when every thing else is really small.

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[3DDave]

Just to clarify - there's already some problem with requiring the inspector to reposition this part to find a condition that meets the FCF requirement and a hostility to the idea that in some of the available positions the FCF requirement will not be met.

The proposed solution is to add more requirements that are based on the same feature and that inspecting the new requirements is difficult because establishing the basis datum precludes simple access to the dependent features and does nothing to eliminate the original problem with repositioning the part or that this will add more feature requirements that will also not be met in alternative orientations.

This has gone pretty far in losing sight that controlling tolerance zones is of benefit when such control duplicates realizable parts and that standardizing those controls is beneficial when it's a labor reduction for commonly encountered good practices.

 

[CheckerHater]

semiond,
The turntable of cylindricity / roundness measuring macine does NOT rotate around the same axis all the time. That's exactly the point.
Runout is measured by rotating around one fixed axis. Your feature may have perfect cylindricity, but runout of several millimeters.
So, no, this is not what I described in figure B

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[pmarc]

CH,

However strange it may sound to you (and to others), and just for the purpose of this academic discussion, I am saying that your case B is not self-referencing. Why? It is because datum axis A is derived from UAME (Unrelated Actual Mating Envelope) of the cylinder, and the total runout itself controls condition of the surface of the datum feature.

In other words, if the cylinder is manufactured with any kind of form error, for example as a saddle that has the valley in the middle of the cylinder length, the total runout reading will show a quantifiable error. My definition of datum self-referencing is that self-referencing is when inspection reading of a geometric characteristic defined relative to a datum is always 0, no matter what, just like in the case I described yesterday - width chosen as datum feature C and then its center plane controlled with parallelism tolerance to the C. The measurement result will always have to be 0.

With that said, I am not suggesting that this kind of callout should ever be used. It is simply because I believe, like semiond, that it does nothing else that cylindricity would not do. And in fact if you go to the paragraphs in Y14.5 about cylindricity and circularity you will notice that word "axis" shows up there.

One more thing... could you give a reference to a paragraph or figure in any old ISO/national standard that would show that the practice you showed in figure C was ever allowed?

 

[greenimi]

CH,
Let me ask you: do you see differences between your posted cases A and B? If yes, what would those be?
I am sure Evan (and maybe pmarc too) will have their own input about your cases A and B.

 

[3DDave]

There was no requirement to constrain the parallelism; it's being added with no particular benefit. You already rejected a simpler method to attain the functional goal that I proposed in favor of a semantic discussion diversion.

If parallelism is to be controlled - pick one surface as a datum reference and apply the tolerance to the other - no need to sum the results and very easy to inspect. I'm sure many people have told you how this works.

And now you deleted the post. Why?

 

[greenimi]

could you give a reference to a paragraph or figure in any old ISO/national standard that would show that the practice you showed in figure C was ever allowed?

I think (in ISO) was alllowed something similar with CH sketch case C, but one of the arrows must have a succsion cap (datum feature symbol). I do not have a good memory anymore, but I remember was something along those lines- datum feature symbol to be shown. Am I correct, pmarc?

 

[semiond]

I deleted the post because i no longer think it's a good one. Two parallel planes will show bad parallelism between each other if you sum their parallelism results to a datum that is not parallel to either of them. On the other hand if they are parallel they will not likely to produce a bad datum - but it's too messy.

 

[greenimi]

My friend "google" said DIN ISO 1101-1985 shows parallelism within some tolerance with a succsion cap on one end and an arrow on the other pointing to two different surfaces (paragraph 8.3, fig 37)
Not exactly as CH picture, but ..... I do not say "close enough"