Pattern hole position 4

[SeasonLee]

I thought a pattern hole’s measurement can be started from any one of the hole, and the position calculation result will be the same no matter which hole was chosen as origin, but the sample shown with a different result, what did I missed on the hole location measurement and position calculation?

Season

 

[Belanger]

I think it's because your measurements are all based on a falsely-introduced orienting of the plate to some external edge datums. The print only shows a reference to datum A, so whichever hole you call the zero, the next hole to be measured is just a straight-line connection, not both X and Y.

In the second scenario where hole #3 is the origin, there is no way all of the other three holes would have deviation in both the X and Y directions, because you can only build the X-Y system as you progress around the 4 holes (actually just the first two to be measured would be enough to square up the X-Y system). There is no perpendicular and parallel until after the second hole is measured for location.

For instance, if I start with #3 and then measure down to #4 as my next hole, the Actual Location would be 0 in the X direction and -21.0238 in the Y direction. (This comes from the -1 and -21, but using the Pythagorean Theorem to make a straight line between those two centers.) Thus, there is no deviation in two directions because it's just a straight shot center-to-center.

I haven't figured out all the numbers to really see if it all works out, but that is where I believe the issue is.

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

 

[SeasonLee]

J-P, Thanks for your comments, you did caught my points, the new post have fully defined the DRF, if the position calculation still with a different result, what is the possible causes?

Season

 

[pmarc]

SeasonLee,
In the second attachment, why are you still using one hole as an origin? The origin of measurements is in the bottom left corner of the part.

Additionally, I would say that the drawing (in the second attachment) is incorrect. There is no basic distance in both directions between pattern of holes and datums B and C.

 

[Belanger]

Yes, if you introduce B and C on the outside edges, then basic dims from those edges would be needed, and then the data given in the table would not start from any hole as the 0,0 point.

So the only reason you could start with any one of the four holes as the zero point is if A is the only datum. And if that is the case, then my previous post would explain why the data seems to be different.

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

 

[pmarc]

Actually, even with A as the only datum, I would say that one should not start with any one of the four holes. They are equally important and they all must fall within their positional tolerance zones simultaneously.

Making one of them as an origin of measurements may lead to false insepction results in certain situations -- for example, pattern that does not meet the positional requirement may be assessed positively.

 

[Belanger]

But pmarc, think of a CMM -- where would it establish a "zero" origin? The answer is anywhere.

Take a simpler analogy: If I want to measure the length of a board, I can hook a tape measure on the left side and measure across. But I could also hook my tape measure on the right side and measure the other way. It doesn't matter, because we're only looking at a local length; there is no false result. And the same applies here: if we are only looking at the local position of the holes, you can "hook your tape measure" onto any one of them. Then you would measure to the next hole. The 3rd and 4th holes would then follow from there, eventually forming the four cylindrical zones (there is still tolerance around the first hole, and that first zone should remain floatable so that its tolerance doesn't get bunched up elsewhere).


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

 

[pmarc]

J-P,
Have a look at attached picture. This is what I meant in this case. For simplicity, I drew only 2 holes and considered their relationship in horizontal direction. I think it can be noticed that pattern in the upper picture (that satisfies drawing requirement) will be assessed as out-of-spec when one of the holes serves as an origin of measurements (the bottom picture).

http://files.engineering.com/getfil...e-35a6-4add-8e70-d90725503bb3&file=origin.JPG

 

[Belanger]

But that second picture would not be correct, because it fails to include the tolerance on the first hole. Notice how my post stated that the first hole would still have to have a "floatable" tolerance zone. Or in your case, since there's only two holes that first hole's tolerance would be bunched up into the second hole's tolerance.

So if it's inspected correctly, there is no issue.

If that still doesn't make sense, then just think about my first question in the previous post: where in your second picture is the origin?

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

 

[pmarc]

But that second picture would not be correct, because it fails to include the tolerance on the first hole.

Of course, the second picture is not correct exactly because of the reason you mentioned. But this is what happens when one of the hole in the pattern is selected as the origin. And this is how most of CMM operators I have met try to deal with such type of callouts -- in order to find the origin one of the holes is arbitrarily picked and the rest is measured relative to the selected one.

That is wrong, but this is exactly what seems to happen in SeasonLee's pictures.

 

[Belanger]

Yes, we agree then. My point was that if you insist that the first hole is not an origin, then you imply that something outside of the 4-hole pattern is the origin. And that is introducing a false datum (i.e., the edges of the plate).

SeasonLee, I know that might not be the answer to the initial question you were looking for, but I still think the situation you ask about is based on that "false datum" issue.

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

 

[SeasonLee]

pmarc
Your comments also caught my points, would you mind to guide us on how to measure the pattern holes by CMM correctly ( in case A is the only datum ).

J-P
The 2nd post (with DRF fully defined) more or less revealed the designer intent, i.e. use the edges as datum features, and I believe the CMM operator took the edges as datum too.

Season

 

[PaulJackson]

Most CMM software packages have pattern fitting routines where rotation and translation "if permitted by the specification" are iterated to solve for the size of the tolerance zones that contain the feature's attribute (aurface, axis, median plane, etc.) deviation. The presentation attached shows the typical open layout progressions chosen to fit rotation of a simple pattern.
Paul

 

[pmarc]

SeasonLee,
As shown by Paul, this would have to be an iterative, best-fit, CMM software analysis.

Paul,
I have a question. Page 4 of your document. If the features shown are shafts, as stated on page 2, could you clarify how the values of "Allowable Individual Feature Position Tolerance" for Features #1, #2 and #4 are calculated?

 

[PaulJackson]

could you clarify how the values of "Allowable Individual Feature Position Tolerance" for Features #1, #2 and #4 are calculated?

PMARC,
It is the difference between the individual feature’s MMC size and its pattern constrained oriented actual mating size as the collective pattern is simultaneously expanded/contracted among the pattern features until expansion is limited by at least one feature, then sequentially expanded/contracted among the remaining features until all remaining pattern rotations have been eliminated, then verified that no individual feature local sizes violate their LMC limits.
Or
It is some abbreviation of that process that makes… sometimes intelligent… sometimes ignorant assumptions about each individual feature’s form and orientation then computes the difference in measured size from the MMC size.
Paul

 

[SeasonLee]

Paul / pmarc
Thanks for the post, If we are using an old CMM software and the package don’t have that feature, then how to start to measure the pattern holes?

Season

 

[pmarc]

Thanks, Paul.
What I don't get on page 4 is the numbers.

Take Feature #3:
If the size of its actual mating envelope is 9.04, you have 9.4-9.04 = 0.36 of bonus.
0.36+0.36 gives 0.72, not 0.52.

Features #2 and #4 look OK, but there seems to be the same issue with Feature #1.

What am I missing here?

 

[PaulJackson]

Yes you are correct the sizes of #1 and #3 are switched around.
Sorry I never caught that mistake thank-you.
Paul

 

[pmarc]

No problem, Paul. I find your presentation valuable because it nicely emphasizes importance of the problem and shows how to deal with such kind of situations. I wish more and more CMM operators were aware of the issue you described.

I would say the numbers are less important aspect in this case - I just wanted to make sure that I got everything right.

Besides, I believe there is no need to provoke other picky bastards like me

 

[SeasonLee]

Well, let's back to my question, how will you start to measure the pattern holes(in case A is the only datum)?

Season