Calculation of position tolerance with reference to hole 1

[jefrojo]

Hi all,
When calculating the positional tolerance of a pattern of holes that are on a bolt circle and are located by a shaft going into a hole, do I get any bonus/minus because of the value of the adjustment between the hole and the shaft at MMC?
(See attached)
With the fixed Fastener equation, I would get a positional tolerance of dia0.1 at MMC for the clearance holes (In reference to my DIA50g6)
Because I have a clearance of .009 at MMC between the hole and shaft, do I need to remove it from the positional tolerance?

Thanks

 

[Belanger]

You've got the right idea, and that figure in the standard is a good way to look at it -- for that special case.
Evan and I were just having fun debating a fine point about the terminology and how we as trainers present the subject.

The 0.5 that you correctly derived will be the maximum permissible offset between the two actual axes. You're safe in just calling that "offset" or "deviation"; calling that 0.5 (or the 1.0 diametrical equivalent of 1.0) the "total tolerance" is what raised that side discussion.
But it must again be emphasized that if you go to other examples where the datum shift is not like this simple case of a single feature, acting in the same direction, then the concept of showing the numbers as in Fig. 5-48 is not even possible (e.g., Figs. 4-6 or 5-2 from that same edition of Y14.5).

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

http://www.gdtseminars.com

 

[pmarc]

Fig. 5.48 is actually a good example to show why datum shift should not be understood as addtional tolerance for positional requirement.

1. Imagine what would happen if datum feature A was referenced at RFS/RMB and not at MMC/MMB: If both axes were perfectly parallel to each other, the maximum allowable axes offset would be 0.45, equal to half of total positional tolerance available (0.9), right? But this would also mean that in the other extreme case, when axis of bigger cylinder is extremely tilted, the linear vertical distance between its beginning and its end could be no more than 0.9.

2. Now, if datum shift was indeed to be interpreted as additional positional tolerance, for datum feature A referenced at MMC/MMB the maximum total positional tolerance of axis of bigger cylinder would be 1.0. And that is not correct since the tilt of the axis can still be 0.9 maximum. Datum shift does not allow it to be bigger.

 

[pmarc]

The last sentence should have been:
"Datum shift will never allow it to be bigger".

 

[Belanger]

Understood, pmarc. (And Evan, I know you were also alluding to the orientation stuff.) But I still say that it's not a huge problem to say that the maximum effective tolerance between the cylinders could be as much as 1.0 for Fig. 5-48.
Notice that I'm just summarizing the possible effective tolerance; it's not a blanket statement about what the position tolerance is always and everywhere!

As an analogy: Consider how Rule #1 controls flatness across the top of a rectangular block. And suppose the height tolerance is ±0.5. You guys wouldn't have a problem with me saying that the maximum possible flatness tolerance could be 1.0. Now if that top surface were tilted to the extreme permitted by the size, then we would say the flatness tolerance is now 0.
Simply paste that analogy onto this coaxial example! If you start tilting things, certainly that eats into the other variation that might be possible.
I think it comes down to this: you saying that "the position tolerance would be..." (which I agree is not a correct blanket statement)
versus my use of "the position tolerance could be..." (which can be fine if understood along the same lines as the Rule #1 example given).

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

http://www.gdtseminars.com

 

[pmarc]

J-P,
I am having a problem with grasping the analogy. In your flatness example there is no "external" factor (like datum shift in case of coaxiality) that could (or could not) influence the value of acceptable flatness error. It is obvious that flatness actual error can be 1.0 max and nothing can change that.

Additionally for flatness example the tolerance zone can tilt itself since it is not tied to any datum, while for coaxiality the tolerance zone is fixed in rotation to datum axis A... I am really having difficulties to see your point...

 

[Belanger]

The flatness error is a variable depending on an external factor (the size). But I goofed in my statement about the flatness being reduced; of course a tilted surface could still be bumpy. My bad! And a bad one even if I were a GD&T beginner <blush>

But what I meant is that when an external factor drives a tolerance, any other variation thrown into it may reduce what's really available. But it's still permitted to say that there is such a thing as a "maximum possible" value.




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

http://www.gdtseminars.com

 

[SeasonLee]

Evan and J-P

I can find out at least two more examples from Alex's training material that datum shift can be added as a total allowable tolerance, and they are not cylindrical coaxial part---the special case as we mentioned earlier.

SeasonLee

 

[Belanger]

Right. As long as we have caveats in mind for other not-so-easy examples, the terminology used there is not out of line.

I'll try another one of my goofy analogies: Suppose my car is at rest in the garage, and you ask me how much space it takes up. I'll give a certain answer (such as "x" sq. ft.).

No suppose I'm driving my car in tight circles and you ask me how much space it takes up. My answer will be different, and it will be a larger number.

So with shift tolerance, we realize that the position tolerance itself doesn't grow (my car didn't grow), but having the position tolerance move around effectively contributes to the tolerance, if that word is properly understood.

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

http://www.gdtseminars.com