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

 

[SeasonLee]

The bolt circle hole position callout:

|POS|Ø0.1(M)|A|B(M)
Where A=Primary datum
B=Shaft hole
0.1=Tolerance zone diameter

Total Allowable position tolerance = Position tolerance + Bouns + Datum shift

The clearance 0.009 you mentioned is a datum shift, you need to add it the allowable position tolerance.

SeasonLee

 

[drawoh]

jefrojo,

I say no.

Compared to the clearance holes for your holes, your locating feature has a clearance of approximately zero.

The second, and more serious problem is that you have more than two holes. Ignore your locating boss. Your first screw hole locates the part in X and Y. Your second hole stops rotation. Everything else has to pass through the resulting locations. A two hole pattern can be sloppy because there is only one direction of error. The third hole makes accuracy more critical.

JHG

 

[Belanger]

Can we see the actual GD&T callout? But in general "shift" tolerance (the possible looseness between the datum feature and the item it assembles with) is NOT directly additive to the position tolerance (careful SeasonLee!), because this makes it look like each hole gets that shift. Rather, the shift tolerance only applies to the group of holes.
Sometimes on easy textbook-style examples where there is only one feature being toleranced in-line with one datum, we may speak of the shift tolerance as being additive. But the actual tolerance zone isn't getting bigger. It's the "effective" tolerance that looks bigger. At any rate, even that idea totally falls apart when discussing a pattern.

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

http://www.gdtseminars.com

 

[SeasonLee]

You are right, J-P, I know the shift can not be always added to the total allowable tolerance, there is a rotational shift for this case, Paul Jackson did gives us a good illustration on this.

http://www.eng-tips.com/viewthread.cfm?qid=199921

SeasonLee

 

[jefrojo]

Thanks for your answers

J-P, I attached another quick sketch with the actual dimensions I would like to put on the drawing.
So from your post, the datum shift is not "usable" to give me a bonus on the tolerance zone of each hole.

Season Lee, I'll take a look at your link.
Thanks

Jefrojo

 

[Belanger]

Thanks Jefrojo... so yes, there may be shift tolerance coming and its source would be the variation on datum feature B. For now we can say the maximum permissible shift tolerance is .034 (the size tolerance on B). However, a GD&T stickler would say that we also need to know the relationship between datum feature B and datum A, such as a perpendicularity callout.

But to the main point: this shift tolerance does not mean that the position tolerance available to each hole grows. (That statement is true for bonus tolerance, which comes from the size variation of each of the 8 holes.) The shift tolerance means that the entire pattern of 8 holes may "shift" as a group a little bit off center from the actual axis of B. But notice how that's different that saying that each hole may get its own extra shift.
Also, FYI, I'm not sure why there is a basic dimension of 45º. If the hole pattern needs to be "clocked" relative to the two rounded slots, then another datum is needed. Datums A and B don't help when it comes to clocking. Just an idle thought.

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

http://www.gdtseminars.com

 

[jefrojo]

Thanks J-P.

 

[SeasonLee]

J-P
A minor error noted, the datum feature B size tolerance is -0.009 and -0.025, they are all negative, so the maximum permissible shift tolerance is 0.016. Anyway, you made an excellent interpretation.

Jefrojo
I will give different tolerance on the position callout, normally the counterbore tolerance is larger than the through hole, so you may consider different tolerance callout for holes and counterbores.

SeasonLee

 

[Belanger]

Right-o! Thanks, SL

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

http://www.gdtseminars.com

 

[pmarc]

Jefrojo,
There is one very important thing you have to keep in mind. Your 0.1 tolerance on 8 holes will work (will assure proper mate with the other part) only if positional tolerance defining spacing between counterpart threaded holes is not greater than 0.1. Or in other words, the sum of both positional tolerances is not greater than 0.2.

This means that 0.08 for this part and 0.12 for threaded holes will work too (or 0.06 and 0.14), but not 0.1 and 0.12.

 

[jefrojo]

pmarc,
Yes, for the tolerance on the corresponding threaded hole, I based it on the Fixed Fastener equation, so it will be .1 .

 

[axym]

SeasonLee,

Despite what some GD&T textbooks say, the shift should never be added to the total allowable tolerance. Never. Even in the special case of a single considered feature that is coaxial to a single datum feature, the shift does not allow the tolerance zone to grow.

The figures always seem to show features whose axes are exactly parallel, with no orientation error. This oversimplification creates the illusion that the shift is equivalent to enlarging the tolerance zone. If a feature with orientation error is shown instead, it is obvious that shifting the zone is not equivalent to enlarging the zone. Adding the shift to the total allowable tolerance would accept a feature with a larger amount of orientation error, which would not fit.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[SeasonLee]

Evan

I think we can add datum shift as a total allowable tolerance if they are coaxial part, here is a good example from Alex's workbook.

SeasonLee

 

[axym]

SeasonLee,

I wasn't able to get the file from your link, but I believe I know the example that you're referring to because it's in several of Alex's books. Unfortunately, this misleading special case is the only example that the book shows. I believe that this has led many people to the incorrect conclusion that datum shift is, or at least can be, additive. The book literally says "the actual amount of datum shift for a part can be additive to the stated tolerance in the feature control frame". I maintain that this is fundamentally incorrect. Even in the single-coaxial-feature special case, the tolerance zone only shifts around - it does not get larger. I'll have to work up a diagram to show this.

In a side note the book says "datum shift is not available on features of size that are gaged simultaneously. For example, there is no datum shift between the holes of a pattern of holes". But no example or further explanation of this is given. I talked to Alex about the datum shift issue at a Y14.5 meeting years ago, and it seems that we've agreed to disagree.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[Belanger]

Evan, I know what you're saying, but for the special case of a single diameter related to a single datum axis acting in the same direction, it's OK to say that the effective tolerance zone gets larger. Recall that the word tolerance is simply "how close to perfect we must be."
So if you factor in the stated tolerance, the bonus tolerance, and the shift tolerance, in that special case the total offset between the actual diameters will be exactly one half of that total. Thus, the tolerance has been increased.

That said, it is somewhat dangerous to get to comfortable with this idea, since it only applies in that special case! So I agree with most of your post, other than that part where it's "fundamentally incorrect." For this special case it's OK: A tolerance zone that shifts around is essentially carving out a larger tolerance zone. Or maybe I should say a larger effective tolerance zone. Heck, maybe we're just splitting hairs at this point...

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

http://www.gdtseminars.com

 

[SeasonLee]

Evan

Sorry for I didn't check the link, I am trying to re-upload it again. The 2nd link is Google document in case the 1st link still failed.

I know what you want to say, and I know the difference between bonus and datum shift, the 3rd link is a snapshot from James Meadows textbook.

ETI Workbook page 221
Difference between bonus and datum shift

SeasonLee

 

[axym]

J-P,

I have a strong opinion on this one, and I'm sticking to my guns. I maintain that it's not OK to say that the effective tolerance zone gets larger, even for the special case.

I don't agree that the word "tolerance" is how close to perfect we must be. In dimensioning and tolerancing, the "tolerance" represents the difference between the allowable extremes. So for a spec of 10 +/- 1, the allowable deviation from perfection is 1 but the tolerance is 2.

I agree that in the special case the total allowable offset between the axes of the actual diameters could be no larger than half the total of stated tolerance + bonus + shift. But the next step of concluding that the tolerance has been increased is not correct, because a Position tolerance is not defined in terms of allowable offset between the considered feature's axis and the datum feature's axis. It is defined as the size of a zone, centered at true position, that the considered feature's axis must lie within. This zone allows orientation error as well as axis offset, which is what the textbook example overlooks. So shifting the zone around does not make the tolerance zone increase in size, even in the special case example. If we treat the zone as if it did increase in size due to the shift, then we would pass a feature whose axis was excessively tilted and would not fit.

A tolerance zone that shifts around is carving out a larger something (you called it an effective tolerance zone), but we must be careful to not confuse that effective tolerance zone with the tolerance zone itself. We should call it something else, to avoid this confusion - perhaps the term "spatial domain" might do. The considered feature component (in this case the axis) can span the entire tolerance zone, but it cannot span the entire spatial domain.

The geometry is different, but the underlying issue is the same as one that we disagreed over in Frank's thread about the sphere positioned to a single datum plane. I maintained that the tolerance zone is not the volume between two parallel planes, it's a spherical zone that can freely translate. Using my new term, the volume between the parallel planes would be the "spatial domain" of the sphere's center point.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[Belanger]

True -- the way the standard describes a "position tolerance" is the zone within which the feature is "permitted to vary from a true (theoretically exact) position," which is of course based on a theoretically exact datum. So the position tolerance zone doesn't get larger, because the true datum hasn't changed. IOW, datum shift does not allow the "position tolerance" of paragraph 7.2 to increase. That's what you were thinking, and that is indeed true.

But using the more generic term "tolerance" (para. 1.3.60) we can say that the datum shift contributes to the total amount by which the actual axes can be off, as you concurred, because this is a comparison back to the axis of the datum feature, not the theoretical datum.

So I'll agree on the general premise: when teaching "position tolerance" we shouldn't lump datum shift into that term. I'm just saying that if we use the definition of tolerance from 1.3.60, in a courtroom setting we could wiggle out of your strict insistence on paragraph 7.2's term (only for this special case of the simple coaxial example) and say that the maximum effective tolerance has increased -- otherwise the two actual axes wouldn't be further apart!

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

http://www.gdtseminars.com

 

[greenimi]

Evan and J-P Belanger,

A special case of a single diameter related to a single datum axis is depicted in fig 5.48 page 144(Y14.5 -1994 standard) and since position could be defined also as coaxiality of features, then the maximum allowable distance between axis of datum feature and axis of considered features is given in the table below the picture.
Therefore, at MMC for that particular example 0.5 = [0.4 pos tol at MMC + (14 MMC-13.9 LMC)+ (25 MMC-24.5 LMC)]/2 right?
Only in these simple cases, doesn't matter if you add the datum shift to the bonus and the positional toleance, you get the same results for the maximum allowable distance between axis, so the same result for the coaxiality and position.
What am I missing here?
Is it a matter of a surface verus axis interpretation?
Thank you