linear dimensions on geometricly toleranced print 3

[mack52]

If a print has a section view with a linear dimension
describing a feature, does the view need to be set up for measurement purposes according to the geometric datum scheme or can I just measure it as a independent feature
because of the linear dimension callout?
This print has a combination of both linear dimensions and geometric callouts on the print.

 

[KENAT]

axym, the more I think about this issue the more I feel like you that perhaps the more important thing when deciding which dimensions to heavily sample is process capability.

Of course, for this to make sense one has to assume that the person creating the drawing paid attention to the tolerances and that effectively they've done it in a way where they are all important.

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[Belanger]

Tolerance values should be as large as possible while still allowing the part/assembly to work. Thus, a smaller tolerance value should be driven by the function. I've seen crazy things like 0.5 mm on the contour of a foam cushion -- gaaaa

http://www!!

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

 

[drawoh]

Tolerance values should be as large as possible while still allowing the part/assembly to work. Thus, a smaller tolerance value should be driven by the function. ...

This discussion about critical dimensions bothers me. To me, as I prepare drawings, you are absolutely correct.

The fabricator is supposed to meet all the tolerances on the drawing. The sloppy dimensions can be just as critical as the tight ones.

I used to work with a guy who left notes on his drawings telling fabricators to pay particular attention to certain dimensions on the drawing. The obvious implication is that the other dimensions and tolerances do not matter, and there is no need to meet them. How does our inspector reject shoddy work?

I have run up against sort of an interesting drafting problem. I was dimensioning some holes tapped in a machined plate, for cable ties. I actually do not care where cable ties are located. A positional tolerance of Ø5mm would be adequate for me. On the other hand, there is no way for a machinist to take advange of this. There is the significant possibility that someone will look at the the holes, and use them for a fabricated bracket, thus imposing a requirement for accurate tolerances, not met by my drawing. I called up a tolerance of Ø0.2mm.

JHG

 

[KENAT]

drawoh, I tend to agree with you on the 'critical dimensions' issue.

The problem in your last paragraph is also familiar. I tend to pick a fairly loose figure well within anticipated process capability, even if functionally it could be looser.

This 'future proofs' it so some extent, for later modifications.

It also shouldn't increase cost as it's 'well within' process capability.

Now the gray area is that if they happened to slip up, and exceed your already generous tolerance, you probably wouldn't reject the part, which generally one might say suggests the drawing should change.

I've got the process selection wrong before too. I'd set it at something like +-.030", which for most machining processes I anticipated to be ample. Well they had a capacity issue and decided they wanted to rough the outline dimensions on the bandsaw (or maybe it was a round part and they flame cut it - my memory fades). For this the relatively loose tolerance wasn't loose enough so we had to do a drawing change.

I'm a bit more generous with such tolerances now.

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[dingy2]

drawoh:

All dimensions must meet the criteria on the drawing but ones that have a function or mating relationship should be monitored (controlled) more often. No manufacturing company will confirm each dimension on each product and that is a fact. How one dimensions the part provides the information needed to control functionally important dimensions.

Your example of only requiring a diametrical tolerance zone of 5 mm but placing a tolerance of 0.2 mm is a typical example of the kind of drawings floating the shop floor today. If you had placed a MMC with the positional tolerance, then manufacturing may end up considering it a functionally important feature and have a checking fixture made for the location of the holes. What a waste of time and $$. If you left the extremely tight positional tolerances in RFS, then I know it would only be checked using variable measuring methods on a limited basis.

Has the application of positional tolerances on these holes assisted in making the drawing clearer or has it added to the confusion?

Dave D.

http://www.qmsi.ca

 

[ewh]

Yes, some dimensions require monitoring more than others, but I disagree that the drawing defining the part is the best place to note which is which. QA often makes their own documents to control part quality, such as ballooned inspection drawings, based on the original part drawing. There is nothing to prevent them from doing the same defining what dimensions get monitored how often.
This comes back to the purpose of the original drawing; to completely define a part without ambiguity. Part definition, not manufacturing process definition. Adding notations of any kind that do not aid in part definition are superfluous and should be documented elsewhere.

"Good to know you got shoes to wear when you find the floor." - [small]Robert Hunter[/small]

 

[dingy2]

ewh:

You are absolutely correct about QA developing a Control Plan. We would get together with the Customer QA and ask such questions as "How does the part fit and function?" since the drawing did not always reflect the function and mating relationship of the part.

I would never ask about adding notations but would suggest that design intent including applicable GD&T is reflected on the drawing.

Dave D.

http://www.qmsi.ca

 

[drawoh]

dingy2,

We subcontract fabrication, but we inspect in-house. I can tell our inspector all sorts of things I would never tell the machine shop. My assumption is that Ø0.2mm positional tolerances are easily achieved. The cost of unneeded accuracy in my situation, is tiny.

In your environment, I would make the tolerance sloppy.

The holes are tapped, so MMC does not help me.

JHG

 

[axym]

We got a bit sidetracked with the discussion of inspection frequency, which is a very important but separate issue. We didn't really answer the "plus/minus versus GD&T" questions from the discussion of Figure 3-29.

Dave, you said that the linear tolerances in Figure 3-29 are comparable to profile of a surface tolerances and posed the question of why one would pick one method over the other. Caseynick replied that profile of a surface more clearly defines the drawing. I agree, but what are the details?

Surface profile tolerances would each designate one feature as the datum feature and create a well-defined tolerance zone for the considered feature. This tolerance zone would control the form, orientation and location of the considered feature but the datum feature would not be controlled in any way. The meaning of surface profile is well defined in the Y14.5 standard.

If linear tolerances were specified, as they are in Fig 3-29, exactly what do we get in terms of control of the features? This is where it gets interesting. These features are not well-behaved features of size, because the surfaces are not fully opposed:

The 35 36 and 25.4 25.5 dimensions are between surfaces that directly oppose each other in some areas but not everywhere.

The 41.1 41.3 dimension is between surfaces that point in opposing directions but do not directly oppose each other anywhere.

The 9.4 9.6 dimensions are between surfaces that point in the same direction and are therefore not opposed anywhere.

The 6.1 6.6 dimension is between a surface and a cone-cylinder intersection line.

So exactly what do these linear tolerances define in terms of control of the features? Can any of these dimensions be considered "size dimensions" to which the Rule #1 perfect-form boundary would apply? Can the actual local size requirement be applied to any of them? If the answer to these questions is no, then what does apply?

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[dingy2]

Evan:

I would set up a plane on the bottom surface which is datum A and then sweep any plane that has a linear dimension from that surface just as one would using profile of a surface. This concept is shown in fig. 2-5.

Dimension 9.4-9.6 is shown from the other end so I would create a plane on the smaller end and the check around the shoulder in various areas confirming that they do, indeed, are not outside the 9.4-9.6 criteria.

That fact that these dimensions are not profiles indicate that they may not have anything to do with the part's function and relationship and would be confirmed infrequently. All dimensions shown in the feature control frames must have an influence on its function and mating relationship and would be confirmed on a regular basis.

Dave D.

http://www.qmsi.ca

 

[MechNorth]

Dave,
Once again, if you're going to keep posting your recommendation to use a defacto datum, please justify it based on a standard. You keep challenging others to prove their case, and we either provide the proof or retract, but you have yet to prove your case on this matter. This is a recurring theme and scenario. It concerns me because there are comparatively inexperienced people using this forum to up their game based on the Y14.5 standard, not a personal opinion. As you are a GDTP, your voice carries extra credibility but pulling in your own opinions, uncorrelated to the standard.

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[dingy2]

Hi Jim:

Each time I respond, I reference a figure or figures as I have in the last post but I do not align myself with the philosophy of "every dimension must be covered by GD&T". I did mention figures 2-5 and 3-29. As a matter of fact, you mentioned that the following "Fig 3-29, yup, I think this is a good mix for this part's functionality".

In the foreword of the standard it states "it is even more important that the design more precisely state the functional requirements". I am attempting to follow that concept.

I was asked how I would measure that linear dimensions shown in figure 3-29 and I answered based on both the standard and my years measuring parts using a CMM.

So, Jim, how would you measure that linear dimensions shown in figure 3-29? The 9.4-9.6 dimension would be most interesting.




Dave D.

http://www.qmsi.ca

 

[MechNorth]

The closest measurement to the intended point-to-point would be using a set of calipers across the top and bottom. Not truly a point-to-point, but the closest you'll get that I've come across. Measuring from a defacto datum is not indicated in the standard. To advocate that method is not valid to the standard.

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[axym]

Dave,

Jim is right, your last post contained several statements that represent your personal opinion. The connection you make between geometric tolerances, function and mating relationship, and inspection frequency does not have any basis in the standards.

I still challenge your use of Figure 2-5 to justify choosing a de facto datum for linear tolerances. The Dimension Origin tool shown in Fig 2-5, with the circle at one end of the dimension, is not interchangeable with baseline dimensioning. We don't see the Dimension Origin circle symbol on the linear tolerances in Fig 3-29, so the high-point-plane origin described in 2.6.1 does not apply to them.

So what does that leave us with?

There is a note on Page 27 of Y14.5-2009, right underneath Fig 2-5. The final sentence in the note is:

"Locating features using directly toleranced dimensions is not recommended."

Why would that statement be there? I think it's because directly toleranced dimensions like the ones in Fig 3-29 don't provide a clear definition of what the control over the features is. There is specification uncertainty. There are no rules in the standards to clarify what the considered feature is, what the tolerance zone is, or what the origin of measurement is. The inspector must invent something based on other knowledge of the part, "tribal knowledge", assumptions, or guesswork, as Dave and countless others have over the years.

So how are non-rigorous specifications like these still workable in industry? Any thoughts?

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[dingy2]

Jim & Evan:

Figure 2-5 reflects the surface using the origin symbol where the 12 +/- 0.4 dimension. The bottom line of 2.6.1 states "Without such indication, the longer surface could have been selected as the origin, thus permitting a greater angular variation between surfaces." Please note that the feature is not a feature of size in any stretch but a distance between 2 planes.

Using that concept, I wonder which surface I should take my readings from on fig. 3-29? Obviously, I would take them from datum A. I really don't know where this point to point measuring concept is derived but I see a surface (origin plane) to a surface (plane).

Evan:

You should have stated that complete note where the statement "Locating features using directly toleranced dimensions is not recommended." was taken out of context.

Bottom on page 26 it states:

Note: When basic dimensions are used, the is no accumulation of tolerances. A geometrical tolerances is required to create the tolerance zone. In this case, the style of dimension (chain, baseline, direct) is up to the user. Locating features using directly toleranced dimensions is not recommended.

The full statement was discussing geometrical tolerances using basic dimension in which certainly makes sense. Certainly, one would not use directly tolerances dimensions in this situation.

8.2.3 Profile Tolerance as General Requirements

This covers your default profile tolerances but 8.2.1 states "A profile tolerance may be applied to an entire part, multiple features, individual surfaces or to individual profiles taken at various cross section through part." So, there is an option to place a general profile tolerance but it is not mandatory. If you use a general profile tolerance, is the datum structure valid for all the features??

I have read (was implied) where if one does not place a default positional and profile tolerances, your company may not be modern (with it). It appeared to be blasphemy that I would actually question this concept since I am a GDTP and have been training in this subject since 1988. Possibly, I might be just old fashion or out of step but I am one of the few that offer live online GD&T training. I don't think that is old fashion.

Does your drawing truly reflect the design intent? Are the part's function & mating relationship clearly shown using GD&T in FCFs or are they all part of the default tolerances?

This, thankfully, is my last input on this subject so beat away guys.

Dave D.

http://www.qmsi.ca

 

[MechNorth]

Dave, again, you avoid answering the question directly. Where do you find your position of using defacto datums in any standard?

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[dingy2]

Gentlemen:

I am now going to eat dirt.

In the forward on page vi, I have noted "design more precisely state the function requirements." but there is more that I should have stated.

It states below the above statement "To accomplish this it is becoming increasingly important that the use of geometric and dimensioning (GD&T) replace the former limit dimensioning for form, orientation, location, and profile of part features. This revision contains paragraphs that give a stronger admonition that in the past that the fully defined drawing should be dimensioned using GD&T with limit dimensioning reserved primarily for the size dimensions for features of size."

This the concept that my fellow colleagues were arguing but could not find anything in the standard to support their claims but it is here.

For that, I will give both Jim and Evan a star.

Dave D.

http://www.qmsi.ca

 

[MechNorth]

Tks Dave. Very gracious, and appreciated.
Jim

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[axym]

Dave,

That is very gracious of you, and you don't need to eat dirt. I'm not sure that Jim and I deserve a star though - obviously neither of us stated our case well enough to convince you.

Y14.5-2009 does admonish the use of limit dimensioning for purposes other than size dimensions, but that use is still illustrated in the standard. The infamous Figure 3-29 uses limit dimensioning for the location of part features.

That is why we need to examine for ourselves what the consequences are when plus/minus or limit dimensions are used on features that are not features of size. It's case-specific. There are consequences, such as the drawing not being "fully defined" in terms of bulletproof definition of what is controlled. In some cases we can live with these consequences and in other cases we can't.

That's why plus/minus tolerances are used extensively and are not going to go away. The pros outweigh the cons in many situations. Opinions?

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca