Basic Dimensions Documented in First Articles 3

[DCarnley]

thread1103-215826

I have recently had an internal discussion at my company over whether or not to include Basic dimensions in an AS9102 First Article. The answer is yes you do. Pure and simple. To do otherwise is not complying with the AS9201 Standard. Basic Dimensions are part of the design characteristics on the drawing.

I read in one post "BASIC dims have no tolerances by themselves. They are nominals. Measuring their location directly violates the design intent and provides no useful information within the context of their design intent. A direct measurement of a BASIC dim does nothing to say if the feature is in tolerance. If one feels it is necessary to measure directly, they need to invoke some other method of tolerancing instead of BASIC."

I'm not sure why that person stated that, but measuring them directly is what you need to do to find Position. True that are nominals but so are that non-Basic dimensions. If I have a part that has a Hole and uses Cartesian dimensions showing the Hole Diameter, the X location and Y Location with ± tolerances everyone agrees that we record the Hole Diameter, the X and Y dimensions on the FAI Plan and measure directly from their respective surfaces on the FAI report. Fine we all agree on that principle. Why does anyone think it's different with geometric tolerancing. , Document the Hole Diameter with a ± tolerance, Basic Dimensions for X & Y direction, and Position tolerance Feature Control Frame. Yes the X & Y BASIC dimensions are theoretical however they ARE directly measured from the DATUM features called out in the Feature Control Frame. Those actual dimensions are recorded and compared to the basic dimensions to find the delta from each X & Y Basic Dimension (True Position). The outcome after the position calculation is called Position (Actual Position from the theoretical True Position in a diametrical circle/cylinder). So the Hole is measured DIRECTLY and related back to the DATUM Schema in a Cartesian manner (can be Polar too). The Basic dimensions have a tolerance when combined but not on there own. Both measurement directions must be ascertained to compute the actual Position within the Tolerance specified in the Feature Control Frame.

There are more cases of Basic Dimension use per ASME Y14.5-2009 that can be discussed. For instance, Basic Dimensions used to locate DATUM targets. Best Practice is to include those in the FAI to show the FAI approver that they were achieved by using whatever device to achieve those Basic dimensions( like Gage Blocks). Simply record the dimension you created with the gage block that you created the DATUM Targets . Now some may come back and say that's redundant but I can tell you it will always pass and external audit without question. Not putting characteristics that are on the Drawing (excluding Reference Dimensions per AS9102) can lead to auditors asking questions and you having to justify why that they may or may not accept.

 

[3DDave]

If the argument is getting reports accepted by auditors who don't understand Dimensioning and Tolerancing but think they do, then reporting raw data is what you are doing and no doubt will be accepted. Just like in school, graders often like it when people show their work.

On the side of those who do understand D&T, there's no point to it. If the inspector can't do the job correctly then it doesn't matter if they do show their work and it's a distraction from evaluating the product variations.

Basic dimensions can't be measured. They are, at inspection, set-up instructions and the errors in them are in the measuring equipment, not the part. They are used as a source of comparison. It's only required to measure the deviation of a feature from e.g. its true position, et al, based on the basic dimensions.

The entire point of basic dimensions was to separate the nominal dimension from the tolerance and to allow the tolerance to be coupled to the feature size, as required, thereby allowing the creation of functional gauges.

Let's say two holes are [5.000] from each other perpendicular to a datum feature reference, but the first is [1.000] away and perpendicular to the same datum feature referenced in the feature control frame that controls them both.

Measuring the distance ([5.000]) between the holes is useless. What's important is the difference between [6.000] inches and the actual distance from the datum simulator/datum feature.

IOW one would indicate the CMM on the datum reference/datum simulator and set the start value in that direction to -1.000 and -6.000 to measure each of the holes, respectively. Only the deviation is measured. Since allowable deviations are typically diametral or width, the measured deviation would be doubled for the report. For X, Y conditions, the measured deviations can be combined to produce diametral values or the CMM could be switched to polar measurements to get the radial deviation directly.

If you feel like reporting raw measurements, that's OK. But the measured values aren't the 'basic' ones and can be the wrong distance/dimension to inspect at all. A linear pattern of 100 holes on a [1.000] spacing rarely means making 99 measurements of 1.000 nominal.

This sort of reporting and getting acceptance from unskilled auditors is more in line with

http://www.eng-tips.com/threadminder.cfm?pid=286

.

At the end I am unsurprised that SAE fails to include detailed guidance and that examples from aerospace companies include in their AS9102 reporting requirements, pictures based on what looks like the 197x and earlier versions of Y14.5.

For example, from Goodrich

I do see a document about this which includes the following (

https://www.sae.org/iaqg/publications/faq9102.pdf):

F11. Question:
Does "Reference Characteristic" (as defined in 9102) include both, "Basic" dimensions and "Reference" dimensions (as defined in ASME Y14.5-2009)?
F11. Answer:
The 9102 definition of Reference Characteristic is; "The characteristics that are used for information only” or to show relationship. These are dimensions without tolerances and refer to other dimensions on the drawing." Both basic and
reference dimensions fall under the definition of reference characteristics.

This page is similar:

http://www.mbccmm.com/cmmq/index.ph...ata-reporting-for-geometric-tolerances-part-1

(it has some delayed opening software.) "For conformance assessment, there is no requirement to report measured values that correspond to basic dimensions."

(added) And this document:

https://www.cobham.com/media/654302/supplierfai_tutorial_0312.pdf

"GD&T characteristics and their measured/calculated values must be recorded in the FAI. Basic dimensions may be omitted from the FAI."

 

[Nescius]

More discussion here:

http://www.eng-tips.com/viewthread.cfm?qid=411373

...Document the Hole Diameter with a ± tolerance, Basic Dimensions for X & Y direction, and Position tolerance Feature Control Frame. Yes the X & Y BASIC dimensions are theoretical however they ARE directly measured from the DATUM features called out in the Feature Control Frame. Those actual dimensions are recorded and compared to the basic dimensions to find the delta from each X & Y Basic Dimension (True Position). The outcome after the position calculation is called Position (Actual Position from the theoretical True Position in a diametrical circle/cylinder). So the Hole is measured DIRECTLY and related back to the DATUM Schema in a Cartesian manner (can be Polar too). The Basic dimensions have a tolerance when combined but not on there own. Both measurement directions must be ascertained to compute the actual Position within the Tolerance specified in the Feature Control Frame.

This is not true. You are not measuring the "location" of the hole...ever. There is no such thing. The hole must satisfy a prescribed state of its unrelated actual mating envelope, or a prescribed state of the hole surface.

Reporting the location of a hole with XY coordinates is not rigorous. It's very poor, actually, and falls apart quite readily, even on nominally rectangular/orthagonal/planar parts. On more complex, free-form parts, it's laughable.

You're speaking as if a position feature control frame is some upgraded method of XY tolerancing, when the geometric tolerance is actually a completely different concept.

 

[dgallup]

Quote 3DDave "The entire point of basic dimensions was to separate the nominal dimension from the tolerance and to allow the tolerance to be coupled to the feature size, as required, thereby allowing the creation of functional gauges."

I disagree with the above quote. Frequently the point of basic dimensions and position tolerances is to ensure mating parts fit. Many times the positional tolerances are too small to be accurately checked with functional gauges. One must use CMM or other measurement techniques. In the process of making the positional tolerance calculation, the CMM is measuring the X & Y basic dimensions as well as the hole diameter (or what ever the position tolerance is applied to) and then making the position calculation. I find it very useful to know the actual X & Y values especially when the position tolerance is exceeded and the part fails inspection. Knowing the actual location and size of the hole is very useful in figuring out how to correct the manufacturing process to make a good part. While I agree it's not necessary to "show your work" to accept or reject the positional tolerance, every machine on our floor does it and it's very useful when trouble shooting bad parts.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[Belanger]

dgallup -- I think I know what Dave was saying, but yes that statement isn't quite true. In addition to the CMM point that you made, we could also consider profile tolerances: They use basic dims, but the purpose isn't "to allow the tolerance to be coupled to the feature size, as required, thereby allowing the creation of functional gauges."

Considering all the other good stuff he wrote in the post, we'll let that technicality slide

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

 

[Nescius]

I find it very useful to know the actual X & Y values especially when the position tolerance is exceeded and the part fails inspection. Knowing the actual location and size of the hole is very useful in figuring out how to correct the manufacturing process to make a good part.

Rhetorically: Which end of a hole do you probe to report the XY "location".

 

[greenimi]

Rhetorically: Which end of a hole do you probe to report the XY "location".

The entire hole axis must be into the tolerance zone as "Unless otherwise specified, all tolerances apply for full depth, length, and width of the feature."
From here it is risk management and control plan.

How much risk you want to take to make sure the therory is fullfilled.

 

[Nescius]

Greenimi,

Yes, that's the point. A single XY pair is not necessarily meaningful at all.

A rhetorical question is one that is asked "with the aim of producing an effect or making a statement rather than eliciting information." In this case, I'm asking a question to highlight a flaw in what dgallup is proposing.

 

[greenimi]

Nescius,

I understand what you are saying, but I don't think dgallup was so far off in the weeds..... specially when you have to provide feedback for manufacturing.

Quote :"how to correct the manufacturing process to make a good part. While I agree it's not necessary to "show your work" to accept or reject the positional tolerance, every machine on our floor does it and it's very useful when trouble shooting bad parts"

And how do you know it is a single XY pair? Maybe XY pair down / up and another XY pair at cetrain length/depth is also used/reported.

However, I don't think basic dimensions should be reported, but some details on the CMM report (beside if the feature is IN or is OUT) could be very usefull.

 

[djhurayt]

I'm in the report the value camp. As it can be used by manufacturing to help identify the cause of the error.
i.e. technician says "All the error seem to be in the X direction, maybe that slide is worn, I should look into that. Maybe we can get some better control on our process"

Now, what I do wish be could get to happen is have some of our vendors stop reporting the basic dimension with a pass/fail.

 

[greenimi]

Per Evan J.
Y14.45 is in process to be released shortly.


"Y14.45 is not released yet. We're still working on it.
One of the reasons that I'm asking the question about reporting basic dimensions is to assess whether this is a practice that should (or even can) be standardized. As GD&T "purists", many of us have our preconceived opinion that basic dimensions are theoretically exact and therefore do not have measured values. At the same time, measured values of basic dimensions are often asked for and reported in industry (see JNieman's post above, for example). I've heard of similar things going on at a lot of other companies as well."

http://www.eng-tips.com/viewthread.cfm?qid=411373

Evan Janeshewski
Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

said:

Quote: "Yes, I'm a member of the Y14.5.1 subcommittee and also the Y14.45 subcommittee for Measurement Data Reporting. Part of the reason that I participate in GD&T and CMM forums is to get a better idea of what challenges industry professionals are facing, and what their opinions are on various issues. A new revision of Y14.5.1 is in the committee review stage and Y14.45 is nearing completion as well. There should be public review drafts coming out later this year (I'll keep you posted). I'm hoping for lots of feedback from members of this forum, especially on the profile section."

 

[DCarnley]

Hi Dave,
I have read your initial response and want to point out that the passing an audit is ancillary and a perk to the actual requirement if AS9102 in recording Design Characteristics on a drawing. If an inspector doesn't understand GD&T they should not be a precision part inspector. You seem to believe that it does not matter whether the process of machining yields 3 of 6 holes are -.003 in X and -.004 in y and the other three +.003 in X and .003 in Y as long as they pass the Position tolerance of .010. Yes you have 3 holes in the lower left quadrant and 3 holes in the upper right quadrant. The question is, is the process for making the holes and position satisfactory. The answer is yes for the FAI. However is the process in control? The answer is no. How will any Quality or Manufacturing Engineer be able to understand the process from the benchmarking of an FAI without the much needed information that shows the out of control condition? Sorry I'm big on continual improvement from the baseline FAIRs. Basic dimensions in GD&T position tolerancing have tolerance. It's is an aggregate of the position tolerance from the Feature Control Frame. The Basic dimensions are a function of that tolerance just as the Diameter is a function when LMC or MMC is when specified. Deviation from LMC or MMC has function to increase the tolerance in the Feature Control Frame thus changing it. I have seen where some inspectors actual were training to reduce the actual position by the factor of LMC or MMC and report the position as "0" if it reaches "0" or below. How would you respond to a actual position on the report as .000. I requested where in ASME Y14.5 it told them to do that. They could find it. That's because it doesn't exist. What does exist, is an addition to the position tolerance (most people and books affectionately call "Bonus Tolerance")

Dave I truly disagree with your position. Seems to me that not reporting the X&Y actuals of a position if a hole is a shortcut that Companies try and take to save time. I believe this shortcut takes away from the intent of the FAIR as benchmarking the production "Process" to continually yield acceptable material. Most likely Quality Engineer or MFG Engineer or the Company paradigm doesn't care about process improvement yield measurements with an acceptable Ppk.

 

[Nescius]

And how do you know it is a single XY pair? Maybe XY pair down / up and another XY pair at cetrain length/depth is also used/reported.

How many extra lines are we going to put on the FAI? One for every XY hole "location" at some number of Z depths? For every hole?

A single basic dimension might serve to locate 1 hole...or 1000 holes. What about dimension that are not shown on the drawing, like implied 90 degree angles, or implied zero distances? What about dimensions that we might not be able to measure, like small segments of a radius? There are countless things that might be defined with basic dimensions that demonstrate the silliness of the whole exercise.

There is a big difference between a FAI and a metrology report used to investigate bad parts. Anybody that blindly wants to include all basic dimensions on a FAI doesn't know what they are asking for. They're missing the point (and power) of geometric tolerances altogether.

As an aside, who's waiting until a FAI to detect a failure mode that "measuring" basic dimensions would show? In my world, we're not doing our job if something like that makes it to the FAI stage. It should have been caught much sooner.

 

[Nescius]

Seems to me that not reporting the X&Y actuals of a position if a hole is a shortcut that Companies try and take to save time.

I'll ask you directly: Which end of a hole do you probe to report the "X&Y actuals" of that hole?

 

[3DDave]

DCarnley -

I noted that a number of companies specifically tell their suppliers to omit basic dimensions and so does an opinion from the AS9102 group.

Any manufacturing engineer should be looking at the across-the-shop raw data for all the process to see what the shop capability is in various areas, as well as individual features on individual parts. That level of detail isn't appropriate for an FAI report.

As a customer, I would like to see long term statistics across the shop. It's been my experience that shops change equipment, operators, bits get dull, cutters get replaced. Unless there's a new FAI done every time that happens, FAIs really only mean that the shop could make one passing part and it's not a guarantee that the rest will conform. I have never seen a shop include Cpk data with their bid package or in their website, though perhaps some have; they mostly list the machines they own.

For holes, there are two main questions, 1) How far from the target location and 2) how far from the target diameter?, are your X-sigma values for the equipment and operators the part will be run on and what are the corresponding values for this particular part for the FAI?

If I worried that much I would witness the inspection as well as the entire fabrication process and would have required a series of pre-qualification operations. It's all good stuff, but not for a snap-shot report.

 

[DCarnley]

Nescius,
Thank you for your insight. I agree that some things don't need to be on the FAI like implied 90 degree angles, or implied zero distances. those thing are not on the Drawing and therefore are not required. I also agree that What Company waits till the FAI to fix the process and repeatability and reproducibility. Unfortunately many don't. But if I have a requirement to measure of a hole position in Y as 12X 1.000 ± .005 you would record is once, range it from Min and Max, or treat each hole individually and record the result for each hole? You as how many lines is necessary. I always fall back to the AS9102 to tell me.

3.6 DESIGN CHARACTERISTICS
Those dimensional, visual, functional, mechanical, and material features or properties, which describe and constitute the
design of the article, as specified by drawing or DPD requirements. These characteristics can be measured, inspected,
tested, or verified to determine conformance to the design requirements. Dimensional features include in-process locating
features (e.g., target-machined or forged/cast dimensions on forgings and castings, weld/braze joint preparation necessary
for acceptance of finished joint). Material features or properties may include processing variables and sequences, which
are specified by the drawing or DPD (e.g., heat treat temperature, fluorescent penetrant class, ultrasonic scans, sequence
of welding and heat treat). These provide assurance of intended characteristics that could not be otherwise defined.

3.10 FIRST ARTICLE INSPECTION (FAI) - [highlight #FCE94F]ALSO REFERRED TO AS PRODUCTION PROCESS VERIFICATION
[/highlight]A planned, complete, independent, and documented inspection and verification process to ensure that prescribed
production processes have produced an item conforming to engineering drawings, DPD, planning, purchase order,
engineering specifications, and/or other applicable design documents.

3.13 MULTIPLE CHARACTERISTICS
Identical characteristics that occur at more than one location (e.g., “4 places”), but are established by a single set of drawing
or DPD requirements (e.g., rivet hole size, dovetail slots, corner radii, chemical milling pocket thickness).

3.19 VARIABLE DATA
Quantitative measurements taken on a continuous scale (e.g., the diameter of a cylinder, the gap between mating parts).

c. The organization should consider the following activities during FAI planning and coordinate planning with the customer,
if required:
1. Determination of design characteristic inspection and sequencing for inspection of characteristics not measurable
in the final product.
2. Extraction of DPD design characteristics required for product realization that are not fully defined on 2D drawings,
including tolerances for nominal dimensions.
3. [highlight #FCE94F]Determination of objective evidence to be included in the FAIR for each design characteristic.[/highlight]

 

[DCarnley]

Nescius,
You are probably not going to like my answer on this because I normally do not take a 2D diameter on a three dimensional surface unless the a small depth of the cylinder precludes otherwise. Per ASME Y14.5-2009 and back to ANSI Y14.5-1973 version, the ø symbol in the Feature Control Frame is not a 2D diameter. It is a 3D cylindrical tolerance zone. In the old day (before CMM) I would use a surface Plate, Height Gage and test indicator to measure hole position. I measure the worst case deviation of the cylinder on both sided and use the side for the worst position. Now we have CMMs. We can project the cylinder axis. I program the CMM to measure each hole as a cylinder and bound the cylinder to the intersecting surface and the intersecting planar depth (Yes depth is a required measurement if it's on the drawing). CMM programs will calculate the worst case position at the intersection points. The shortcut is, if we believe the the machine used to create the cylinder is always exactly perpendicular to the adjacent surface or datum feature (not always the adjacent surface), then you only measure a 2D circle.

I know these shortcuts are timesavers. And I'm OK with Companies taking this approach to save money but they should stipulate so, like in a Contract or Procedure/ Work Instruction. To me the Normative Standards like AS9102 and ASME Y14.5-2009 are clear on their own. The grey area is created by individuals. Thus the continued debate over whether or not to document those Design Characteristics called "Basic Dimensions" on the face of the drawing.

 

[3DDave]

DCarnley -

From 3.10 - "have produced an item conforming to engineering drawings" Just one item. Not a way to the verify the process will remain in statistical control or act as a benchmark for getting better or worse results.

The implied 90s are on the drawing and are described by "and/or other applicable design documents." Failing to measure them is typical and the reason that large default angle tolerances are still on drawings, even though compliant parts would often be useless for purpose were they built to the limits of those tolerances. (<rant> Typical excuse is that extreme parts would be rejected for 'workmanship,' and so it doesn't matter what the tolerance value is, they won't be inspected anyway, closing the loop of ignorance.</rant>)

I don't know about the gray area - the AS9102 committee issued an opinion saying the basic dimensions are considered reference information. Since the basic dimension has no tolerance, inspecting it is of no value.

Is there anyone that publishes AS9102 guidance specifically requiring this reporting?

 

[Nescius]

...some things don't need to be on the FAI like implied 90 degree angles, or implied zero distances. those thing are not on the Drawing and therefore are not required.

The implied 90's and zero dimensions are as "on the drawing" as anything else. The information is required to make the part and it is absolutely communicated by the drawing. If it's not "on the drawing", where is it? Going further, what about model based definition? No drawing, no dimensions "shown" anywhere.

I measure the worst case deviation of the cylinder on both sided and use the side for the worst position. Now we have CMMs. We can project the cylinder axis. I program the CMM to measure each hole as a cylinder and bound the cylinder to the intersecting surface and the intersecting planar depth (Yes depth is a required measurement if it's on the drawing). CMM programs will calculate the worst case position at the intersection points.

As the hole straightness and orientation error become worse, the conclusions you are drawing from this procedure are worth even less. Distilling the "location" of a hole down to an XY coordinate is impossible. Tolerancing the location of a hole with +/- dimensioning is and always has been complete madness. The only way any of this is valid is if you assume perfect form and orientation...which leads me to:

The shortcut is, if we believe the the machine used to create the cylinder is always exactly perpendicular to the adjacent surface or datum feature (not always the adjacent surface), then you only measure a 2D circle.

So, you convince yourself that the machine drills perfect holes, perfectly oriented, but in the wrong spot...then you measure the location of a 2D best-fit circle at some cross-section of the hole...and consider this information so valuable that you spend tremendous amounts of time and money to gather and report it?

You're crippling the power of geometric tolerancing and the mathematical certainty of part definition that it tries to provide. Nothing is lost by not including basic dimensions on the FAI. There was nothing to lose. Thinking about the location of a hole in terms of XY coordinates is nonsensical. It's not limited to hole location, either. Most geometry is incapable of being described rigorously with only +/- dimensions, precisely why geometric tolerancing exists.

 

[axym]

Hi All,

Excellent discussion - it's great to see input from both sides of the issue.

To me, the discussion reinforces the idea that verifying conformance to the specified GD&T characteristics and providing data for process analysis/correction are two different things. This really helps to clarify the purpose and scope of the different standards and quality documents.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca