General orientation for linear dimensions

[steveapathy]

I have designed a part containing both GD&T and ordinate dimensions. I have the edges of the part (an irregular shape, but mostly orthogonal) all defined by ordinate dimensions from a common origin, which is the primary locating hole in the part. Since a hole provides no rotational reference, the orientation of the part during measurement is not totally clear.

The part itself is assembled using this hole (near the bottom right corner of the part) as the primary locator, and a slot (near the top left corner of the part, in line with the hole) as the rotational control. For the purposes of the GD&T positional callouts, I have the hole as datum B, and the slot as datum C. Datum A is the large flat surface of the part.

When this part is inspected, they would most likely just use one of the long vertical edges to align the "y" axis, and measure all linear dimensions based on that orientation. Since the edge isn't what aligns my part in my assembly, that's not really capturing design intent. What I really want them to do, however, is align the part based on |A|B|C| prior to measuring linear dimensions from the alignment hole.

I know that a line profile all around my part would fulfill this need, but the vendor we receive the part from is insistent on having it defined by linear dimensions for inspection and qualification purposes.

Does anybody have any advice on how I can make it clear how to orient the part during measurement?

I included a made up part that illustrates what my part is like (made quickly, so ignore the fact that it's not fully defined... the real part has other features with a positional GD&T callout using |A|B|C|).

 

[3DDave]

GDT includes ordinate dimensions. What you are creating are FCFs - Feature Control Frames.

To make it clear use profile of surface all around with basic dimensions to locate the perimeter. And promise to pull the contract if more clarity is required.

 

[steveapathy]

Dave,

Thanks for the reply. As I mentioned, yes I believe a profile with basic dims would fill the need, and I understand FCFs. Sometimes it's just not worth the fight to force GDT upon a vendor that does not have expertise in it, if linear dimensioning can be used sufficiently in its place, and it's not always practical to just switch to another vendor.

If we put GDT aside, and let's just say that I have nothing but ordinate dimensions on the part, do you have any tips for defining orientation of the part during measurement? The vendor is comfortable with using the hole and slot to align the part at this point, and I'm mostly looking for any advice on if there is a standard way of defining that requirement.

Thanks in advance,

- Steve

 

[CheckerHater]

The "standard" way would be to use GD&T, but you already rejected it.

Either way, show the angle between slot location and your "zero" coordinate. This will orient your measurements with relation to datums you chose.

The angle should be basic, but I am afraid your vendor will not like it either

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[TheTick]

Checkerhater needs to check the OP. The vendor is the one rejecting GDT.

 

[jassco]

Hi, Steve:

I don't think there is such a thing named "standard way". You just need to create your print(s) in accordance with ASME Y14.5.

Profile with basic dimensions is way to go. Alternatively, you can use +/- tolerances for those dimension that you specified. "Orientation" of the part does not exist. You have to control "orientation" (virtual) via those surfaces which are touch points of your CMM stylus.

Best regards,
Alex

 

[drawoh]

steveapathy,

Draw a line between the holes to show explicitly that they line up. SolidWorks does this for you.

The shop cannot use your holes as datums without making a fixture. If this were my piece to inspect with the tools I can get at, I would pick up your round hole and one edge. I would measure to everything else. I would covert my results to your co-ordinate system. If I am the fabricator, I will line up the one edge, whether or not you care about it.

--
JHG

 

[steveapathy]

I just browsed through ASME Y14.8-2009 and found several examples (figure 2-1 is the first) with the following note:
1- DIMENSIONS ARE RELATED TO DATUM A (PRIMARY) DATUM B (SECONDARY) AND DATUM C (TERTIARY)

This is related to machining operations on a cast part, but seems to be sufficient for my application.

EDIT: 14.8 also has a section for general notes (5.3) that includes an example for "ALL TOLERANCES APPLY WITH DATUM FEATURES D AND E CONSTRAINED"

Not sure why 14.8 has these examples, and 14.5 does not. There is nothing I've seen that makes them specific to the principles in 14.8, or that excludes their validity in 14.5.

 

[pylfrm]

steveapathy,

You might find thread1103-411755 to be of some interest.

If your vendor will not accept standard ASME Y14.5 profile tolerances, why would they accept a note from ASME Y14.8 that essentially turns your directly toleranced ordinate dimensions into something equivalent?

I will say that it's probably none of your business how the vendor orients the part during measurement. If they want to use some arbitrary surfaces that are convenient for them, and then do the appropriate coordinate transformation, why not let them?


pylfrm

 

[steveapathy]

Thanks, pylfrm. That thread is directly relevant to what I'm looking to do.

What some people have a hard time understanding is that not every vendor is educated in ASME Y14.5, or exactly how to interpret profile tolerances. What every vendor DOES understand, is how to pull a direct dimension from one feature to another. This is a vendor that supplies thin film, cut on steel rule dies. They don't have a CMM, and they have no need for one. They don't even have an automated video measurement system (i.e. SmartScope). We could work through it and translate their direct measurements into a profile compliance, but why go through all the extra effort if direct dimensioning can sufficiently define my part? Why make them measure every edge to the same tight tolerance, or confuse them with multiple start/stop points for profile callouts, if I only truly care about the position of a few critical edges?

The problem with not specifying an orientation, is that all of the dimensions are coming off of a hole. With a hole, what stops them from have the part rotated 10° on their measurement table? What does a rotation of 10° even mean? What's the reference point?

Sure, they can put the part on their table any way they want, but how do they perform an appropriate coordinate transformation if they're given no reference for what to set as the 0° reference angle? How do they know their y-axis should be set at xx° off of the hole-slot alignment without some guidance from the drawing?

It's very easy for me to describe these things in words to them, and to get them to measure the parts appropriately. That part is already done, but I need it baked into our contract (aka our drawing) as well. I could translate those words onto my drawing in a dozen different ways. I prefer to specify it in a way that is indicated on one of our relevant standards (14.5 and 14.8).

 

[Belanger]

why go through all the extra effort if direct dimensioning can sufficiently define my part?

What constitutes a sufficiently defined part? The entire purpose of the GD&T system is that I could show you a simple rectangular plate that most people would say is sufficiently defined, yet there could be some loophole that allows for a non-functional part to be made. The vendor could point to the drawing and say, "Well, you didn't indicate that that surface was only the secondary datum!"

Profile (and other GD&T symbols) are not there to make the drawing seem more complicated -- they are there so that the drawing cannot possibly be misunderstood. In other words, it covers your butt. There are all sorts of horror stories where hundreds of thousands of dollars had to be eaten by the customer because the vendor's parts were no good -- yet they met the requirements of the drawing to the letter of the law.

Why make them measure every edge to the same tight tolerance?

Who says it has to be tight? You can use profile but simply dial the tolerance number up to be larger. It's all about ensuring that every portion of a part is defined without making assumptions.

I prefer to specify it in a way that is indicated on one of our relevant standards (14.5 and 14.8).

And you're absolutely correct. Any vendor who doesn't want to learn proper interpretation of Y14.5 isn't going to be in business very long. Sure, you don't want to overwhelm them with GD&T if they're not familiar with it, but at the same time the drawing must be air-tight. So take a gradual approach, I guess.

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

 

[mkcski]

J-P:

I totally agree with your last comment. Knowledge and training - the lack thereof - are the issue and not the drawing. Its like saying if one person cannot read no one should write books - geez. We had a similar GDT issue with a supplier we did not want to loose. We ran and GDT class for their staff and opened up a line of communication for them to ask GDT questions - problem solved.

Certified Sr. GD&T Professional

 

[CheckerHater]

What every vendor DOES understand, is how to pull a direct dimension from one feature to another.

Exactly how?

http://files.engineering.com/getfil...-0d94-4f39-a70b-4472d653614f&file=Capture.JPG

With a hole, what stops them from have the part rotated 10° on their measurement table?

The slot, the slot! Establish an angle between your hole-to-slot direction and your "zero" direction. Your slot is your clocking datum.


"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[steveapathy]

What constitutes a sufficiently defined part?

Design intent and good engineering evaluation does. While I 100% agree that GDT is a better method of designating design intent, it's not always necessary to ensure a properly functioning part.

Who says it has to be tight?

Again... design intent. Some edges need to be more tightly-toleranced than others due to required clearances. Others have a mile of clearance to the rest of my assembly, so I wouldn't want to apply the same tolerance to the entire perimeter of the part.

Sure, you don't want to overwhelm them with GD&T if they're not familiar with it, but at the same time the drawing must be air-tight. So take a gradual approach, I guess.

That's exactly what we're doing. We have position callouts for parts of this sub-assembly, we're just not pushing the profile callouts on this, because there's no functional need. We can't afford to teach every vendor GDT.

Its like saying if one person cannot read no one should write books

No, it's like saying the vendor can't speak Finnish, so I translate the book into English for them.

 

[steveapathy]

The slot, the slot! Establish an angle between your hole-to-slot direction and your "zero" direction. Your slot is your clocking datum.

Yes, I know that.. but how do THEY know that if I don't supply them with those instructions? That's the whole point of this thread...

To be clear, I've already resolved my issue.

This thread is becoming more of a "Why wouldn't you just use GD&T for everything?" discussion. While no longer helpful to my initial question, I'm hoping to shed some light as to why others may choose to continue using linear dimensioning.

 

[Nescius]

This is a vendor that supplies thin film, cut on steel rule dies.

Profile tolerance + optical comparator + transparent overlay

This is old technology, not complex or expensive, and can work quite well. If the part will fit on the machine, the comparator is arguably the perfect tool for checking things like this.

Why make them measure every edge to the same tight tolerance, or confuse them with multiple start/stop points for profile callouts...

Start/stop points are not needed. You can draw the non-uniform tolerance zone boundaries with phantom lines, defined with basic dimensions. See section 8.3.2 in the 2009 standard.

 

[steveapathy]

Profile tolerance + optical comparator + transparent overlay

This is actually how we had them inspecting parts early on. Unfortunately, that was only giving us a pass/fail, and we prefer to have more concrete information on the true shape of the part for qualification purposes. I suppose I could have basic dims to the relevant edges, and have them report on that in addition to pass/fail... but I see another thread is already on-going regarding whether or not basics can/should be reported, so let's leave that one alone in here...

Start/stop points are not needed. You can draw the non-uniform tolerance zone boundaries with phantom lines, defined with basic dimensions. See section 8.3.2 in the 2009 standard.

That's a good suggestion for future applications. Thank you.

 

[Belanger]

Steveapathy -- one quick follow-up. I asked "Who says it has to be tight," and you replied:

Again... design intent. Some edges need to be more tightly-toleranced than others due to required clearances. Others have a mile of clearance to the rest of my assembly, so I wouldn't want to apply the same tolerance to the entire perimeter of the part.

The solution is not to avoid GD&T; the solution is to simply have two (or more) different profile tolerances, one being the required tight tolerance and the rest having a generous profile tolerance.
If you don't use profile then you're still going to have to tolerance the perimeter of something, yet the tolerance would be of the plus/minus type, which is an ambiguous method.

It's not enough that we make drawings that can be understood. We must make drawings that cannot be misunderstood. (Big difference!)

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

 

[steveapathy]

Nobody needs to convince me that GDT is a superior method of defining a part. I've been convinced of that for a long time. But...

the tolerance would be of the plus/minus type, which is an ambiguous method

I disagree with the notion that a plus/minus dimension is always ambiguous. Often, yes... but not always.

I think section 2.7.1 of Y14.5-2009 nicely lays out plus/minus interpretation for form.

Section 2.6.1 lays out how you can specify an origin to remove ambiguity in a distance tolerance.

Are there going to be cases where it's still not totally clear what the intent was? Absolutely.

If you'd like to point out how you think there's still ambiguity in the examples in Fig 2-5 and 2-6, I'm ready to learn. Otherwise, can we agree that plus/minus tolerances are not ALWAYS ambiguous?

 

[pylfrm]

steveapathy,

I don't think anyone means to suggest that directly toleranced dimensions are always ambiguous, just that they usually are for cases like the outside profile of your part. As I alluded to in my previous post, the Y14.8 note mostly eliminates that ambiguity.

If the vendor is already on board with position tolerances, then presumably they understand basic dimensions and datum feature referencing. From that baseline, it shouldn't take much more than 30 seconds on the phone to explain profile of a surface for simple applications.

Actually, I'd say profile is rather straightforward compared to position tolerancing. There's no need to worry about unrelated actual mating envelopes, modifiers such as MMC and LMC, or two conflicting interpretations in the standard.

pylfrm