GD&T Datums on Ordinate 0-0 1

[sgjoachim]

Hi Everyone -- I have a question about specifying a GD&T datum at the 0-0 of an ordinate dimension scheme. The 0-0 is specified at the center of a hole, so is it correct to call out the horizontal 0 as datum B and the vertical 0 as datum C, assuming a perpendicular surface is already datum A?

We are having many a discussion about this at work, but I can't find examples of this in the Y14.5 spec. According to the spec, I would think you'd want datum A to be a perpendicular surface, datum B to be the hole itself, and datum C to be a vertical or horizontal surface to lock down rotation. That said, we have found examples in released industry specs showing the datum on the 0-0.

So is this a correct place to put a datum? If so, what is the datum? Is it a theoretical vertical (or horizontal) plane through the axis of the hole?

To help visualize the part, you can think of a flat metal plate that has several mounting bosses and features on it. The drawing view would be looking head-on at the mounting bosses and trying to use one of them to define the features control frame (in all directions except into/out of the page).

 

[KENAT]

I would think you'd want datum A to be a perpendicular surface, datum B to be the hole itself, and datum C to be a vertical or horizontal surface to lock down rotation" while the order of the letters is not technically relevant, you're correct that normally 'A' sets your primary 'perpendicular' 3 point plane, 'B' is derived from the axis of the hole as your secondary datum and then when you add in a third clocking datum you can derive 2 perpendicular planes through that axis oriented to 'C'.

Posting guidelines faq731-376

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

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

 

[drawoh]

sgjoachim,

A datum must be a real feature you can contact and measure from. In your case, the base of your part is the primary datum. The hole is your secondary datum. Your tertiary datum is another hole, or some face that controls rotation. If you are using a feature of size as a datum, such as your hole, it really ought to be an accurate feature of size.

This scheme allows you to draw ordinate dimensions from your hole.

I think I have re-phrased what KENAT said.

--
JHG

 

[JNieman]

You x and y axis, or horizontal and vertical vectors, are theoretical and intangible. How do you use a cylinder to ascertain 3 o'clock and 12 o'clock without features of alignment? You cannot.

Just to clarify the root of the issue KENAT and drawoh are addressing.

_________________________________________
NX8.0, Solidworks 2014, AutoCAD, Enovia V5

 

[powerhound]

If you're asking if the axis of the hole can serve as both datums B and C then the answer is no. KENAT and drawoh have offer two suitable solutions. Make the hole a secondary datum feature and either another hole or an edge of a planar feature as the tertiary datum feature. You can't put a datum identifier on a centerline in any situation.

John Acosta, GDTP Senior Level
Manufacturing Engineering Tech



SSG, U.S. Army
Taji, Iraq OIF II

 

[Belanger]

To the OP: I agree with all previous posts. I often come across prints where a horizontal center line is labelled as B and a vertical center line is labelled as C. This simply isn't allowed by the Y14.5 standard. While it's true that the axis of the hole is the theoretical datum, the triangle symbol for a datum must always be tagged with a physical feature (not an imaginary line).

See paragraph 3.3.2 (a) through (f) in the 2009 edition of the standard in order to help steer your colleagues in the right direction.

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

 

[3DDave]

More exactly, it's the axis of the theoretical datum simulator, which may not coincide with with the axis of the hole for a related datum feature. And, in the case of a feature of size referenced with MMC it doesn't require contact with anything. If using a fixed gauge there can be clearance between the gauge and the feature.

I'm really curious how the center of the hole is specified as 0,0. That specification is usually handled by the datum features referenced in the FCFs.

To answer part of the original question - yes, there are at least two mutually perpendicular theoretical datums that can pass through a theoretical axis, but they are typically not labeled. Their location and orientation are a composite of the choices made in the datum reference part of the feature control frames.

Look at the figure for an inclined datum feature to see that a theoretical datum doesn't have to be coincident or parallel to the datum feature that is key to its location.

 

[sgjoachim]

Thanks everyone for weighing in. Definitely sounds like there's a clear consensus! I'm a young engineer and new to GD&T so I was wondering if maybe the datums on each "0" of the ordinate origin was an industry accepted practice that I just didn't know about. It sounds like we should steer clear of that though. I can think of several other ways to establish the datum reference frame that are more in line with my understanding of the Y14.5 spec, so I'll stick with that going forward. Thanks again.

 

[JNieman]

Just to expand on the topic, I thought I'd throw in a couple cents on the topic of coordinate/ordinate dimensioning and datum schemes with GD&T controls. It's not 'illegal' to use datums that coincide with your coordinate zeroes. In fact, it's often wise, when "design[ing] for manufacturing" to keep these benefits in mind for some manufacturing methods. For CNC controlled machinery, they are all working from a zero, anyways. Depending on the size of the part, deviations from nominal often get worse as you get further from your programmed origin. This isn't universal but it's often a thing to keep in mind for some manufacturing situations.

I'm not suggesting you move your datums to coincide with your coordinate origin, but rather the opposite. Your datums should be the wisest choice based upon the function of the part. Your zero does not need to be external surfaces such as if you were to but a part up against two walls (or a fixed vice jaw and a part-stop for the sake of visualization). Your zero can just as easily be the center of a cylindrical feature such as a mounting-boss in your case. Your Z-zero may be the mounting surface level. Your coordinates can origin there. However your Datums may be a derived from two mounting bosses (either a line tangent to both, or center-to-center).

You don't have to sacrifice convenient/logical coordinate origin location to suit Datum schemes. If they overlap, it can have some benefits. Also keep in mind that ordinate dimensions can be "basic" as well.

_________________________________________
NX8.0, Solidworks 2014, AutoCAD, Enovia V5

 

[powerhound]

You don't have to sacrifice convenient/logical coordinate origin location to suit Datum schemes.

IMHO...sometimes you do. Convenience should always take a back seat to function. In those cases where convenience has an equal outcome as function then fine. The problem is that it most often does not work out that way. I'm not saying convenience should be disregarded but it should be considered only after functional requirements.

John Acosta, GDTP Senior Level
Manufacturing Engineering Tech



SSG, U.S. Army
Taji, Iraq OIF II

 

[KENAT]

sgjoachim I think maybe you misunderstood, or we all misunderstood you.

It may be entirely appropriate to put the axis of the hole as 0,0 of your ordinate dimensioning scheme.

Since your tertiary datum is only used for 'clocking' then you arguably don't need to worry about your basic dims chaining back to it.

So 0,0 of your ordinate coinciding with hole CL is probably correct.

Posting guidelines faq731-376

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

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

 

[tbuelna]

I would agree with powerhound. When it comes to GD&T tolerancing schemes, the primary consideration should always be minimizing rejection rates of out-of-tolerance parts. Making things easy for the designer is no justification for how the tolerancing arrangement on a drawing is established. It costs very little to change the tolerancing on a drawing before release. But it can cost a huge amount of money if the tolerancing used on the drawing results in an excessively high rejection rate of the associated part production.

 

[JNieman]

IMHO the primary is not eliminating out of tolerance parts, alone, but rather eliminating parts that won't function. The tolerance should seek to accomplish that goal. But without when designing a part; applying tolerances, selecting datums, etc, you have to keep function in mind as the absolute primary requirement.

I do not disagree at all that there are times when coordinate origins cannot overlap with Datums. I didn't meant to imply otherwise. I meant to say that you should design for manufacturing as much as you do for the engineering requirements. Making the manufacturing easier, combined with inspection, will reduce rejected parts. Making the coordinate origin overlap with the datum scheme (when possible) is not about convenience for the designer. It is no more or less difficult to origin your dimensions on any other feature. It's just clicking around to place dimensions, no matter the origin. It's about making the manufacture as seamless and simple as possible in order to get the best rate of quality parts.

Apologies if I was vague or misleading, before.

_________________________________________
NX8.0, Solidworks 2014, AutoCAD, Enovia V5

 

[sgjoachim]

KENAT I'm with you. My plan would be to keep the ordinate origin at the center of the most important mounting boss. The datum reference frame would be established with three datums, in order primary to tertiary: the plane coincident with the top the mounting boss (where it mounts to the mating part), the hole at the center of that boss (RFS), and a horizontal or vertical surface at the outside edge of the part for clocking.

The ordinate dimensions would be basic, locating other mounting holes, and we'd have a position tolerance on those holes (and size if they aren't tapped). Other critical features would be dimensioned/toleranced as appropriate relative to that same datum reference frame. Then we want a note to tolerance all undimensioned (CAD) geometry to a standard range relative to that datum reference frame. I'm thinking the feature control frame in the note would be a profile tolerance, as I've seen in some examples.

On a related note: There are multiple coplanar mounting bosses. Would it be more correct to specify the top surface of one as the primary datum and THEN establish coplanarity with the rest? Or would you establish coplanarity and make that coplanar surface the primary datum. The latter makes more sense functionally, since they all mate to the same part, but I'm not sure what effect (if any) it would have on my datum reference frame (or if it's even allowed!). Note: I'm establishing coplanarity by having a profile tolerance that points to a phantom line connecting the bosses and a note "[X] SURFACES" as shown in the spec. Maybe I'm overthinking it, but the idea of coplanar surfaces as a datum was confusing for me.

 

[3DDave]

You mean like how a chair can't reliably be oriented to the floor with multiple legs that have coplanar ends?

 

[KENAT]

"Or would you establish coplanarity and make that coplanar surface the primary datum."

Yes, see 4.5.7.1 & 6.5.6.1 of the 94 edition Y14.5 std. Not explicitly shown but I believe readily extrapolated.

I believe the newer version of the std may allow other options but don't have a copy to say for sure.

Posting guidelines faq731-376

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

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

 

[akitasumo]

If you are using Y14.5-2009, there is a new concept of using a hole pattern as a datum, so you would have only Datums A and B, see fig 4-26.

 

[KENAT]

ASME Y14.5M-1994 also allowed patterns of holes as datums akitasumo.

Now how many people really understood/understand the implications is another matter.

Posting guidelines faq731-376

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

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

 

[3DDave]

The use of patterns of features as a datum reference was in 1982 as well.

What the 2009 version offers is the suggestion that multiple perfectly located and coordinated expanding features with a frictionless interface that allows the physical datum simulator to glide over the surface of the part and be perfectly constrained in 3 degrees of freedom would allow the use of the RFS modifier to a pattern of holes.

It doesn't go into all that detail because it's just a theoretical inclusion in the standard, but that is what would be necessary to make use of the concept. I still have seen no real mechanism that depends on multiple interference fits and allows understanding where a part will move to when constrained in a similar fashion without taking a significant amount of stress/strain analysis to determine where the deformed interference fit pieces would deflect to and how that would alter the location of the mating parts.

I asked for an example during the public review and received only the suggestion that it was there to allow the concept, not to make use of it.