Does the Primary Datum always have to be perp to a hole? 2

[randy64]

For the last 25 years or so I was told and believed that the primary datum for locating a hole should be the datum that is perpendicular to that hole. So, if I have holes on two surfaces that are perp to each other, one could have datums A,B,C, while the other could be C,B,A (order of datums in Feature Control Frame).

A year or so ago I was "converted" to a new way of thinking, namely, that if the holes that are perp are related to each other, then they should have the same primary, secondary and tertiary datums (both A,B,C).

This makes sense to me as it was explained that if you inspect to a different datum reference frame for each set of holes, you lose the relationship between the holes that are drilled into different surfaces.

Is this generally accepted now? Are there any sources or resources that give a good explanation of this? I'm asking because I'm meeting some resistance based on "we've always done it that way!"

Any and all help is appreciated. Thanks!

 

[weavedreamer]

There is always a relationship between the holes, and if properly defined it can be quantified. Changing datums may introduce additional tolerance.

Functionality should be the factor that helps guide the decision of which datums to use, and which order of precidence is required.

 

[looslib]

Look at the datums as inspection references. Place the part on Datum A, push it up against Datum B, and align it with Datum C. Now inspect as much of the part as you can from this orientation. Additional setups for inspection add cost to the inspection process and the part.


"Wildfires are dangerous, hard to control, and economically catastrophic."

Ben Loosli

 

[pmarc]

Primary datum does not need to be always perpendicular to a hole. First of all datums should be selected to reflect a function of a part as close as possible.

If you need resources, this one might help:

http://www.tec-ease.com/gdt-tips-view.php?q=121

 

[fsincox]

In a word, NO! It is a common way used and shown as applied to fasteners, but, it is not a requirement. The design should dictate.
Frank

 

[drawoh]

randy64,

No.

Your datums tell the fabricators and inspectors how to fixture your part. This has little to do with where your holes are or how they are oriented.

On a drawing properly dimensioned to ASME Y14.5, most if not all your feature control frames should show the same set of datums, in the same sequence.

JHG

 

[pmarc]

On a drawing properly dimensioned to ASME Y14.5, most if not all your feature control frames should show the same set of datums, in the same sequence.

I would not formulate such conclusion. It strongly depends on how complicated a part and its function are. I am closer to say that parts out of the real world (not the ones from GD&T handbooks) tend to have more datum features than 3. Of course I would agree that the less different setups for measurement the easier for inspector, but sometimes it is simply impossible to live with only one datum reference frame and satisfy all the requirements.

So it does not mean, the drawing where 10 part features are referenced to e.g. 5 different DRF's is not properly dimensioned to Y14.5. If proper functional relationships are captured then everything is OK.

 

[Belanger]

In simplistic terms, datums tell you how to grab onto a part to hold it for checking. So there's no reason why a hole's datum must be perpendicular to the hole. The reason we often think that is because many holes have pins assemble into them and the mating part then flattens out on the surface. But if there is a need to arrange the datum sequence differently, then go for it.

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

 

[dingy2]

Datums reflect how the part assembles on its mating part with the primary datum on a non-cylindrical part, being its mounting surface. Holes on that surface are usually perpendicular but that is not absolutely necessary.

The secondary and tertiary datums on a non-cylindrical part are usually holes since holes are more important to its function than the perimeter feature.

Dave D.

http://www.qmsi.ca

 

[fsincox]

pmarc,
I am in agreement with you. I do believe we over simplify for OUR own convience many times.
Frank

 

[Belanger]

Just thought I'd throw this in...
One question that I often use during GD&T training is, "What is the maximum number of datum features allowed for a single part?" There really is no maximum. Some folks say 3, but that is only the maximum number of datum references in a single feature control frame. Some folks say 23, because that is the number of letters of the alphabet that are available for use. (But double letters and beyond are allowed.)

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

 

[dingy2]

J-P:

Depends upon the part

Dave D.

http://www.qmsi.ca

 

[drawoh]

pmarc,

Most of the time, I use the same set of datums for all the GD&T on a drawing. Sometimes, I need to connect two features, and then I need an additional datum. For me, where I work, this is unusual. Your job may be different.

It sounds to me as if the OP is assuming that the datums change for each face with holes in it. In this context, his question makes a lot of sense. This approach makes a lot of sense, but it is not what the standard requires.

JHG

 

[pmarc]

drawoh,

That's why I said it depends how complex the part is and indeed I do work with parts where 3 datums are simply not enough.

By challenging your statement I just wanted to warn before generalizing and simplifying different things. This is how misconceptions are born and spread. 'Primary datum always perpendicular to a hole' and 'The same set of datums in the same order in every FCF' are just two examples of those.

 

[Belanger]

Not sure what you mean, Dave. The Y14.5 standard does not prescribe a maximum number of datum features that can be identified on a part. Thus my statement.

I suppose if you have a sphere, for example, then there is a limit. But I was simply talking about the interpretation of the standard, not a particular part.

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

 

[dingy2]

J-P - I agree with you that the standard does not prescribe a maximum number of datum features on a part. My statement (rather short one too) only notes that the number of datum features on a part depend upon its complexity. That's all.


Dave D.

http://www.qmsi.ca