G D and T: Can three planar surfaces be chosen as Datum Features?

[alexengineer]

I have a question to ask u folks. Can three planar surfaces of say a rectanglar block be chosen as datum feature A, B and C? Say this is a plastic part... we know that the actual planar surface will not be perfectly mutually perpendicular to each other.. Aren't we overconstraining the object by using three planar surfaces as datum features?

 

[EnglishMuffin]

The convention in GD&T is that the order in which the datums are specified in a position or other callout is important. Lets say they are A, B, & C in that order. For the purposes of that callout only, datum plane A is then understood to be defined by three points, totally fixing it in space. Datum B is defined by two points, and C by only one. To be really precise, it is possible to specify the location of the points, but if this is not done the points are taken to be as far apart as possible in directions which are consistent with the intent of the callout.

 

[alexengineer]

What u are saying is that I should define the three datum A,B and C are by Plane, Line and Pt?

Say a rectangular block. Shouldn't it be that I define by three planar surfaces are datum features? While the actual part surfaces(due to surface defect, warpage..etc) are not perfectly mutually perpendicular to each other, using the order of callout A, B and C, my understanding is that we are still able to construct three mutually perpendicular planes(thus getting the Datum Reference Frame)

From CMM pt of view, the CMM operator will check on first planar surface by clicking CMM stylus on more than 3 pts. For this, he/she is able to establish the flatness of Datum A. From the CMM machine, a best fit plane A is generated. Then he click on planar surface for datum feature B. Using A as a reference, The best fit mutually perpendicular plane B(Imaginery Datum B) is formed. Similar for datum C using Datum feature Planar surface C.

Is my understanding correct or must we use only Plane,LIne and Pt?

 

[desertfox]

Hi alexengineer

Yes you can have three datums A,B,C on three mutually perpendicular planes and these would represent axis X,Y,Z.
Each surface might require a tolerance of say flatness in its own right but this is not mandatory. The three datums
are theoretically assumed to be at 90 degrees to each other.
From these datums other datum features can be established for instance datum D = datum A + datum B + datum C.
When you insert two or more datum reference letters into a tolerance box in which each letter is in a box of its own,then the order in which they are specified is the order of their importance or priorty.
However reading your last post if datum B is checked against
datum A for perpendicular then why make it a datum just give
datum B a tolerance for perpendicularity to datum A. Similar
for datum C just give a tolerance for squareness to either
datum A or datum B .

hope this helps

regards Desertfox

 

[alexengineer]

Desertfox,
As to your answer"if datum B is checked against
datum A for perpendicular then why make it a datum just givedatum B a tolerance for perpendicularity to datum A.Similar for datum C just give a tolerance for squareness to either", the reason why Datum B is referred to A is because that is the only available datum that is referred to. I am doing that cos it's only when all Datum A, B and C are firmly referred to(C related to A and B and B related to A) that the Imaginery Datum Reference Frame(DRF) is established.


Based on your answer.. say in the example of a rectanglar block, we can select the three(supposedly in theory mutually perpendicular) faces as datum features although the actual part is not mutually perpendicular. A set of imaginery mutually perpendicular planes A, B and C is only established by best fit in order of its relation(B related to A and C related to A and B)

I am asking this question cos there was a question that came up from my work. It was suggested by my peer that I need jus Plane, Line and Pt to establish my imaginery DRF. Anything more than that is "overconstraining" my part. My take is that Plane, Line and Pt is the minimum to constrain a part.I could specify Plane, Plane, Plane or any other combination so long as all the degree of freedom is taken care of. While in reality, three planar surfaces of an actual part is not perfectly mutually perpendicular,
an imaginery mutually perpendicular DRF can still be established from the actual part planar surfaces through best fit in order of A, B and C. Is my understanding correct?

 

[desertfox]

Hi alexengineer

Can you give more detail on why you need three datums in terms of the component your dimensioning you only need to tie three faces upto each other if its important.


regards desertfox

 

[EnglishMuffin]

alexengineer: Your understanding is basically correct. I quote the following from GeoMetrics III by Lowell Foster (which I recommend by the way) :

The PRIMARY datum surface A establishes the first relationship of the part for proper orientation. The three (minumum) extremities or contact points of the surface establish the primary datum plane A
The SECONDARY datum surface B establishes a further reationship of the part for proper orientation. The two(minimum) extremities or contact points of the surface establish the secondary datum plane B at 90deg BASIC to datum plane A
The TERTIARY datum surface C completes the part orientation. One (at least) extremity or contact point of the surface establishes the third datum plane C at 90deg BASIC to both datum planes A and B.

End quote

In other words, it is not necessary, indeed it is unlikely, that planes A, B and C are actually exactly at 90 degrees to each other. Only the first plane (A) is fully utilized. Also note, as I said previously, that it is very important to get A, B and C (or whatever symbols you are using) in the correct order in your callout. If you get involved with CMM checking much, you will find that you have to do this sort of thing all the time.

 

[EnglishMuffin]

alexengineer: Also, since there seems to be some confusion on the issue, I should stress that it does not matter what other geometrical relationships may or may not have been used in defining A,B or C in isolation - squareness to each other for example. For the purposes of a particular callout (typically position) where you wish to use all three to create a datum frame, the convention given in my previous posts applies.

 

[alexengineer]

Thank you all so much for the reply. Really appreciate that...

 

[imagitec]

As has been noted, the perpindicularity of the surfaces - as GD&T defines perpindicularity - is not controlled by defining them as datums. Nor is the flatness of the individual surfaces. It's best to imagine how the part would be inspected: Start with three mutually perpendicular, accurate inspection surfaces. In ABC order, bring the datum surfaces of the part into contact with the inspection surfaces. The part surfaces can be rough as pavement (most GD&T manuals exaggerate the surface roughness to make this point), yet this procedure can be repeated over and over with the same result. That's the point.

If the flatness of a datum surface is functionally important, apply a flatness tolerance. If perpindicularity or parallelism between it and another surface is important, apply the appropriate tolerance (this also applies an implicit flatness control on both surfaces).



Rob Campbell

http://www.livejournal.com/users/robcampbell

 

[TheTick]

Placing a part on three planes "A", "B", and "C" for inspection.

This assumes "A" is primary, "B" is secondary, and "C" is tertiary. Imagine a fixture on a table where "A" is the horizontal surface, "B" is a side wall perpendicualr to "A", and "C" is a rear wall perpendicular to "A" and "B".

1. Part is placed on horizontal surface. "A" surface rests on "A" plane. Theoretical 3 point contact.

2. Part slides over to "B" wall so that "B" part surface contacts and is oriented by "B" wall. Theoretical two point contact.

3. Part slides in contact with "B" wall until "C" surface contacts "C" wall. Theoretical single point contact.

Measurements/gaging of feature positions are now actually taken from the fixture walls, not the part datum surfaces.

Due to illness, the part of The Tick will be played by... The Tick.

http://www.EsoxRepublic.com