Cylinder surface as a datum

[cwdaniel]

I have an old drawing showing a cylinder in profile and specifying the side of the cylinder as a datum. That datum is referenced by another cylinder elsewhere in the weldment, making a side of that cylinder parallel to the datum.

Is this an allowable call out? I'm thinking the datum and feature control should be referencing the axis, not the sides.

 

[ewh]

You are correct that the datum is referencing the axis. The location of the callout is also a correct one, if it is simply pointing to or connected to the profile of the cylinder.

 

[drawoh]

cwdaniel,

Your datum must be a real feature that can be fixtured to. With ASME Y14.5M-1994, a datum on an axis is a mistake.

Your drawing probably is correct.

Apparently, an ISO standard somewhere allows you to place datums on axes, but I have not seen it explained. I have no idea what it would mean.

JHG

 

[cwdaniel]

These cylindracal features are actually tubes so I'd say the ID of a tube can be fixtured to if it's important.

Can't holes or other ID's be used as a datum and that applies to the axis of the feature?

In this case though the outside is probably more useful for fixturing.

 

[ewh]

Yes

 

[dingy2]

One can have one side of a cylindrical feature as a datum but that is unusual. If the Designer really wanted this, then there would be an end view with a phantom line tangent to the diameter and the datum designation symbol attached.

I belive that your Designer wanted the C/L of the cylindrical feature and should have reflected this a bit differently.

Holes and other IDs are used as datums and should be. It depends upon how the part assembles.

If the part assembles with bolts protruding throught some holes, then one of the holes is the secondary datum and all dimensions come from that hole. There would be a 3rd datum or teriary that is used to orient (rotate) the part.

Dave D.

 

[ewh]

While it may be poor practice, I don't think the design intent was to designate one side of the cylinder as a datum, but was used as a shortcut rather than tying the datum to a diameter dimension. Dimensions are sometimes (again not preferred) placed such that the dimension lines are within the part boundary, which would be a similar case.

 

[draftsmandon]

I realize I am looking at an older book, but I rely on my drafting book from college which is Engineering Design Graphics, seventh ed., by James H. Earle. It states on page 358 that the "Cylindrical Datum Feature" is the axis of a true cylinder. In other words, two of the three required planes are represented as being the perpindicular centerlines intersecting at the axis of the shaft. This should be noted as a datum surface symbol in the circular view of the shaft. This practice may have changed with newer standards.

 

[hydromech]

As explained to me recently during a course on BS8888 GD&T, you can only use the following as datums...

Points, Edges, Axes' and Surfaces...

Any part of a surface of a cylinder would be classed as a profile and would not qualify as a datum.

The axis of the cylinder is the datum and the other cylinder on the weldment should be referenced to it. The form of the cylinders should be controlled by additional tolerances such as cylindricity and run out etc...

Hyd'

 

[ewh]

draftsmandon,
Time to get the actual standard. The datum may be the axis of a cylinder, but the datum feature is the cylinder itself. As far a locating the datum callout, that is more a matter of preference, as long as it does not violate Y14.5. It is just as correct to link the callout to a cylindrical dimension in a side view as it is to a circular dimension in an end view (of a cylinder).

 

[aardvarkdw]

The best way to think of this is, Where would you dimension the feature from? If you wouldn't attach the dimension to the centerline, you wouldn't attach the datum identifier to the centerline. The Centerline IS the datum but in order to make it a datum you must have a verifiable feature to measure from, hence attaching the datum indentifier to either the extension line of a dimension, a diameter callout or the edge of the feature.

Read section 4 of the ASME Y14.5M-1994 standard.

David

 

[cwdaniel]

Ok I think we're clear on the use of cylindrical features as datums. I'm more clear anyhow. My question is really more towards the proper drafting and interpretation of what is on this drawing. I think this drawing is basically poorly drafted and needs correcting, but is it worth it if the intent is clear enough?

Here's some crappy ascii art to show kind of what is happening on the drawing. The datum and the feature control are applied to surfaces of the cylinders. No dimension or axis is involved.

O------O top view
____
|//|A| |-A
|_ _|
| | | |
| |__| | side view
| |__| |
|_| |_|

 

[aardvarkdw]

There is nothing wrong with that. See figure 6-32 and 6-33 of ASME Y14.5-1994.

http://www.mooload.com/new/file.php?file=files/281106/1164732353/Doc1.doc

David

 

[cwdaniel]

Aardvardw, those examples are exactly how I would like to see this drawn. Right now it's not that way. The reference and control are on the outside surfaces. The ID of the cylinders is what needs to be controlled. The callout of the drawing does not do this in my opinion.

 

[aardvarkdw]

If the inside of your cylinders is what you want to control then yes, you want to change your drawing. That being said there is nothing fundamentally wrong with assigning a datum to the outside of the cylinder. It all depends on what is your critical feature. If you are concerned with the outside cylindrical feature being parallel with the other outside cylindrical feature then dimension it that way or vice versa. Either way is acceptable as long as it meets your specs.

David

 

[ewh]

aardvarkdw is correct on this. The drawing is correctly dimensioned as is IF THAT WERE YOUR REQT. If your require that parallelism be controlled to the ID, then the drawing is wrong, and the ID should be the datum.

 

[ringman]

Hydromech,

I am a little unclear as to the geometric definition of an edge. Where was this course you took, given?

Thanks

 

[hydromech]

Leamington Spa...Warwickshire...England.

Definition of an edge...intersecting point of two or more surfaces.

I aint saying it's right or wrong...it's just what i was told.

H

 

[ringman]

Hydromech,

Very interesting to say the least.