Parallelism to cast surface designated as datum 1

[Dyskolos]

All,

I have a part that my company is manufacturing for a customer, and have a question regarding the GD&T that appears on the print. I have prepared a drawing showing a similar setup to the part we are making; please see attached. This drawing is not fully dimensioned; I have only included what seemed relevant to my questions; if I have left any relevant information out or can provide further information please let me know.

Part info:
The part is a casting that we machine to print
The datums marked A, B, C are noted on the drawing as "cast datum location surfaces"

Question:
The bottom surface of the part is machined, and calls for a flatness of .002 and a parallelism of .001 to datum A. It is the parallelism to datum A that confuses me. Since this is a cast surface and is likely to be somewhat uneven, I am unclear how to apply the parallelism of the machined surface to the cast surface. I have seen (on other parts)datum targets specified to show how the datum plane is to be established, but this part does not have these. Am I correct in thinking this should have datum targets, or is there some other principle I can apply to establish the datum plane on the cast surface, despite the surface irregularities?

Thank you for your consideration

 

[Dyskolos]

I appear to be having trouble uploading the attachment. Trying again here.

Edit to add: Attachment shows up for me now. I think I failed to click the "insert link into your post" button on my initial submission.

 

[CheckerHater]

It all depends on how bad the surface is. Datum targets may be good idea.
Unfortunately nothing is attached. Please make sure you didn't use symbol "&" in the file name.

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

 

[chez311]

As CH said datum targets may be a wise idea on any cast surfaces depending on the casting tolerances and flashing etc..

However I do not believe your datum C is a legal datum as it is shown - a datum target line(s)/point(s) would have to be utilized instead.

 

[pmarc]

Like others said, due to inherent irregularities of cast surfaces datum targets are usually used to simulate cast datums.

With regard to the parallelism tolerance itself, as shown now it is illegal because its value (.001) cannot be less than the flatness tolerance value (.002).

 

[Dyskolos]

To clarify, this is not my design, nor that of my company; we are a contract manufacturer, and thus have only a limited ability to effect drawing changes.

I don't have castings as yet to see what the surface will be like; I am in the preliminary stages of this project, but I am near-certain they will not possess flatness within .001", thus making parallelism of .001 to that face suspect.

So it appears from the comments that datum targets, while they are probably a good idea, are not strictly required to make this "legal". This being the case arguendo, how could I establish the plane of datum A to which I can check my parallelism? Since they did not specify targets, can I pick any old 3 points on the datum surface to establish a plane for datum A?

Pmarc,
Good point. Does it actually make it illegal, or does it just render the .002 flatness superfluous (since it would automatically be satisfied if the parallellism is met)?

Thanks for the comments; I very much appreciate it!

 

[pmarc]

Well, let's say that with the given tolerance values the two callouts create a conflict. So either parallelism of .001 is illegal (if we assume that the flatness callout reflects the design intent properly) or flatness of .002 is illegal (if we assume that the parallelism callout reflects the design intent properly).

 

[chez311]

As I said, I do not believe your tangent plane datum C is legal per Y14.5 - maybe someone else has a differing opinion. That needs to be changed to be established with datum targets, or utilize the entire surface (since it is 180 degrees and has opposing points) as a FOS datum feature.

I do not believe you may pick "any old 3 points" - if datum targets are specified then you must use those locations exactly as they are specified. If they are not (as in this case) then your datum simulator which establishes datum A is a flat planar surface only. Maybe someone else knows of an accepted alternate practice but I do not know of one.

 

[Dyskolos]

pmarc,
That makes sense. I will ask the customer to clarify.

Chez311,
I agree datum C is suspect; will ask the customer to clarify.
Regarding datum A: Ok, using a flat planar datum simulator (such as a surface plate) makes sense. Question: Hypothetically, if the surface ends up irregular in a certain way (such as convex, with the high point in the middle of the plane) it would be able to rest on a flat plane in more than one (or an infinite number) orientation, correct? What controls this in a) this drawing, or b) a correctly-made drawing, if the answer to a) is "nothing, because the drawing is wrong".

If I'm asking silly/basic questions I apologize; I am not formally trained in GD&T and may well have gotten cross-threaded at some point(s).

 

[Dyskolos]

I think I may have answered my own question regarding datum A possibly being convex; I would have to apply flat planar datum simulators to all three datum features specified on the drawing (may end up somewhat different depending on how the customer wants me to handle datum C, but I think I've got it).

Thanks all for your help!

Edit: Slight changes made for clarity.

 

[chez311]

Dyskolos,

Please refer to paragraph 4.11.2 of the Y14.5-2009 standard regarding stabilization of datum features. It will direct you to the math standard Y14.5.1 - I am unfortunately not as well versed with this standard as I would like so I wouldn't be able to provide you much direction past that. Again not to sound like a broken record but if you are concerned about rocking or significant irregularities/variation your best bet is to have datum targets specified on a surface that you expect to have issues with significant irregularities like a cast surface.

 

[drawoh]

As I said, I do not believe your tangent plane datum C is legal per Y14.5 - maybe someone else has a differing opinion. That needs to be changed to be established with datum targets, or utilize the entire surface (since it is 180 degrees and has opposing points) as a FOS datum feature.

The datum[ ]C feature on the OP's drawing is weird, but it is fixturable as per the definition of datums in ASME Y14.5. The primary datum feature is a face picked up by three points. The secondary datum feature is the side, picked up by two points, and the tertiary datum is picked up by one point, presumably the quadrant of the radius. It is a much better datum feature than the centre of the radius.

A datum target on the radius quadrant would be more proper, but as shown, probably it shows design intent. We are dimensioning from and end that happens to be round, sort of. The radius is not particularly accurate or critical.

--
JHG

 

[chez311]

drawoh,

What it seems to me is that it is almost specifying a datum target without actually utilizing a datum target callout. I've also seen people try to specify a tangent plane datum like this before (utilizing the actual "tangent plane" symbol) but its been defined as non-compliant by several experts on this forum on several occasions. It doesn't seem to me that as shown, it is supported by the standard - can I understand what is being communicated? Yes, but I was just trying to say I don't believe its actually strictly allowed.

Can you point to where in Y14.5 this datum, as drawn, is supported?

 

[drawoh]

chez311,

I cannot see anything like the OP's drawing. Look at Fig-4.35 for tangential attachment to datum features. Remember, this is a tertiary datum.

--
JHG

 

[chez311]

drawoh,

Figure 4-35 applies to irregular FOS - I don't believe the tangent datum plane C as shown constitutes an irregular FOS. The circular section might, since it is a full 180deg and has opposed points, however not the plane tangent to it as shown.

I don't mean to nitpick - as I said I'm just saying that strictly going by the letter of the law, I don't think the datum as drawn follows the definitions provided in Y14.5, but it could easily be made compliant if it were modified to be a datum target. Thats not to say one couldn't easily design a fixture to hold, gauge, and measure this part per the drawing.

 

[pylfrm]

Dyskolos,

If datum feature C is referenced as tertiary with A as primary and B as secondary, I agree with the chez311 that the intent is clear, but that a datum target would be the proper way to specify it. Your drawing shows datum feature C referenced as primary though, which doesn't make much sense either way.

I think I may have answered my own question regarding datum A possibly being convex; I would have to apply flat planar datum simulators to all three datum features specified on the drawing (may end up somewhat different depending on how the customer wants me to handle datum C, but I think I've got it).

If the tolerance only references datum feature A, then datum features B and C are not involved at all.

You didn't state what drawing interpretation standard applies. If it is indeed ASME Y14.5-2009, which refers you to ASME Y14.5.1M-1994, then you can find a relevant excerpt in thread1103-441860. For the parallelism tolerance, you can choose whichever candidate datum minimizes the error. A convex datum feature would actually the tolerance easier to comply with, although more complicated to inspect.


pylfrm

 

[chez311]

pylfrm,

Agreed - the position callouts do not look correct, I'm not sure exactly what the intent is but whatever it might be is probably not being accomplished. The first FCF only controls perpendicularity of the axis and the second FCF only controls position in one direction relative to C. Position/translation relative to B is not controlled.

 

[Dyskolos]

Pylfrm,
-Looking back, I am confused my own self what I was thinking regarding the interaction of the datums. Please write that off as nonsense.
-Regarding C as primary datum; what about the drawing marks it as such?
-The original drawing carries a note to interpret the drawing in accordance with ASME Y14.100-2004, as well as a separate note/chart stating that various "Geometric Characteristic Symbols" are as per ANSI Y14.5-1973. Perhaps this is a source of confusion? I am not familiar with the changes made to Y14.5 over the years, and this is the first time I have seen a reference to Y14.100.
-Choosing from among candidate datums to minimize the error makes sense. I had considered this, but got hung up on the fact that I won't be able to make sure my customer chooses the same candidate datum. That said, I believe I can document my process sufficiently to cover us.
-The other thread you provided is very helpful, thank you much!

 

[chez311]

Dyskolos,

Your position callout is sort of a disjointed attempt at multiple single segment. You have two separate Feature Control Frames (FCF) each with a single (primary) datum, the first of which is to datum D and the second is to datum C. The first datum in any Datum Reference Frame (DRF) in a FCF is always considered primary.

In regards to the position callout itself, as I said it does not constrain translation in one direction as datum B is not in either FCF. Making some assumptions in what is trying to be achieved (ie: position relative to B can be held to the same size zone and does not need to be tighter/looser than relative to C) a better callout might be the below. You are also missing diameter symbols in each FCF next to the tolerance value to denote the zone is cylindrical as it is applied to an axis.

Perpendicularity | ∅.003 (M) | [D]
Position | ∅.015 (M) | [D|B|C]

Edit: sorry I realize you were referencing the machined face D instead of the cast face A in your original callout, I have changed my suggestion to reflect that.