True Position without tertiary datum? Establishing a proper rotation

[basselope]

I know I've seen this before, but I'm drawing a blank. My task is to layout pieces using CMM, Comparitor and basic hand instruments. There is no dedicated gaging available.
In the attached pic is a part with Datum A being the material surface, Datum B being an outside radius, and a third feature (a similar outside radius) having a .010 positional callout back to datums A and B. Given that the two radii are only connected through two Basic dimensions intersecting at a virtual point, how does one go about establishing a proper rotation in order to evaluate the positional callout?

My gut feeling (beyond this being a poor way to dimension this part form an inspection point of view) is that the best one can do is trig out the angle between the virtual "Y" axis (the 9.17 dimension) and the direct feature-to-feature line between the two radii and then rotate my part coordinate system accordingly (in this case 2.2 degrees.)
In other words, establish a three axis coordinate system to the three Datums shown in the drawing, and then rotate 2.2 degrees to create the .350 basic distance, thereby losing any variability the positional tolerance might allow along that axis.

While moderately effective, it still irritates me that I am partially checking the position of the feature to itself.
Any ideas?

 

[Jim Wynne]

Re: True Position without tertiary datum?

I know I've seen this before, but I'm drawing a blank. My task is to layout pieces using CMM, Comparitor and basic hand instruments. There is no dedicated gaging available.
In the attached pic is a part with Datum A being the material surface, Datum B being an outside radius, and a third feature (a similar outside radius) having a .010 positional callout back to datums A and B. Given that the two radii are only connected through two Basic dimensions intersecting at a virtual point, how does one go about establishing a proper rotation in order to evaluate the positional callout?

My gut feeling (beyond this being a poor way to dimension this part form an inspection point of view) is that the best one can do is trig out the angle between the virtual "Y" axis (the 9.17 dimension) and the direct feature-to-feature line between the two radii and then rotate my part coordinate system accordingly (in this case 2.2 degrees.)
In other words, establish a three axis coordinate system to the three Datums shown in the drawing, and then rotate 2.2 degrees to create the .350 basic distance, thereby losing any variability the positional tolerance might allow along that axis.

While moderately effective, it still irritates me that I am partially checking the position of the feature to itself.
Any ideas?

You're correct--that dimensioning is seriously messed up. Because of the circular (i.e., datums that refer back to themselves in feature control frames) datum references--a sure indication that the designer and/or CAD person is clueless--there's no way to tell what the designer intended. There's no easy way to tell a customer that his drawing is significantly wrong, but you need to explain your plight and ask for guidance, and as I always recommend in situations like these--get it in writing.

 

[basselope]

Re: True Position without tertiary datum?

Thank you Jim. I knew somebody had their head up their backside, I just wasn't positive whether or not it was me!

I feel better now.

 

[Jim Wynne]

Re: True Position without tertiary datum?

Thank you Jim. I knew somebody had their head up their backside, I just wasn't positive whether or not it was me!

I feel better now.

I occasionally remove my own head from that allegedly anatomically impossible position just long enough to make sense. Maybe we need a new smiley here.

 

[basselope]

Hehehehe...
Talk about circular references!
Explain to me how a smilie get it's head up it's backside?

(though I'm sure someone has already figured it out!)

 

[Miner]

I know I've seen this before, but I'm drawing a blank. My task is to layout pieces using CMM, Comparitor and basic hand instruments. There is no dedicated gaging available.
In the attached pic is a part with Datum A being the material surface, Datum B being an outside radius, and a third feature (a similar outside radius) having a .010 positional callout back to datums A and B. Given that the two radii are only connected through two Basic dimensions intersecting at a virtual point, how does one go about establishing a proper rotation in order to evaluate the positional callout?

My gut feeling (beyond this being a poor way to dimension this part form an inspection point of view) is that the best one can do is trig out the angle between the virtual "Y" axis (the 9.17 dimension) and the direct feature-to-feature line between the two radii and then rotate my part coordinate system accordingly (in this case 2.2 degrees.)
In other words, establish a three axis coordinate system to the three Datums shown in the drawing, and then rotate 2.2 degrees to create the .350 basic distance, thereby losing any variability the positional tolerance might allow along that axis.

While moderately effective, it still irritates me that I am partially checking the position of the feature to itself.
Any ideas?

Jim is correct about checking with the customer. My guess is that they meant to specify the center to center distance between B and C. You can do this by omitting the diameter symbol in the callout, but most GDT semi-experts don't realize this. By omitting the diameter symbol, rotation does not matter since you are measuring a radial distance. In any case, verify the customer's intent in writing.

 

[basselope]

Yea, I've already contacted the ship-to customer. His sentiment mirrors the rest of us here.
But as their company is global and "Engineering" only exists as some distant, mythical group like the Druids, getting an official interpretation is a very time consuming task. Time is something the customer and myself simply don't have much of on this one.
We've both decided to get the opinions of the smartest folks we know (yes, that would be you folks) and go from there.:

I feel pretty comfortable with the information I have now, and I'm not afraid to stand by my results. Thanks again, Y'all!

 

[Jeff Mathena]

Long time lurker, but I haven't posted very much. Thought I might add my $0.02 on this subject, since I have an abnormal interest in GD&T.

The symbology shown in your drawing for denoting Datums is not supported in ASME Y14.5M 1994, but that from the ANSI Y14.5M 1982 standard. Since I don't have a '82 standard anymore I can only go off the '94 standards and assume the drafter just hasn't converted their symbology over to the newer style.

Position callouts with 1 and or 2 elements in the datum reference frames are allowed by the standard. If you only have 1 element in the datum reference frame, such as -A- in your drawing, the only thing that position can control is perpendicularity to -A-. With the position callout you are questioning, the only control the datum reference frame can control is the perpendicularity of the feature's centerline to -A-, and since -B- is only an axis, it can only control a radial distance away from -B-. The fact that the diameter symbol is still shown can be confusing and is shown in some examples in the standard. But since the callout is controlling perpendicularity to -A-, the diameter symbol seems to make since in this instance.

The assumption made to trig out a coordinate system from the 0.625 DIA hole would not be correct, as it would then be creating a coordinate system that isn't supported by the given datums.

Just my input. I'm happy to discuss my points if any of them seem out in the weeds.

Thanks,

Jeff

 

[basselope]

Yea, I agree with you that creating a three axis coordinate system is not proper to the callout. It sounds to me like we both agree that this drawing needs a revision.

As far as your abnormal interest... I was always under the impression that "Abnormal" was a prerequisite in this business.

Thanks for your input. I always like to other points of view.

 

[True Position]

The drawing is bad, but generally if I see a drawing like that that I'm laying out on the CMM the general method:
Align first to the A datum surface
Measure datum B, assume it is correct.
Using a special rotated alignment, rotate the part to bring the X or Y axis to the correct length. Measure the error in the other direction as your true position error.

In your example, you'd pick up the top surface.
Pick up hole B. (In this same I'm assuming you're making B the origin)
Rotate the part until hole C in Y is 9.170.
See what this makes C in X. Any change from -.350 is the TP error.