Help on positional inspection method

[QE Mike]

I have two identical holes on center line drilled on opposite ends of a part. One hole is datum A Ø.189/.195 and the other is the feature under consideration, also Ø.189/.195 with a position requirement of φ|Ø.002(M)|A(M)|. We are trying to verify conformance by using the tightest fitting gage pins in each hole, placing the pin in Datum A in the rollers of a concentricity gage with the indicator tip on the pin placed inserted into the feature under consideration. Using this method, I know we can add the bonus tolerance from the feature calculated by (Øgagepin-.189) to the positional tolerance of .002 and get up to .008. Can we also add the datum shift as well calculated by (Øgagepin[A]-.189) for a potential total positional tolerance of .014? My instinct says NO! But this has sparked a healthy debate and I wanted to get some feedback. We would prefer to check this with a functional gage or CMM, but that is just not in the cards right now.

 

[Belanger]

Yes, if I'm understanding the situation correctly then you can add in that datum shift. Usually datum shift isn't added directly in, so a purist might give me some flak for my answer. But yours is a special case of one feature, to one datum, and where the feature and the datum act in the same direction, inline with each other. So there's a direct relationship and they essentially stack together.

Think of it this way: If the datum were tagged to the other hole, and the position tolerance tagged to the opposite hole, the results would be the same. So bonus and shift in your specific example are kind of the same thing, but coming from different ends of the part.


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

 

[mkcski]

I agree with John-Pauls' assessment. These coxial relationships are unique when dealing with the compounding affect of bonus and shift. There are many good examples in GDT books. Suggest section 7.6 page 148 in the 2009 version of Y14.5 for a reminder.

Certified Sr. GD&T Professional

 

[QE Mike]

Thank you both very much. I would love to see those examples in the 2009 version, but I only have the 1994 version. Maybe I can get the company to splurge on the new one.

 

[mkcski]

You are very welcome. Yes, a few shelves full of GDT reference material is worth it. I buy everything I can get my hands on. I recommend buying the 2009 version at a minimum - around $195 ????. Tec Ease, ETI, James D. Meadows and many others offer books at reasonable pricing.

Certified Sr. GD&T Professional

 

[Belanger]

QE Mike -- in the 1994 standard, check out Fig. 5-50 on page 146. That's pretty much what you are dealing with.

Notice the table for that figure has "bonus" tolerance going downward, and the datum "shift" builds as the numbers go to the right. That shows that they essentially add together (again, for this special case where there is one feature, to one datum, and where the feature and the datum act in the same direction, inline with each other.)

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

 

[QE Mike]

I saw that table in the 1994 version, but for some reason, I have two big name customers with very similar parts, one aerospace, one medical both saying they don't allow us to add the datum shift when using the concentricity gage, only when checking on a CMM. I wish I knew what they were thinking.

 

[powerhound]

Typically datum feature shift is not added to any bonus tolerance. This is a special case though where it actually does work out that way. Your customer may not be considering this as a special case.

John Acosta, GDTP Senior Level
Manufacturing Engineering Tech

 

[pylfrm]

I'll present the expected flak: The previous answers consider only translational datum feature shift. They ignore rotational datum feature shift, which could be quite significant depending on you part geometry.

Imagine a part with total length 11.000 and diameter .194 holes in each end to a depth of 1.000. This would allow a rotational datum shift of about (0.005 / 1.000) = 5 mrad. If the axes of the holes were perfectly parallel, they could be offset by about .051 and still meet the requirement.

Make sure there aren't any simultaneous requirements preventing you from using all of the potential datum feature shift for this one feature.


pylfrm