GD&T Question

If I level F as the "primary" and then check G runout to F, if F is crooked, when it projects over to where G is, the centerline of the actual F will be no where near the centerline of G.
Same going the other way. Even though the X and Y at the mouth of both F and G holes might be identical, the axis being crooked, projected way over to the other hole will give a really awful runout.

If I am putting something in F and using your part as a connector that then holds G, that's what I'd want. Rod in-line to F, your part connecting in-line to G, and the rod in G being straight out the other end.



If I have a solid part with a clevis sticking up, and pins in the ears of that clevis... Your part sits in the clevis and the pins insert from the two sides.
Then, What I care about, is that when I put the two pins in, neither side is so crooked that it binds.

So, cylinder F-G thought of as a single thing, and check the runout of each half to that compound feature.

If either side runs out (I personally don't like runout, but that's me) too much, it will bind, wear excessively or restrict motion and not function.



Neil's solution is built in to PcDmis and more elegant.
Contructing a cylinder from the points measured in F and G individually, then checking is bulkier.
They should net the same answer.

I tend to do the second way, constructing, because I'm old, PcDmis didn't used to have the thing to do that build in, and I don't remember the 1 time every two years I have to program something for this, that's 100% on me. I'd like to remember to do it as Neil said.

If I level F as the "primary" and then check G runout to F, if F is crooked, when it projects over to where G is, the centerline of the actual F will be no where near the centerline of G.
Same going the other way. Even though the X and Y at the mouth of both F and G holes might be identical, the axis being crooked, projected way over to the other hole will give a really awful runout.

If I am putting something in F and using your part as a connector that then holds G, that's what I'd want. Rod in-line to F, your part connecting in-line to G, and the rod in G being straight out the other end.



If I have a solid part with a clevis sticking up, and pins in the ears of that clevis... Your part sits in the clevis and the pins insert from the two sides.
Then, What I care about, is that when I put the two pins in, neither side is so crooked that it binds.

So, cylinder F-G thought of as a single thing, and check the runout of each half to that compound feature.

If either side runs out (I personally don't like runout, but that's me) too much, it will bind, wear excessively or restrict motion and not function.



Neil's solution is built in to PcDmis and more elegant.
Contructing a cylinder from the points measured in F and G individually, then checking is bulkier.
They should net the same answer.

I tend to do the second way, constructing, because I'm old, PcDmis didn't used to have the thing to do that build in, and I don't remember the 1 time every two years I have to program something for this, that's 100% on me. I'd like to remember to do it as Neil said.