Perpedicular issues

[bboyd]

Can someone please explain how my perpendicularity is out. Measured top plane and cylinder and it is pependicular to .001 or less (great) measure bottom plane (part is only 1" thick) compare to same cylinder out .0045". Ok check parallel manually between the 2 faces with less then a .001 variation. What is going on here Origin is on the top face and center of cylinder.

[#2#]

Can someone please explain how my perpendicularity is out. Measured top plane and cylinder and it is pependicular to .001 or less (great) measure bottom plane (part is only 1" thick) compare to same cylinder out .0045". Ok check parallel manually between the 2 faces with less then a .001 variation. What is going on here Origin is on the top face and center of cylinder.

Workplanes? Isn't perp. dependent on the workplane?

[bboyd]

Not sure I understand what your saying here

[#2#]

Workplanes? Isn't perp. dependent on the workplane?

That statement is false. Perp. would have to be independent of workplane otherwise you would never get a true perp. from a cylinder to a plane.

Let me re-think.

[JoeRZ]

There's an explanation of Perp in the Wilcox website has something to do with projection distance. Check that out.

Joe

[#2#]

Ok, I have a question or two. Is your cylinder an ID or an OD? Are your planes measured similarly? By that I mean, is one composed of 40 hits over a 5 by 5 area and the other 3 hits over a 1 by 1 area?

Is this a machined part? Is the flatness of each plane very good?

[cmmguy]

Perp is "view" ie workplane and alignment(rotation) dependent if that has any bearing on the values you are getting.

[bboyd]

The cylinder is and ID measurement.
The planes are measured the same amount of hits over approx the same area.
It is a machined part with both faces and hole done at the same time.

Could one or 2 bad hits on the plane affect this that much?

[#2#]

Perp is "view" ie workplane and alignment(rotation) dependent if that has any bearing on the values you are getting.

How can it be workplane dependent? In order to check the perp. of a cylinder to a plane you would have to rotate the workplane through 360º around the cylinder to get a true 3d perp.

[#2#]

The cylinder is and ID measurement.
The planes are measured the same amount of hits over approx the same area.
It is a machined part with both faces and hole done at the same time.

Could one or 2 bad hits on the plane affect this that much?

That is what I was wondering. If your planes were only a few hits it is possible that the hits fell in "bad" areas. You said you checked the parallelism of the planes "manually". Does that mean on a plate with an indicator? Did you check the parallelism with the CMM?

[bboyd]

I did a quick check with a micrometer (not as accurate as an indicator but the area is not that large and I was looking for a quick confirmation) and the 2 planes are parallel to less then .001

[#2#]

I did a quick check with a micrometer (not as accurate as an indicator but the area is not that large and I was looking for a quick confirmation) and the 2 planes are parallel to less then .001

I would dimension the parallelism using the CMM - or check your perp. on the surface plate. See if you can get some numbers that agree.

[Jeffrey Lovely]

I think the order that you pick the features make a difference too

[#2#]

I think the order that you pick the features make a difference too

Ah, yes, feature first datum second. I think I have seen that in my travails.

[cmmguy]

How can it be workplane dependent? In order to check the perp. of a cylinder to a plane you would have to rotate the workplane through 360º around the cylinder to get a true 3d perp.

Perpendicularity of the cylinder to the plane is evaluated with axis compared to two perpendicular planes and IF and only IF the tolerance is preceded by the Circle symbol(Ø) then the zone is cylindrical shape. If the symbol is not shown then default is two planes. If symbol is missing then perp is "view" (and therefore alignment) dependent based on the view that the tolerance is depicted on the drawing.

Now, if the circle symbol is present, then ignore what I said earlier - I should have been more clear.