Cylinders Issue

[RussL]

I have culinders in a program, they all inspect fine when viewing part on machine (actual hits all look good).
I take four hits at two levels.
On screen, vector of cylinder after inspection is 90 degrees out (centerline shows parallel to surface, cylinder is normal to.
I change cylinder to check three hits at three levels.
Cylinders check fine.
Centerline is correct, true location within .010
Has anyone else seen this, is this a known problem.
PC D-MIS V3.7 MR2, Windows XP.

[\v/inston]

i remember being told by the instructor at the training course to always use atleast three levels when measuring a cylinder.

[Matthew D. Hoedeman]

Been a problem for a LONG time. You need a minimum of 3 levels of this to keep the cylinders on the right axis.

[RussL]

Cylinders Issue

I have a hole going through a corner of a part.
Both top and bottom surfaces hole goes through are not normal to hole.
I make a cylinder and the vector always ends up normal to the surface.,
not along the centerline of the hole.
I had to go into Catia, get the actual hole vector and update the cylinder in PC D-MIS.
Is this a known problem? perhaps there is a setting I need to change?

(This was supposed to be a separate post)

[56flh]

If I take hits at four levels, PC-DMIS seems to always create two cylinders. Not much experience here, so I'm probably missing something obvious.

[tking]

Thanks folks. I've been having fun with cylinders too. 3 or 4 points at 2 levels. I'll try the 3 level trick. No training here, just hard knocks.

T. King

[BKulpa]

8 is the worst number of points for cylinder

8 is the worst number of points to use on a cylinder. If the distance between levels is close to the diameter, the algorithm can construct the vector in 6 possible directions. And I agree that 3 levels definitely works better.