Capability & GDT Features - Surface Profiles, Line Profiles & Symmetries

[Peter A]

Ok, I have read a few threads on the forumn but feel I need my own thread.
We have been asked to measure Surface Profiles, Line profiles & symmetries on injection moulded components. All have a unilateral tolerance ie 0-0.1mm.
How do we correctly upply capability to these? Currently we are using MiniTab to calculate a PPU value, only having an upper boundry. Our customer has queried this thinking we should be using 0 +/-0.05mm.
Any views on this?

[Stijloor]

In Reply to Parent Post by Peter A

Ok, I have read a few threads on the forumn but feel I need my own thread.
We have been asked to measure Surface Profiles, Line profiles & symmetries on injection moulded components. All have a unilateral tolerance ie 0-0.1mm.
How do we correctly upply capability to these? Currently we are using MiniTab to calculate a PPU value, only having an upper boundry. Our customer has queried this thinking we should be using 0 +/-0.05mm.
Any views on this?

Welcome to The Cove Forums!

Do you have a sketch that shows the actual callouts? The default for profile of a line and profile of a surface is that the tolerance zone is equally disposed about the true profile. But without an actual drawing, I can not tell.

[bobdoering]

It would be nice to see some data, but my guess is your process variation distribution is a beta or Weibull distribution, non-normal - unless you are very far away from 0. Using Ppu alone is incorrect. It has a totally different meaning than Ppk. You most likely will need to transform the data to get an accurate understanding of your capability with "centering" - Ppk.

[Paul F. Jackson]

Profile is a tolerance with max and min boundaries that the surface must reside within. Those boundaries may be oriented and/or located to datum references or free to float.

No matter whether the limits are specified equal bilateral, unequal bilateral, or unilateral relative to the basic they can be treated as equal bilateral by figuring the offset for contour boundaries that have a specified basic displacement from datum references.

Since profile specifications may often be used to control non-generic forms... each measured point can be examined for its variability within the bilateral zone to an offset target mean (if necessary).

When a profile tolerance is applied to co-planar (or other) surfaces without datum references... then it is treated similar to flatness where one boundary "ZERO" is only approachable but never attainable... whether the surface is curved, flat, or other... as long as there are no opposed surfaces in the contour tolerance zone (eliciting a size situation) this would be a "one-sided" tolerance (zero to max).

If the profile zone is linked to a datum reference with a basic or is not linked with a basic but the contour wraps around any contour where there are opposed points it then it is not a "one-sided" tolerance.

Paul