Re: little bit off topic (about inspection)

From: Bert van den Berg (
Date: Fri Mar 03 2000 - 14:55:41 EET

Steve Pitt wrote:

> For inspection of freeform surface, what should be inspected?

This depends on what you need, and the resources available. Consider
a car fender. Do we need to determine if the fender will:
- fit onto its mounting points ?
- blend smoothly with its neighbour ?
- show a (large or un-even) gap to its neighbour ?
- have any undesirable visual marks or features ?

To answer some of these questions you might:
- measure some points on some slices through the fender, and
assume that the rest must be OK (low reliability)
- measure points where the sculptured features interface to mounting
features to ensure that the shapes blend smoothly,
- try to measure exactly along part edges to make sure the shape will
blend with the parts it is mounted to (difficult)
- scan the part so that you have possibly millions of points, and
have likely captured all the shape information of the surface
(most thorough).

> There exist a lot of inspection features such as plane, cylinder, etc.
> In that case, sampling several points is enough.

Not really. ISO standards for measurement call for Max/Min material
sampling. This is ~equivalent to functional gauge testing. To approximate
the same thing with a probe you need _many_ points. The problem is
that the sample points have to capture the maximum / minimum points
on the feature surface, rather than an average (e.g. least squares) estimate
(as done with most feature fitting software), and you don't know which
points to measure to capture the Max/Min points.

> But I think that freeform surfaces are different from the features.
> Just is it enough to see the difference between point data and the original
> surface?

What is needed is information that defines the translational, rotational, size,

and form differences between actual features, and their design specification
(in general).

> Or the surface which is reconstructed from point data must be compared with
> the original one?

This is as challenging (i.e. ambiguous) problem. Constructing a surface from
the measured points might allow:
- the "size" of the sculptured surface to be specified and analyzed,
- visual features of the measured part to be better evaluated
- features synthesized from the surface (e.g. feature curves). to be better
Depending on the function of the surface, an LS or MMC surface
construction might be required.

Bert van den Berg                 
Applications Engineer        
Hymarc Ltd.                 
35 Antares Drive                        Tel: (613) 727-1584
Ottawa, Canada, K2E 8B1                 Fax: (613) 727-0441

For more information about the rp-ml, see

This archive was generated by hypermail 2.1.2 : Tue Jun 05 2001 - 23:02:57 EEST