Name X

Manipulating the secret settings - sounds like fun  ;D

I will test it asap, thx for the hint!

'Cut Holes with Curves' could be a function you can use. Just create a circle from your cylinder in IMInspect, edit it numerically to 1 and import it into your IMEdit project. After closing both sides of the model surface you can use the curve to create a watertight cylinder.

The other option would be to import the numerically edited cylinder as NURBS, create a polygonal model from it and integrate it into the scanned model. 

??? no other idea...

No other chance than doing it with the 'normal' fitting stuff.

In cases like you described I sometimes start to create all my nominal on the stl just by picking (e.g. a circle with three points), starting with the 'perfect' stl as data object. So I get all my nominal primitves in a fast - not accurate way, but can use the autoextraction to create measured ones. Now the trick: Delete all nominals, switch the measured to nominal primitives and put the previous data stl as reference.  ;D

In IMEdit I would try it with surface patches...

Do you have a CAD software and the NURBS modul in PolyWorks? Because then you can create two NURBS patches on the left and right side of the blade, export them and create a intersection in the external software. Then import the new CAD, and merge the original and the new model. After that you have a perfect sharp edge and you can use the rolling ball to get a nice radius.

As far as I know the WinRDS is not working on 64bit, but Hexagon is working on it.

I think it is problem concerning the compensation...

Please check the displayed tip-diameter, perhaps start a new calibration anyway.
And if you use the orientation of the probe switch to 'compensation point' for compensating.

One sided tolerances can only be used with features. No solution for the report of the cp, but for displaying it I use the Auto-Range Color-Scale and define a one sided range...

If you use nominal and measured ones I think you can only use the angle feature between the two planes.

For using only measured ones the quick way is to define nominal and measured inside one feature with pick&fit so you can use the 'normal' feature controls.

Three ideas:

GD&T circularity
- create a plane at the height you want to measure (numerically, enter the z coordinate (x+y=0, normal z=1, x+y=0).
- crop the plane so it covers the outer area of the scan
- select data points to primitive (plane) without using an angle
- fit a circle into the points selected (without an angle)
- activate the circularity in feature controls/GD&T

Idea two and three are just methods to get the points:

I think thats what I would choose to get the points: creating calipers all around with the outside tip anchored. A fast way to get the calipers would be to create one (e.g. numerically) then duplicate it and rotate it around z (if you want I can write a macro for duplicate/rotate it every 0.1 degree or something like that).

The last one: Profile gauges - in V11 you can now create PG also radial, perhaps you can get the points like this.

No reason to apologize ;)

Normally I also use cross-sections on scanned pointclouds. But sometimes accuracy is not enough with the scanner, so probing is the only way...

Hi PW-Users,

using the 'probing cross-section' I got two problems:

1: during probing you miss a area, going back to the part the cs follows the sequence of points, so the result is a weired cs
2: probing the cs you have no chance to get the original points, e.g. for changing the projection distance onto the cutting plane or to make a alignment using the cs points (as data points)

My solution 1:
- create a polyline from the probed cs
- export the points of the line to a text file
- import the text file as data
- extract the data normals
- update the cs

Using a macro for it you can probe additional areas.

Perhaps someone with another idea for problem 2? Because now I use the points projected onto the cs plane, it would be perfect to use the probed points, compensated but not projected...

Have you tried to change the properties of the scan (interpolation step, max. edge length)? Perhaps the parameters were reset to standard...

The Best Fit, Min or Max are using exactly the same points, so the GD&T measurement is also the same. Best-Fit with reject outliers turned on reduces the points - so you get a 'better' result  ;D

I would suggest to take a look at the smoothness of the scan data - if there are no outliers (from reflections or something) I would not use it, if the data is very rough I would turn it on. And I would turn it on if the normal result is out if tolerance  ;D

It is located in the management of the coordinate systems (located in the alignment menu), there you can choose if you want to display the 'world' coordinate system (and all the other ones you have created). Don't have my dongle right now, will look for the exact path tomorrow...

Align>Coordinate System>Manage

I agree with GoBeavs, the only way to get the correct measurements would be GD&T, cause normally you need the Tschebicheff fitting methods (not sure if I got the name right...). If you use GD&T in PolyWorks the original points for creating the plane are used for a complete refitting of the parallel min/max planes.

Do you need another idea to get the right points for starting GD&T?