Jump to content
You are not logged in. Please click here to Sign In ×
Software Release 2026 of ZEISS Inspect & ZEISS PiWeb

Evaluation method selection

---

By [Zo...] ,
in General

 Share

Recommended Posts

---

[Zo...]

Posted
Posted
Not having much experience with Calypso I was amazed to see how different the results are when applying Outer Tangential Element vs. LSQ. After talking to one ZEISS tech he suggested to always use Outer Tang El setting ( for the most part...) but a different tech just yesterday said that I should not do that unless I am making a part to fit something. Well, I am supposed to make the part to the print, ( I create my CAD model, program the CNC mill, set up and run the part and check them on the CMM - sometimes just for internal use, sometimes to get a report for the customer) and don't always know if there is a matting part or what type of fit is intended.
Are there rules of thumb that one can learn when to use which setting for evaluation? I find it confusing that the readings are inconsistent, for example, a R 0.06" +-0.002 corner radius that has no matting part measured multiple times and it varied within 0.002-0.0035. S changing the method would give different results, but how do I know the absolute value? I don't just want a report that the customer accepts. I want to know the ACTUAL size, if that is a real thing in the first place.
I went to basic training, we touched on evaluation but not in depth. Apparently it is more important than that. Is there a way to learn more about that ( can't seem to find it in the book...) ? Tutorials focused on evaluation methods and perhaps constraints?

Thanks
Link to comment
https://qualityforum.zeiss.com/topic/14952-evaluation-method-selection/
Share on other sites
---

[Cl...]

Posted
Posted
150_820c89eab18dbbb5328609830f28e9d6.pdf
This PDF is directly from Zeiss, and should help you understand
what the appropriate evaluation method should be.

Regarding that radius; You will always have a problem with radii and circles that have a <180° surface to measure. This is were constraints come in. If it's say, in the XY work plane, click on "Evaluation" on the feature window and tick (constrain) the X & Y location. That should give you a more accurate size. Conversely, if you constrain size (Radius) you'll get a more accurate location.
---

[No...]

Posted
Posted

Please sign in to view this quote.

Let me ask you a rather philosophical question:
What is the "actual size" of a feature that deviates in a more or less random way from its nominal geometry?
What if that radius is not even a circle? How could anyone tell you the actual size of it then? How do you define that?
So, measuring results are always just an approximation of the real geometry. We TELL Calypso to measure a radius, even if it isn't one. The fitting algorithms are just a way to return an approximated result of the kind we asked for (usually a regular geometric element), so we don't have to deal with the "something" that was really measured.
---

[Me...]

Posted
Posted
Clarks PDF outlines the reason for LSQ. Zeiss's Cookbook also suggests LSQ for the measurement of Features. The explanation from my Training & Zeiss would be because it is more stable and repeatable for measuring FEATURES. When it comes to evaluating, that is recommended to be applied through Characteristics.
---

[Cl...]

Posted
Posted

Please sign in to view this quote.

Absolutely, a sound a sound strategy. This is part of the "Best Practices" document for creating a Calypso measurement plan
that I created for our programmers to follow. There are a set of standard protocols to follow in the creation of a program, that ensure uniformity and consistency (among other things).
---

[Cl...]

Posted
Posted
I'm sorry, I can't share the actual document. I can tell you what's in it.
I think most of it is just common sense stuff.

In no particular order;

1] Always use a start Alignment w/instructions in the feature comments.
2] Create loading instructions with a picture of the part as it is in the fixture. This is available
under User information. (We use autorun on all our machines).
3] Run from the features side.
4] Group all features (Start Alignment/Base Alignment/All Features/All constructions.
This keeps everything separate and organized
5] UOS, use the compact printout (Characteristics only) A very simple report showing
among other things XYZ coordinates for the machine operators to make moves.
5] All reports to be saved (using a simple PCM line of code) on the network drive under
- Resources - Name for output files - for this measurement plan.
6] Set evaluations from the characteristic side.
7] Program naming convention is; 13-XXXXX OP-10_Rev_A
8] There are a few PCM related things that are also included.
9] Characteristics are simply numbered as #13, #14, etc.
10] Always heal the model (Some may disagree with this one)
11] Naming convention for stylus should include A rotation - B rotation - direction of styli - Size of styli E.g., 90_-90_X+_1.0
12] (For autorun programs; after establishing the Start Alignment, Run form Base Alignment under "Selection".

There are a few more, but I can't remember what they are.

I'm open to adding things if anyone has their own?
  • Like! 1
---

[Zo...]

Posted
  • Author
Posted

Please sign in to view this quote.

Correct me if I'm wrong, but isn't the measuring process absolute ? Whereas the evaluation is "chosen" by the user, based on the interpretation of the function or the intended use ? And that would make the evaluation relative. Therefore there is a difference.
 Share
Go to topic listing
×
×
  • Create New...