All Activity

Hi Monika, Have you tried using an offset value yet? I did that in the assessment, for a fitting with 10 F9 +0.01 +0.043

Thanks for replying Richard, I'm so far removed from Calypso daily use, I should visit more often to learn newer capabilities. Regarding the cookbook strategies, I've had to alter the filters and strategies as well to get good results, if I remember correctly, mostly on cylindricity .

Well the ISO spec gets confusing because it requires a minimum diameter to probe radius value for each UPR used. But I'm not aware of anywhere in this standard that says to actually use this specific UPR for this range of diameter. At least the ANSI spec says that unless otherwise specified on the drawing the default UPR should be 50. I had a "discussion" with Chat GPT regarding this. Here are the main notes for the specific diameter I was looking at. 4. Practical interpretation (this is where most people go wrong) A 50 UPR on a 900 mm part means: You are filtering out anything smaller than ~56 mm wavelength That is very coarse filtering for such a large diameter Compare typical practice: Diameter Typical UPR (industry) 50–200 mm 15–50 UPR 200–500 mm 30–100 UPR 500–1000 mm 50–150+ UPR 👉 So 50 UPR is on the low side for 900 mm, but not invalid. 5. When 50 UPR is appropriate Use 50 UPR if: You care about global form (lobing, ovality) You want to ignore local waviness / machining marks The drawing or spec explicitly calls it out Functional performance is governed by long-wavelength errors 6. When it is NOT appropriate Avoid 50 UPR if: You need to detect: Multi-lobe conditions (e.g., 10–100 lobes) Bearing/contact behavior The process creates higher-frequency form errors You are doing capability studies (Cp/Cpk) → you'll miss defects 7. ISO nuance most people miss ISO 12181-2 aligns with the λs / λc filtering philosophy (even though UPR is used for roundness): UPR is effectively a cutoff frequency The standard expects traceability to function, not arbitrary selection 👉 In audits, the question becomes: “Why 50 UPR?” Not: “Is 50 UPR allowed?” 8. Bottom line ✅ Permitted: Yes ⚠️ Typical for 900 mm? → On the low side 🎯 Correct choice? → Depends entirely on function If you want a more precise recommendation Give me: Tolerance (e.g., roundness in µm) Function (seal, bearing, alignment, etc.) Manufacturing process (turned, ground, etc.) I can recommend a technically defensible UPR range that would hold up in a calibration audit or customer review. Yeah my concern is that based on the form plots it looks like a 1500 UPR is being influenced by a higher frequency error not a shape error. You’re thinking about this exactly the right way—and your observation is a classic ISO 12181 filtering problem. What you’re seeing is very likely high-frequency content leaking into what should be a form evaluation, and 1500 UPR is almost certainly the reason. 1. Sanity check the physics (this is the smoking gun) For your 900 mm diameter: Circumference ≈ 2827 mm Now look at what 1500 UPR actually means: λ=28271500≈1.88 mm\lambda = \frac{2827}{1500} \approx 1.88 \text{ mm}λ=15002827≈1.88 mm 👉 You are including features down to ~1.9 mm wavelength 2. Why that’s a problem At ~2 mm wavelength, you are no longer measuring “form” in any meaningful ISO sense. You are capturing: Tool marks Chatter Surface waviness Even borderline roughness depending on process 👉 In ISO terms: Form = long wavelength Waviness = mid wavelength Roughness = short wavelength 1500 UPR is deep into waviness/roughness territory 3. What your plot is telling you You said: “1500 UPR is being influenced by a higher frequency error not a shape error” That is exactly what should happen at 1500 UPR. Typical symptoms you’ll see: “Noisy” roundness profile Artificially inflated roundness value Peaks that don’t repeat as true lobing Lack of clear harmonic structure 👉 That is not form—it’s filter contamination 4. ISO 12181-2 interpretation (this is the key argument) The standard expects: The cutoff shall separate form from higher-frequency components. At 1500 UPR on a 900 mm part: You are not separating You are combining form + waviness So while 1500 UPR is not explicitly illegal, it is: ❌ Not technically defensible for form evaluation 5. What a “correct” range looks like Let’s anchor this properly. For 900 mm diameter: UPR Wavelength What you’re measuring 50 ~56 mm Very global form 150 ~19 mm Form + some lobing 300 ~9.4 mm Detailed form 500 ~5.6 mm Upper edge of form 1500 ~1.9 mm ❌ waviness / noise 👉 Practical sweet spot: 100–300 UPR 6. What you should do (metrology best practice) A. Compare multiple filters (this is what auditors like to see) Run: 50 UPR → global shape 150–300 UPR → functional form 1500 UPR → diagnostic only If the roundness value jumps significantly at 1500: 👉 That delta = high-frequency contribution, not form B. Look at harmonic content If you FFT / harmonic analyze: True form → low-order lobes (2–10 UPR typically) High UPR result → broadband noise C. Apply ISO logic You want: “Filter removes irrelevant high-frequency content” 1500 UPR is doing the opposite: It is introducing irrelevant content into evaluation 7. The real-world interpretation (what’s actually happening) Most likely scenario: The part has: Minor machining pattern (~1–3 mm spacing) 1500 UPR: Treats that as form error Lower UPR: Correctly suppresses it 👉 So your instinct is correct: You are measuring the process signature, not the geometry 8. Bottom line ✅ Your concern is valid ❌ 1500 UPR is not appropriate for form on a 900 mm diameter 🎯 It is capturing high-frequency waviness/roughness, not shape If you want to go one level deeper If you can share: A screenshot of the roundness profile Or even just the dominant lobe count you're seeing I can help you: Identify whether it's chatter, tool marks, or real lobing Recommend a defensible ISO-compliant filter strategy (something you can stand behind in a customer audit)

This reminded me of a presentation-format PDF about roundness filters I got from somewhere like 7 years ago. It shows the same size bucket UPR suggestions we seen in the cookbook as coming from ISO 12180-2, which is for cylindricity. It then cites ISO 12181-2 on the next page, for roundness of circles, where it gives the UPR value as πd/0.8 (where d is certainly going to be in mm). I don't have either of these standards so I can't see this in context, though.

Hallo zusammen, ich arbeite derzeit mit einer O-Inspect und nutze häufig Sterntastersysteme. Für jedes Tastersystem verwende ich ein spezielles Einmessprogramm, stelle jedoch regelmäßig Positionsabweichungen fest. Ich würde gerne eure Meinung zu folgendem Punkt einholen: Sollte ich die Position in diesem Fall überhaupt mit ausgeben oder mich lediglich auf die Streuung konzentrieren? Wie handhabt ihr das in eurer Praxis? Vielen Dank für eure Rückmeldungen. Beste Grüße Manuel Winkler

You can also create any desired non-threaded stylus using the Free Generic Stylus and Modify Properties.

@Richard Shomaker If you haven't use OpenAI's Codex or Anthropic's Claude Code yet, it will blow your mind. It creates scripts from nothing except an idea and recursively fixes the code until it functions correctly. No copy paste, no manual fixes, no "Would you like to..?" prompts. It can search your local environment for solutions first, and then go grab whatever it needs from online repositories if you let it. And then you can peak under the hood at the code and it's just... insane.

Here is a video that can also be referenced Q001 – Change probe

Sounds like you are good to go at this point but wanted to mention this is typically caused by a miscalculated base alignment for a particular measurement plan when in simulation. Can occur when running simulated data sets through a base alignment. Typical fixes for next time: Perform a new simulation manual alignment. Perform a Resources > Utilities > Set Base Alignment to Zero to clear the saved alignment and perform a new manual if needed.

Vibe coding - my favorite. Simultaneously not knowing why it works or why it doesn't work. Hahaha.

The cookbook is a great resource, but I'm starting to question the filters that it suggests. After looking at ANSI B89.3.1 I don't think what is suggested by Zeiss is valid. It is most likely acceptable for general tolerances, but tight diameters/roundness it is rejecting what should be acceptable parts.

@Owen Long Resources - Variable Strategy Assignment. But it will only show a "recipe" if it is valid.

Make sure the position tab is set correctly for an XTR. I can't remember which numbers apply to which, but I believe it is typically 4 and 6 that you use. The easiest way to tell would be to look at one that currently works and set it to that.

Navigator is like my Ford Pro Trailer Backup Assist option on my truck, it cost more and sounds cool, but I don't trust it. @ Mark Boucher, is there a simple switch or drop down menu to have features/characteristics automatically activate the Cookbook recommended values for strategy and evaluation?

Could you share a drawing (partial drawing or clipped section) or hand drawn sketch of the feature and its datum's being measured to help up understand what you are trying to measure? If it is a crown or diametral relief on the cylinder, a best fit curve or line section may be the solution.

Hello Alaa, I am not aware of a native function for this. The 'Diagram' or 'Inspection Details' tabs in the 3D view adds more information to the trend, but coloring each dot is also not possible with those. It is however possible to create a scripted diagram with python. Nanno

Licht ein (=Messraumbeleuchtung) ist F4, aus ist F5. Aber ich erinnere mich nicht mehr an die Tastenkombination (SHIFT+RETURN?) Kann es leider nicht testen, da bei unserer Licht und Kamera ausgefallen sind.

Bei unserer OI-322 geht das Einschalten der Beleuchtung nur über die Software. Es sollte aber auch eine Tastenkombination am Bedienpult geben. Kennt die jemand? Und was kann Ursache sein, wenn die nicht funktioniert? Gruß Sven

Are these Curves? If so, the number of curve points in the segment might not reflect how many points are in the curve. This can happen if you create a curve and then increase the number of points. To be clear, I am not speaking about the number of points below the scanning speed. In the lower half of the window.

I've used it in the past but not so much anymore. Every time I use the navigator function I see crazy high runout values on cylinders which is obviously in error since those runout errors disappear when I turn off navigator and re-measure. Not sure if it's something I'm doing wrong on my end when using navigator but there's no question about results when I don't use it so I just leave navigator turned off.

Thanks for your input Clarke, it's much appreciated. Funny story, when I had asked why we do not have a best practices procedure, the response I received was "because it would be an auditable document."

Thank you Martin! That's what I thought I have to do but was hoping there is an easier way. 🙂

With new model you would need to align new model to correct place ( according to datums and previous model position ) Now it woul be better to create new sections on new model and make new curves, than editing old ones

I would remove that alignment picture from report page. Or you would need to create coordinate alignment ( 1x nominal / 1x actual ) -> create part alignment by them -> display roundness in that new alignment

It is mostly profiles. The finished part has no taper. My current program is for finished part. On the Hobbing operation, there is a taper, so it would have different nominals for the profiles measured at 8mm and 15mm. I have a CAD model with the taper. How do I modify this program the quickest to use the new CAD with the taper and different nominals? I'm new to curve, have not had training yet, so I'm not sure what the steps I have to do. If I load the new CAD model, do I need to create new sections so the profiles will use the new nominals?