All Activity

Our programming team, just two of us, is certainly in a position where programming/debugging at the CMM is counterproductive with the volume of inspections. Top that off with no dedicated CMM operators to help test programs has had some disastrous results. I've quickly discovered some of the quirks with Planner and Simulation already. We'll see just how much more finnicky it can get!

Without the FreeForm license, here is an article on how to get a profile result... https://portal.zeiss.com/knowledge-base?id=2495545

I think you can create profile callout on every feature, then report Max ( or in ASME Max and Min ) from those characteristics. I have no exp without FF so i don't know how others are dealing with this.

This is probably pretty basic stuff, but how do I take a profile measurement of a composite of multiple features? I'm using Calypso 2025, and do not have FreeForm. Attached a snippet of the feature I'm trying to measure. Need a profile of the entire perimeter, despite what the drawing may suggest:

I am trying to recall 3 curves..... with all the errors I was getting I did the following... 1) i copied the recalled curve (no actual data anywhere when i did this but the same nominal curve), 5 times making 5 new features 2) I then de-selected curves just to select 1 single curve for three of the new features 3) I then de-selected curves in two of the new features to just two... one feature has the curve at the beginning and middle, the other has the curve at middle and end All of those came out with no errors... its only when I have selected all three, that the random error shows up...not just at the ends of one curve... some are in the middle of the the curve itself.....

for anyone else visiting this article at a later date, here's a KBA on the topic. https://portal.zeiss.com/knowledge-base?id=3139416

I took out the actual Part Numbers to avoid leaking any sensitive data. The PCM File runs fine except for the loadCADFile() function. when trying to compute it on its own it returns 0 shown in the attached picture. Any help would be greatly appreciated. Mike // IGVAPCM.txt //variable named 'paramFile' is equal to the actual .para file name // Then 'fileExists' checks if the .para file is in the program's directory then reads(readPCMFile()) it if its there. paramFile = "PARAMETER.para" if fileExists(paramFile) readPCMFile(paramFile) else // If file isnt there, this writes the file to the programs directory addToFile("PARAMETER.para","RH = getRecordHead("u_field1")") // part number header addToFile("PARAMETER.para","ORH = getRecordHead("planid")") // plan/program id header addToFile("PARAMETER.para","DWGREV = getRecordHead("u_field2")") // drawing revision header addToFile("PARAMETER.para","MNAME = getRecordHead("operid")") // machine/operator name addToFile("PARAMETER.para","MAT = getRecordHead("u_field4")") // takes into account the material being used endif // Main selection / confirmation loop starts here. // The repeat...until loop will keep asking until exitLoop becomes true. repeat PART = inquireList("Select the correct Program Number","ZA3657","ZA3658") validSelection = false // reset the flag for this loop iteration fileExists("ZA3657CAD.sat") fileExists("ZA3658CAD.sat") clearCAD() setCoordSysVisible(true) // --- PART-specific cases --- if PART == "ZA3657" loadCADFile("ZA3657CAD.sat") setRecordHead("planid", "ZA3657_PARTNO_002_OP40") setRecordHead("u_field9", "ZA3657_PARTNO_002_OP40") setRunID("ZA3657") // runs the named mini-plan validSelection = true endif if PART == "ZA3658" loadCADFile("ZA3658CAD.sat") setRecordHead("planid", "ZA3658_PARTNO_002_OP40") setRecordHead("u_field9", "ZA3658_PARTNO_002_OP40") setRunID("ZA3658") // runs the named mini-plan validSelection = true endif wait(10) redrawCAD() setFit() // If no valid PART matched (user canceled or selected an unexpected value), show a message if validSelection == false message("No valid selection. Please select a part or check automation headers.") endInspection() else // Refresh header variables so the next steps use what we just set RH = getRecordHead("u_field1") // part number we set above ORH = getRecordHead("planid") // program/plan id we set above DWGREV = getRecordHead("u_field2") // drawing / revision read back MAT = getRecordHead("u_field4") // FIX: ensure the header key is a string // Ask the operator to confirm the displayed metadata before starting exitLoop = confirm("ARE THESE VALUES CORRECT?", cr(),"PART NUMBER: ", RH, cr(),"PROGRAM NUMBER: ", ORH, cr(),"REVISION: ", DWGREV) // confirm returns true if operator clicks Yes/OK endif until exitLoop == true

I thought it was just rotation around z. This is version 2018. So I probably have to allow rotation in the freeform. Thanks.

That is not always the case and I suggest to look at the post from Klaus-Dieter here: https://qualityforum.zeiss.com/topic/24879-rules-of-thumb-for-2d-curve-evaluations/#comment-141085 Especially the files he sent.

Kann meinen eigenen Post nicht editieren 🤔 Manchmal steckt man so tief in seiner eigenen Bubble, dass man das wichtigste vergisst: Das führe ich dann an Bauteilen und nicht an Normalen durch, ansonsten kann man sich das sparen. Betrachtet wird hier dann nur die Wiederholpräzision und nicht der Bias.

Despite all the bugs and shortcomings I wouldn't want to be without our offline station. Programming at the real CMM is not an option anymore, since they constantly run programs almost 24/7 today. Yes, it needs a bit of experience to judge whether a simulated program will run properly on the CMM, but even that is better than blocking the CMM for hours when testing out a new program. And of course I'm better off with a false alarm in simulation than with a broken stylus or a damaged CMM. However, a real letdown in Planner was the simulation of our O-Inspect (which is now in CMM heaven). Zeiss sold us an additional 600 € video license dongle (so the camera drivers would also work on the offline station), promising we would be able to fully simulate with CAD models etc. But that never worked a bit. The dongle was completely useless as the simulation only displays a static dummy image that is always the same and has nothing to do with your part. It's impossible to simulate any part that way. Programming was also a nightmare as there are simply no functions to put an optical point properly on an edge. It always needed a lot of messing aroung with coordinates etc. Last but not least the light settings had to be re-teached on the real CMM anyway. I ended up not using Planner for the O-Inspect at all. It was just a big waste of time and money (for the dongle).

Karsten beschreibt das Vorgehen nach Definition. Ich arbeite öfters auch mit Verbundbauteile mit Elastomeren, da hat es sich schon als sehr nützlich erwiesen die MSA 1 (bzw. Vergleichbare Vorgehen im Rahmen der Messprozessentwicklung) mit Umspannen durchzuführen, weil ich hier Bediener bin und dann genau weiss was ich zu welchem Ergebnis getan habe. Diese Info bekomm ich durch die MSA 2 später nur durch viel Glück in der Qualität. Der Hauptunterschied wird hier das Verhältnis aus Toleranzbereich der Bauteile zu Genauigkeit der Messgeräte sein. Bei mir stellt sich in der Regel nicht die Frage nach der Genauigkeit des Messgeräts.

Norbert, I completely agree with you! The current version of Planner is pretty rubbish. You should think very carefully before buying an offline station. Generally speaking, however, Zeiss promises a simulation that works 100% of the time and sells it (at a very high price). You have a lot of experience using Planner and are surely familiar with all the pitfalls, big and small. New users, however, assume that the simulation works 100%. If the software fails to detect collisions that are not immediately obvious, it can take a long time to identify the problem in the real world. Or if the measuring space is insufficient – particularly with RDS – and an end stop is reached in the real world. It is even worse if collisions with the machine’s own geometry go undetected. All of these are basic functions that the current software version only covers inadequately. As I said: if you have plenty of experience, you might not need 100% simulation, and 80% might be enough (though even that isn’t always achieved...)

Have you looked at this? https://stylicleaner.com/

Hello colleagues, We have a number of DuraMax machines in operation in our production environment. Despite intensive extraction, there is always some oil in the air, which builds up on the stylus systems over time. The carbon fibre extensions (ZEISS REACH CFX) are particularly affected. Some of them start to feel a bit sticky after just one year of use. Occasionally, there are problems with the contacts on the old XXT plates that do not yet have gold-plated contacts. How do you clean your XXT plates and, in particular, the carbon fibre extensions? We clean the probes with steel shafts and ruby balls using 99% alcohol. An automatic probe cleaning system would, of course, be a real hit in a shop floor environment. For example, every five times a probe system is used, the probe would first move to a cleaning station. Well, we can always dream 🙂

Maybe it's my lack of imagination, but I never understood why people want to simulate each and every bit of their CMM setup in Planner anyway. I learned very early that the simulation is very limited when it comes to anything other than pure collision detection between stylus and workpiece. And even that only works reliably maybe 80% of the time, and only on the lowest speed setting. I started with Planner in 2009 and many of the bugs I encountered back then are still present today, like the CMM making weird moves especially after a probe change, which do not happen on the physical CMM. That alone was enough to keep me from simulating the whole setup. Because if the virtual CMM can't even move realistically, all the other stuff doesn't make much sense. So most of the time I only simulate the workpiece, the fixture and the stylus, maybe also the head if it comes very close to the part. All the other things are only there to impress the boss 😄

What is checked, then it's used. If there is a profile callout with datums, then you have to look what is free. If you have only A and B in callout's datum, then you allow only rotation - nothing more. If you have a newer Calypso, then use GD&T and fill only used datums - rotation will be automatically added.

UPDATE: Calypso program editor My GIT repository: https://github.com/justyand/zeiss_pgm_editor Direct link: (v.0.1-alpha) https://github.com/justyand/zeiss_pgm_editor/releases/tag/v0.1-alpha

I have a round part with a plane close to the top. Datum B is the plane. Datum A is the OD. Base alignment is the plane as Spatial and Z Origin, OD is X and Y origin, and there is a small slot in the top plane the is perpendicular to the plan, but rotated 15 degrees. This slot is my rotation in the base alignment. On top of the -B- plane is a scrollwork like feature with a profile to ABC. There is also a profile for the scrollwork. This scroll work walls are basically perpendicular to the plane -B-. How do I set up the fitting options for the scroll work freeform in order for the profile to just A and B reads correctly. I think I want to allow rotation and X and Y and that's it (Looking for advice to understand)? When you check the box is that allowing rotation or stopping it?

but i am also looking at the Raw Data.. I see the noise it that even without the GDT evaluation

classic for both... cant really use the new one for this type of eval.

Are you using the new GD&T Line Profile or the Classic Line Profile?

I have done this several ways and i am still getting random noise... I have created noms 3 different ways 1) from recall points , then recalculate.... in order to do this no data can be in the features yet 2) I created a text file with all the nominal points and either "read into" within the curve or use the points set 3) used the cad model to generate the nominals once that was done, i recalled the points using both non-filtered data and checked the filtered data. Every one of those things creates the same random "noise" .... sometimes it is at the end of one curve some times its well into the curve....

It may be "Plane1ba". Martin's data seems to lean on that feature. And I just found the problem. Inside of Plane1ba is an extra Clearance Data and guess which probe is attached to it. I have no clue how an extra Clearance Data was added or how it would have been assigned a probe that isn't being used. 🫣

You can add the probe racks to the simulator, but only virtually. You cannot mount them to your table and do crash detection. (If that changed in version 8, I cant find it!) It would be so nice to be able to simulate and not have to validate on the real machine!