Multiple Calibration Spheres

[[Me...]]

Anybody on here have experience making a program to qualify styli with multiple calibration spheres?
When teaching the reference sphere position can you teach it more than one sphere at a time?
Anybody automated teaching the reference sphere position?
We have a Contura with the VAST-XTR Gold sensor and 24 different stylus systems in all sorts of configurations.

I currently have a "standard" holder with one 30mm artifact, it would take 2 days to calibrate these by hand. I'm thinking of building a "tree" with 5 spheres to avoid having to keep moving the current holder. Does anybody on here have experience with something similar? I don't have PCM.
TIA
-Dave

[[Ja...]]

I use a 30mm and 8mm Cal Sphere. They are set as Sphere 1 & 2. I reference each one with the Master Probe. Each stylus is set to it's respective sphere. When they run in the cal program they know which one to use base on the initial Cal Probing.

[[Aa...]]

Just be aware, you're throwing away some of the accuracy of your machine when you use a second sphere like that. The machine was calibrated (tuned in and verified to be within the machine specs) using the Ref Sphere that came with the machine to qualify. The "calibrated" size of the RS may be off within a certain amount (which is a significant amount in the scope of the accuracy of many of Zeiss' machines). The difference between the errors in the calibrated (entered in Calypso) sizes of the RS's will directly add to the inaccuracy of the machine.

Perhaps you could switch to a single orientation (straight up?) for qulifying all of them? You couldn't do a 180-degree dynamic tensor qualification for a lot of probe orientations, but you don't need a 180-degree dynamic tensor qualification. Even if your probes "drift", their rigidity shouldn't be changing.

[[Er...]]

Hum, I might be slow, but how are you losing accuracy. Obviously he isnt talking of attatching 5 pingpong balls to toothpics. As long as you know the size and roundness, and ridgity of the holder if available, I cant see how accuracy is lost. As long as the masterprobe is used to locate them all.

[[Me...]]

The plan is to buy calibrated sphere's from Zeiss, I'm designing my own tree 3" in dia. with a beefy base to hold the spheres.

[[Aa...]]

Please sign in to view this quote.

Therein lies the issue. There's an uncertainty on the calibrated size that equals half the machine spec for some machines. When the machine is calibrated, parameters adjusted in the machine counteract inaccuracy in the diameter. (They measure a traceable artifact with a stylus qualified on the reference sphere.) In this manner, the uncertainty of the reference sphere doesn't directly contribute to the uncertainty of measurements on the machine. But when you bypass the machine calibration by swapping/supplementing your reference sphere incorrectly, that is no longer the case.

[[Er...]]

But Zeiss never adjust the size of your sphere. I never ever seen or heard about that. And the accuracy of the sphere is grater than your cmm is able to accuratly pickup. What probing error do you have?

What parameter are you talking about? CAA? That has as far as I know, not a thing to do with your sphere. That compensate for unsquareness of the machine etc.

I can get the traceabillity part, but accuracy. I dont. Unless its a shitty sphere 🙂

[[Aa...]]

Please sign in to view this quote.

No, I'm talking about the spheres from Zeiss, which come with a measurement uncertainty of 0.3 microns, (which, by the way, the sphere calibrations don't fall within Zeiss' scope of accreditation, according to the certificates.) Our Micura spec starts at 0.7 microns. The Xenos spec starts at 0.4.

[[Er...]]

Yes, so if you use as an example 5 spheres, with the same accuracy down to the 4:th decimal. How can you possibly loose accuracy? 😱

Let's say I have a sphere with the radius of 14,9883 ±0,3 μm. So it's somewhere between 14,9886 and 14,9880. But in the software you say it is 14,9883. Then you probe it, and your styli radius will be determined based on size calculation fron the given size of the sphere.

Uncertenty range is: 0,06 μm.

Then you have another sphere, let's say its 14,9889, so somewhere between 14,9892 and 14,9886. Probe number two then gives you the same uncertainty as if we used sphere number one.
Uncertenty range is: 0,12 μm

The range we use for determination of size is bigger, yes. But since the size's are known, we don't loose accuracy. Unless we measure the wrong sphere 🙂

But once I was wrong 🙂 (I've heard) So if I am wrong, please try to explain so a stupid Swedish guy understands 😃

[[Me...]]

Please sign in to view this quote.

I could well be wrong but, I thought the tech brought their own sphere to calibrate?


I like what you said about single orientation, but I think I can only do geometry without doing a full 180?
this sensor (VAST-XTR) is new to me. I previously used the RDS-CAA and seemed I had to a full qual about every third qualification as the sigma would start getting too high.

[[Aa...]]

The traceable sphere they brought with them was used for the 10360-5 verification step, not for qualifying probes.

[[Me...]]

So what's the best way to calibrate a .5mm probe that shanks out on a 30mm RS? I was going to get an 8mm second sphere but it sounds like that's a bad idea?

[[De...]]

It is not bad to have the 8mm sphere.

You can still use the smaller sphere to do a dynamic tensor calibration.

Once that is complete, use your original 30mm sphere for calibration using a smaller number for sphere coverage, and choosing tensor re-qualification.

Tensor Re-Qualification will update your tensor data for radius / location but without updating the bend allowances that dynamic tensor calculated when you ran it.

Dynamic tensor should only need to be run for a given stylus once, after that so long as you dont wipe out the the data by running just a regular tensor calibration, the bend allowances should be retained. So basically, do dynamic with your 8mm and let it setup your bend allowances and then run tensor re-qualification with the 30mm sphere to update your size and location data.

Afterwards running a new tensor re-qualification should be all you need.

[[Me...]]

Thanks, Derek. I appreciate the info. I haven't used really small probes with Calypso before that required a smaller RS. I tried to calibrate it on the 30mm RS in RC-CAA Mode but it failed.

[[Aa...]]

It's not that I want to resurrect an old thread, but I've recently had to revisit this subject, and thought I should post what I've learned. I got this email back from Zeiss's Quality Systems & Calibration Manager:

This can be a bit of a complex answer, but I will try. If a different sphere is used to define a stylus, it will affect the end measuring results, similar to changing the value of the initial reference sphere used during calibration. There are ways to minimize this impact, which can include factors such as using your own traceable artifact to conduct a comparison and derive the best value for the secondary sphere. Either way, it is best to conduct an R&R for best results, and to more accurately determine what the actual impact is between spheres.

[[To...]]

Haven't seen Derek on the forum in a long time. Hope is ok. I sent a PM to him back in April 2021 and it is still in my outbox. (which means he hasn't read it, for the new members of the forum)