stylus deviations

[[Cl...]]

Have a question. What would be the "acceptable" tolerances for stylus deviations? Specifically the X and Y values. When I do a stylus calibration, the "S" value is 0.0 but depending on the length of the stylus I could get readings of say -0.054 on the "X" and .504 on the "Y"

[[Ju...]]

What units do your qualifications use and to how many decimal places? I tend to use metric to four decimal places.

[[Cl...]]

We're using metric and to 2 decimal places.

[[Ju...]]

We usually investigate stylus deviations when the sigma value gets above 2 microns. (0.002)

[[Ma...]]

We are using 0.0030 mm tolerance - mainly casted parts

[[Je...]]

S = sigma. Loosely defined as form or "error". This is the critical value for stylus qualification.

X/Y are merely location in space relative to the MasterProbe. They are irrelevant to a "tolerance". Delta of this value between calibrations may be important to some but isn't necessarily important to all (for instance if you disassambled/recreate stylus configs)

[[Ni...]]

Confirm you are on metric by going to measurement tab > Measurement plan editor features > Units > Length Unit

I was told by Zeiss trainer that anything above .0010 metric is not guaranteed by zeiss. So we try to keep it below that. Our RDS XXT probes usually come out below that no problem as long as they're good and clean.

The vast XXT Gold was the only one that would come out at .0010 - .0016 so we raised the tolerance for that one up to .0016
Anything above that then we knew there was something wrong.

The X and Y don't mean anything. It is just telling you the distance away from the master probe and reference sphere. It pretty much is defining where it found your probe. You can't really tell if its bad based on that.

[[Cl...]]

Thank you all for your input!

[[Da...]]

It should be clear, when interpreting the Sigma value, that it doesn't simply reflect "the form error". The form error itself would rather be the minimized range of touch points during calibration, comparable to the range of a measurement of a Chebycheff sphere.

The Sigma value is the square root of the variance value, which represents the squared mean deviation to a normal feature. Since the variance itself is a squared value, it's not very intuitively interpretable. That's why we use the Sigma value, which is the variance's square root.

The range value contains 100 % of all touch values, the Sigma value contains only about 34 % of all values. If you compare the 6s value (2x 3 Sigma), which contains about 99.7 % of all touch values, to the range value, you get a pretty good idea about the relation of normally distributed points to outliers.

So, better don't see the Sigma value as form error, but rather use the 6s or ±3s value (2x 3 Sigma) to get an idea of the real form error without outliers.