XXT Normal Verification Values

[[Mi...]]

Good morning all,

Former VAST user here who recently moved over to XXT, and I'm looking for a final sanity check before deciding whether I need to bring in a Zeiss service technician.

My primary question is: What verification values would you consider normal for a standard RDS-CAA down stylus?

For example, if I check the calibration sphere with the master stylus and then verify using a 2 mm × 40 mm down stylus at A0B0, what X, Y, and Z deviations would you typically expect to see?

A while back I noticed some unusual results on a program for a part I had run before. Well after doing some troubleshooting, I was looking at the stylus and did a verification on the cal-sphere. Those numbers came up pretty high, I forget exactly. Anyway, I cleaned the calibration sphere, verified that everything was tight and clean, and recalibrated the stylus. Calibration completed successfully, although the S value was higher than I would normally expect (I forget what the value was specifically but higher than .0010in (.0254mm). The verification issue remained. Eventually I shortened that stylus and the results improved, but when I checked my other RDS-CAA styli I found what seemed like consistently high verification values. The calibrated sphere size repeats very well, but the sphere center location does not agree closely with the master stylus location.

My expectation would be roughly 0.0005 in. (0.012 mm) maximum deviation, but I'm seeing values that are often 2-3× higher depending on the stylus, generally with a consistent shift in one or two axes. In contrast, my non-RDS-CAA styli verify very close to nominal and repeat well. I understand that an indexing system will introduce additional uncertainty compared to a non-indexing stylus, but I'm trying to determine what is considered normal versus what may indicate a problem with the system. S values are running generally about .0005in (.012mm).

Just trying to make sure I'm not missing anything before making a final reccomendation.

Pics of an example, I have seen that deviation in X as well. Y has been good. 

Thanks in advance

 

As an additional question, do any XXT users incorporate calibration or verification sequences directly into their measurement programs for tighter-tolerance applications? I know this is fairly common in PC-DMIS environments and I'm curious how others approach it in Calypso.

 

[[Ri...]]

From my XXT probe checking program where I use the 5 standard articulations (1 thru 5).

I use the Reference Sphere, XYZ zero is set by the MasterProbe.

I then use each articulation and measure the sphere.

I report the True Position (TP) for each articulation at .0003 Sphere (Shape of zone).

If any are OOT I investigate the issue.

Using a Contura 900x1200x600 I run this program once a month (I do a similar test with the VAST XTR Gold).

Every month, the Reference Sphere is placed in a new location, I use 5 standard location and two levels of height in Z axis (My riser is 6 inches).

This is how I detect nonconformities with a sensor; this method allows the sensor (either one) to move in a wide area around the Reference Sphere while taking measurements and uses a good portion of my volumetric area. Typically, I can detect if a stylus, a sensor or a quadrant has an issue that should be evaluated, after evaluation if found to be a non-fixable issue in house it is escalated to Zeiss.

I have been using this method for almost 20 years, and it hasn't failed me yet.

As an FYI, I also perform this test with no filters or outliers to detect sensor issues, but I use a wider tolerance. When I run without filters and outliers, it is considered a non-quality program and not for suitable part measurements.  

I plot data from both programs in excel which tells me if something is wrong. The second program is only run when I think there is an issue after the all testing of the first program is exahuasted.

Over laying bell curves of the quality program and the non-quality program should reveal shifts that indicate an issue.

[[Mi...]]

OK so you're saying that you set your limit at .0005in true position. Which aligns with my understanding as well for an expectation, maybe a little tighter than I would've said but broadly the same. 

I like your strategy of moving the cal-sphere around the table for subsequent iterations. I think that's a good set-up in general. I've had some ideas as to how to codify the verification moving forward, I think I'll incorporate some of that. 

Anyway, thank you, I appreciate it.  I was really just looking for someone who is very familiar hands on with an XXT to share what their standard is for location verification. Everyone I talked to was somewhat back and forth on what numbers I should be holding that to.