Calibration / Repeatability

[[Wo...]]

Hello.
I have several issues with calibration and repeatability.

I have an Accura with XTR head and hundreds of probe position to calibrate.
I wrote a qualification program with sphere at 180 degrees that I would run thru the night to calibrate all the probes (takes about 12 hours).
I would use a tensor calibration thinking it's more accurate (?) than just geometry requalification.
Then I realized, that tensor calibration doesn't take a probing point on the very tip of the probe.
I've noticed that because I had a probe with a flat spot at least 0.003" on the tip and was qualifying beautifully with tensor (0.0002mm sigma). Only after doing geo. requal. it went up to 0.0015 sigma.

So I've started doing requalifications with geo. requal. in four sphere positions ( to make sure that styli is checked on all main points - can't do Z+ on a horizontal styli with a sphere at 180 degrees).

Now I've started seeing high sigma ( 0.0010mm) on brand new styli with high rigidity (53.8), when doing geometry requal.
I would also get high sigma randomly on the same styli at different XTR rotation. Tensor shows much lower values - usually below 0.0004mm

So I decided to approach the whole thing differently.

I have ordered a 3" ceramic block and brand new 8x65mm probes ( same size as MasterProbe ).
I wrote a program that takes points on the block with that new probe ( not used for anything else) and then measures the exact same points with my regular styli.
The idea is that the difference between measurements should not be more than a specified value. That would detect a flat spot or a deposit on a probe.
The question is - what's the value that the probes should repeat at?
I've set the tolerance for 0.001mm, but my probes have a hard time passing that tolerance - which I find concerning.
I would think that probes that were just calibrated should easily be within 0.001mm from each other on a simple point.
Or do I expect too much? What do you think that value should be?

Thanks for any input.
Wojtek

Capture.PNG

[[Da...]]

Hi Wojciech,

I think you just overthinking mate 😃 Just run 12 position calibration and you will get all the angles calibrated. Probably less accurate method but it will save you those 12hrs which is ridicules.

[[Wo...]]

Hi Daniel,

Can't do 12 positions - it's an XTR not an RDS.

Thanks 😉

[[Ri...]]

I've dealt with this scenario before.
To save time qualifying I spent the time evaluating each program (paid attention to our top 20 products at first), then determined which articulations could be removed and use an existing articulation as close to the Five Standard Articulation's.
Z- (A0B0)
X+ (A0B90)
Y+ (A-90B90)
X- (A0B-90)
Y- (A-90B-90)
(A nd B angles are examples that reflect the axis stated).

After completing this task, and continuing on to the top 200 programs I went from 81 Probe configurations to what I coined as "9 Standard Probes".
Nine probes that perform the same as 81 probes.
This was all on a VAST XT Gold.

I think this same scenario could be used for your issue.

[[Ri...]]

You can also create a measurement plan that scans the reference sphere along each axis of the stylus shaft. This will help detect these things.

[[Wo...]]

Please sign in to view this quote.

Hi Rick,

The amount of probe rotations doesn't really bother me. I can easily run it overnight and have calibrations done by morning.
As for reducing the number of probes - I use mostly four 5-probe stylus system daily on 90degree "A" rotation.
The 15degree increments positions are used occasionally so I calibrate the only every couple of months.
Since we do mostly prototypes I never know which probe position will be used the next day, so I keep the styli up just in case.

What bothers me is that I'm seeing differences over 0.001mm between probes on actual measurement on the block.
I'm wondering which calibration method would provide the most consistent measurements. Choosing one over other is just a few clicks in the calibration plan.

On the block I'm not checking every single calibrated position, just enough to touch each point of a given probe along main axes to check for wear.

Thank you for input!

[[Wo...]]

Please sign in to view this quote.

Hi Richard,
I assume the scanning path would be similar to the one used by dynamic qualification?
Scan along the circumference and then crisscross over the top?
What would be the characteristic that would point to a worn probe, Roundness of the Sphere?

Thanks for your help!

[[Ri...]]

I would start with Roundness, Diameter, and the XYZ deviations.

[[Ri...]]

Please sign in to view this quote.

0.001mm between styli is not a high value.
For the metrically challenged (myself included) that's 0.00004".For CMM's in typical "Labs" (insert eye roll here), I think the values are relative to the SIgma of the styli and Uncertainty of the CMM.