Vision Offset Calibration Order

I think you have to calibrate vision first on optical cmm, and touch first on cmm which have a optical probe (cmmv for example).
In both cases, to the question "did the tool moved", say yes to the first calibration and no to the second...
I hope it's clear for you (not exactly for me ! LOL !)

In both cases, they are using the same PC-DMIS software, and as far as I am concerned, calibrating the offset is to pull their orgins together using the ring tool.
So I am confused for what reason we need to define the calibration order, will it work if we calibrate touch probe first on an optical CMM?

Calibrate the optical first to the ring, when asked if tool moved, click YES. Run in MAN+DCC

Then calibrate your main probe to the ring, when asked if tool moved, click NO. Run in DCC+DCC.

Then calibrate that main probe to the sphere, when asked if tool moved, click YES. Run in MAN+DCC.

Calibrate any other probes to the sphere, when asked if tool moved, click NO. Run in DCC+DCC.


Everything is linked together after you calibrate your vision to the ring.

I think Bfire is correct his way sounds just fine.

I just want to add my two cents.

I write programs daily using both my Optics and TTP.

My order is:
-Calibrate my main TTP to my sphere hit "YES" if the tool has moved.
-Calibrate any other tips to the sphere but say "NO" ( I only do this for my star probes)
-Calibrate your TTP to the ring next and select "YES" again and run MAN+DCC
-Calibrate your optics to the ring and select "NO" and make sure "Calibrate Probe Offset" is selected
-Lastly I calibrate my main TTP to the ring again.

In my experiences I am not sure but I have done this calibration so many times and for some reason sometimes one order works one time and it will not work the next time. This is the only order of operation I have found to be full proof to work everytime.

Good Luck!!

Recently I have a better understanding about this offset order issue.Below is my understanding:
1)When we define a new vision probe , if we go to the probe dialogue and click "EDIT" we can see the XYZ, for example , <0,0,127> ,here 127 is the focus of the camera, and <0,0> should define it's center
2)When we calibrate the vision probe to the ring gage first saying "The tool has moved" then touch probe saying "The tool has not moved", PC-DMIS will change the XYZ of touch probe instead of changing Vision, which may have an influence on the vision measurement

If you're using a Vision CMM, then you want to Calibrate the Vision probe, and then cross calibrate the TTP to the Vision probe. NEVER the other way around.

*the other way around is when you have a touch CMM with an add-on Vision probe, like the CMM-V*

Afotion, you have given this latter procedure WHICH IS INCORRECT. Note that you are slave-calibrating your Vision and setting the master probe as the TTP. Cease this immediately!!

Here is the procedure:

- calibrate Vision Optics, Optical Center, and Illumination (may not be necessary) (result: Vision probe is cal'd)
- calibrate the A0B0 TTP to the sphere. YES, sphere has moved. (may not be necessary) (result: TTP is cal'd, sphere position is reset in machine coordinates)
- calibrate Vision Probe with the Ring Artifact. YES, the ring has moved (result: ring position is reset in machine coordinates relative to the probe)
- calibrate TTP with the Ring Artifact. NO, the ring hasn't moved (result: probe position is reset in machine coordinates relative to the ring = probe offset is calc'd)

NOW

- recalibrate the TTP A0B0 master tip to the sphere! Since we have changed the "position" of the probe by changing the offset, the sphere "position" has moved. So YES, the sphere has moved. (result: sphere position is reset in machine coordinates relative to the newly offset TTP)

- now calibrate any other tips. NO, the sphere hasn't moved, we locked it to the correct position previously.


Cross calibration complete

If you're using a Vision CMM, then you want to Calibrate the Vision probe, and then cross calibrate the TTP to the Vision probe. NEVER the other way around.

*the other way around is when you have a touch CMM with an add-on Vision probe, like the CMM-V*

Afotion, you have given this latter procedure WHICH IS INCORRECT. Note that you are slave-calibrating your Vision and setting the master probe as the TTP. Cease this immediately!!

Here is the procedure:

- calibrate Vision Optics, Optical Center, and Illumination (may not be necessary) (result: Vision probe is cal'd)
- calibrate the A0B0 TTP to the sphere. YES, sphere has moved. (may not be necessary) (result: TTP is cal'd, sphere position is reset in machine coordinates)
- calibrate Vision Probe with the Ring Artifact. YES, the ring has moved (result: ring position is reset in machine coordinates relative to the probe)
- calibrate TTP with the Ring Artifact. NO, the ring hasn't moved (result: probe position is reset in machine coordinates relative to the ring = probe offset is calc'd)

NOW

- recalibrate the TTP A0B0 master tip to the sphere! Since we have changed the "position" of the probe by changing the offset, the sphere "position" has moved. So YES, the sphere has moved. (result: sphere position is reset in machine coordinates relative to the newly offset TTP)

- now calibrate any other tips. NO, the sphere hasn't moved, we locked it to the correct position previously.


Cross calibration complete

Colin,

Thank you for the correction! I believe the the last step you mentioned where you calibrate your TTP back to your sphere again has never been iterated to me correctly as this was not explained to me like that in training either nore from my trainers. Does the calibration work without re-calibrating the TTP back to the Sphere AFTER doing TTP to sphere, Vision to Ring, TTP to Ring. If not I believe this could have been the cause of all of my heartache while troubleshooting wrong measurements.

Thanks Again!

This was posted before by BuckB.. Very good Information please read below:



The question whether the fixture has moved will have two different results. Yes its moved will update the location of the calibration artifact whether the ring or sphere. No it hasn't moved will update the probe file from the last known location of the calibration artifact.

On a vision system always measure the ring first with vision and always say YES its moved. The vision probe is the Master that everything else is offset back to. This will locate the ring in the measurement range. Your vision probe should always have a TIP location of 0,0,#. If X and Y are not 0 then someone answered NO at some point.

Then choose a touch probe tip to be the secondary master. Measure TIP1 only on the Cal Sphere and answer YES. This will update the probe radius but will not update the probe location. Then calibrate the same tip on the ring gage and answer NO. This will update the probe location relative to the camera. If it is the first time you have ever measured the offset always use MAN+DCC. The manual hit is straight down about 5 mm from the inside diameter at 6 O'Clock. If you have done the offset before and know its close then you can run DCC.

Then if using multiple probes or star probes, go back to the cal sphere with the same TIP1 and answer YES its moved, this now updates the location of sphere with the corrected offset of TIP1 to the vision probe. All other tips can now be calibrated back to the cal sphere and answered NO like a normal CMM.

Hopefully this makes sense.

Nice post, Eva !
But be carefull on 2014SP2, I think answering yes isn't enough to zero offsets (Don't know about vision cmm, I can't check it !). You have to "reset to theo values" AND answer Yes...