8 Series - Measuring Setup Calibration Fails

[[An...]]

Hello,

When running the VMR Measuring Setup Calibration, i capture data on the Part and Fixture with both robots and can visibly see missmatch between the LHS and RHS scans.

Adding more scans doesn't improve this and the Calibration and the process fails.

I don't understand why this has happened or how to correct it.

[[Ja...]]

Hello,

the message in the tooltip indicates that something has been measured, that is not modeled as CAD (in the part or the fixture) *and* that is further than 100mm away from the nearest CAD data. This can be an indicator for a danger of possible collision during movement, since the software can only check for collisions if the objects are modeled in the CAD!
Look for mesh that is colored in red (while the "Calibrate Measuring Setup" dialog is open) to find the spot in the 3D view. Depending on what/where the problematic measuring data is, there are two main options:
* Adapt the CAD model (for example if there is something missing in your fixture model, that exists in reality)
* Cut away data if nothing corresponds to it reality (can happen in rare cases caused by reflections or similar)

Note that the precision of measuring calibration is not comparable to the final measuring result, as it does not use any reference points or more complex alignments between individual measurements. Some visible discrepancies are expected and do not worsen your final measuring result.

Regards,
Jan-Henrik Kluth

[[An...]]

Thanks Jan-Henrik,

This is a program that was fully functional (In 2019), but now will not complete this step in (2020).

I've just re-run the whole setup and there is no "Red" out of spec data data visible or captured. There is 15.2% Yellow Not Used, 70.3% Measuring Object and 14.5.

I can still accept, but the setup will not be calibrated.

Best regards,

Andrew

[[Ja...]]

Hello Andrew,

we changed some parts of the "Measuring Setup Calibration" from 2019 to 2020. For example, you can now move the part (or the fixture) while the dialog is open by clicking on it and supply the algorithm with a better "starting solution". With that we also changed some parts of the algorithm that dealt with cases where the CAD was very far away from the measured data, because that often led to completely wrong results (for example completely flipped parts). Maybe that affected you project.

However to make a more definitive statement in your case, we would need to analyze such project directly. If this is possible for you, can you please contact the GOM support (support@gom.com) for a way to transfer a project? The project should contain the data measured by the software for the measuring setup calibration, but the measuring setup calibration itself should not yet be executed (click 'Cancel' in that dialog).

I am sorry for the inconvenience.

Regards,
Jan-Henrik Kluth

[[An...]]

Hi Jan-Henrick,

Multiple projects have been sent to GOM-UK and i believe onto Germany.

We ran the Cell Calibration and Hand-Eye scripts with GOM-UK, when it scans the Robot Rail, Calibration Board positions etc and it appears all of the Slave Side of the Cell is "Out of Position". This Out of Positon is what we see when we run the Set-Up Calibration.

This is backed up when i run the Slave Robot Calibration Measurement Series on the Calibration Board, the Closest Board positon fails 100% of the time. I have to manually update the Sensor position, moving the Robot/Sensor away from the Calibration board to then continue. Master side runs fine.