parid

Question 1

in which cases is it legitimate to use measurements from the last_measurement directory? and why?

Answer 1

ONLY for RECAL.par8s
Otherwise the parid results cannot be reproduced, once new meas (real or Sim) are made

Question 2

There are only few free params, but the accuracy is 0.0mm anyway. What could be the cause?

Answer 2

using last_measurement meas which were done in Sim instead of real ones

Question 3

How does one properly commit parid results that shall be saved?

Answer 3

run calibration, producing results you want to save
save .par8
save all
commit whole parid folder
make sure svn st and svn up are clean
go to parid/calibration/TIMESTAMP
finalizeParid
write results into report directly if necessary

Question 4

Parid produces errors/crap results on a recalsys. How does one troubleshoot?

Answer 4

• get data from IEPC
• run initial residuum
• check dates of meas
• check initial residuum -> if bad, then check why

Question 5

what happens if parid detects dependencies with one variable?

Answer 5

It is fixed at it’s current values and NOT identified -> you should not have dependencies!!!

Question 6

when doing Cal+Ref, and then wanting to produce the same result by running only ref, what do you have to think of

Answer 6

save all
use the robot that was saved by save all (in different dir!!)

Question 7

How does one start with parid?

Answer 7

use ideal robot as start
hold the robot to find mea devices and save those
only then start to identify real robot

Question 8

What 5 problem categories exist in parid?

Answer 8

mea is erroneous
meas are incompatible
residuum too low
residuum is bad indicator
convergence problems