Versions Compared

Old Version 13

changes.mady.by.user Mikhail Dubrovin

Saved on

compared with

New Version Current

changes.mady.by.user Mikhail Dubrovin

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

Table of Contents

Why do I think that 2020-02-25 optical metrology is good

  • metrology was done on 2020-02-25 for new detector which currently labeled on connectors side as ePix10k2M 1, also known as epix10ka2m.1, 2M.1 etc.  
  • very good quality check
  • all z-measurements are around 1 feet, have a lot of repeatable numbers - can't be trusted

...

  • red rectangles - for points measured in optical metrology
  • blue rectangles - panels reconstructed from geometry file

DAQ panel indexes in optical metrology possible detector orientations

Code Block
titleDAQ panel indexes in optical metrology possible orientations
        rot = 0  ^y
        Q0  1  3 |  4  5  Q1
            0  2 |  6  7
        ---------+----------> x
           15 14 | 10  8
        Q3 13 12 | 11  9  Q2

        rot = 1  ^y
        Q3 13 15 |  0  1  Q0
           12 14 |  2  3
        ---------+----------> x
           11 10 |  6  4
        Q2  9  8 |  7  5  Q1

        rot = 2  ^y
        Q2  9 11 | 12 13  Q3
            8 10 | 14 15
        ---------+----------> x
            7  6 |  2  0
        Q1  5  4 |  3  1  Q0

        rot = 3  ^y
        Q1  5  7 |  8  9  Q2
            4  6 | 10 11
        ---------+----------> x
            3  2 | 14 12
        Q0  1  0 | 15 13  Q3

 

Images of panels in data for all detector orientations in metrology processing v1 without quads

  • v1 of the metrology processing script represents simplest approach - each panel coordinates are taken as is to insert in the detector
  • data from mfxc00118 run44 silver-behenate sample
  • data panel 1 (numerated from 0 to 15) is replaced by intensity ascending array, darkest corner is (0,0) pixel.

...

  • numeration of quads at rotation looks symmetric - so seems right
  • one orientation of the detector gives correct image and it shows smallest offsets between panels in quads

Detector orientation in the hutch

mfxc00118 runs 70-73 for Fe55 in the corner of quad 0-3

Code Block
titleepix10ka2m.1 orientation in the hutch, view from the beam (or on open panels)
   ^y          |
Q0 | Q1        |
---+----->x    |
Q3 | Q2        V g (gravity)

Panel 0 data is replaced by the ascending intensity array, (0,0) pixel is in the darkest corner.
Image Added

mfxc00118 run44 silver-behenate

panel 1 is marked

 

mfxc00318 run 10,13 lysozyme

psana direct and mirror image for crystfel

panel 0 is marked on full size image

 

Geometry files centered for mfxc00318 run13 lysozyme

Geoptimizer or how geometry can be screwed-up

...

First, thanks to Olexandr and Marina for discovering how our geometry should be mapped to data array.

...


However, before solution was found, geoptimizer returned geometry which is significantly different from optical metrology, as shown in movie comparing panels before (from opt metrology) and after optimization.

  • It is clear that blind application of geoptimizer may cause >1 mm offset of panels in x-y ... 

 

References

  • It is suggested to use geoptimizer for panels with fixed center and panel orientation in (x,y) from optical metrology and apply optimization for  z coordinate and off-plane tilts. Z coordinates were not measured correctly in optical metrology.
  • Next optical metrology is expected to be with precise (~10um) Z measurement.

References