Fix fix of coord transformation
As discovered yesterday evening by @williamm and me, MR !260 (merged) fixed the filling of in-pixel histograms but broke the calculation in the detector member functions
It turns out that this was previously done correctly, and the problem fixed in !260 (merged) was only related how these member functions are used.
These commits from MR !260 (merged) need to be reverted:
75552e28 was wrong and is hereby reverted. I had to invert in by hand because the separation into
PixelDetectorprevents me from using
c091264a: updating the PASS conditions is reverted via
These commits from MR !260 (merged) contain valid changes:
- 1ec6c7b8: AnalysisDUT: simplified filling of histograms with assoc_cluster->column/row()
- 5f71c04a: AnalysisDUT: correct filling of 2D histogram: column/column -> column/row
Below is a validation of the changes:
I'm comparing the in-pixel efficiency for DESY run 718 from June 2019 with two ATLASpix samples as a DUT and a high threshold of 1000mV which is causing visible efficiency losses in the pixel corners. I repeated the check for run 742, same APX conditions, other sample.
left side: master on f9476e32
right side: this MR on 45c3e835