Skip to content

Bump connectivity scans

Michael Daas requested to merge bump_connectivity into development

This MR adds and improves some things related to source testing and scanning bump failures.

  • Add bump failure scans:
    • A merged bumps scan based on detection of high crosstalk
    • Three different disconnected bumps scans:
        1. Crosstalk-based (fast: 8min + 1min [tuning + scan])
        1. Based on threshold shift upon sensor bias change (slower: 5min + 12min [tuning + scan])
        1. Modified source scan (need source, rather slow)
  • Improve other scans:
    • Detect bad pixels in digital and analog scan
    • Normal source scan: Add stop condition minimum number of hits per pixel
    • Noise occupancy scan: Send much more triggers and properly use statistics for noisy pixel identification
  • (Add optional HitOr-trigger based noise occupancy scan for special noisy pixels not seen with normal scan)
  • (Return uncertainties from S-curve fitting, add optional argument for passing fit start parameter)

TODO:

  • Source-based bump scan: Limit occupancy of noisy pixels and ignore disconnected pixels during estimation of average occupancy per tile
  • Add meta scripts
  • Cherry-pick changes to S-curve fitting
  • Cherry-pick HitOr-trigger based noise occupancy scan
  • Perhaps add ToT-histogram fitting routine to (modified) source scan in order to also catch disconnected pixels that are very noisy?

NOTES:

Estimated errors* of bump disconnection scans are:

  • [25x100]:
  1. Crosstalk: N_{disconnected}^{real} \simeq N_{measured}\!\,^{+6 \pm 5}_{-11 \pm 10} = N_{measured}\!\,^{+0.008\% \pm 0.007\%}_{-0.01\% \pm 0.01\%}

  2. Threshold: N_{disconnected}^{real} \simeq N_{measured}\!\,^{+3 \pm 2}_{-5 \pm 1} = N_{measured}\!\,^{+0.004\% \pm 0.003\%}_{-0.007\% \pm 0.001\%}

  • [50x50]:
  1. Crosstalk: N_{disconnected}^{real} \simeq N_{measured}\!\,^{+2 \pm 1}_{-42 \pm 1} = N_{measured}\!\,^{+0.003\% \pm 0.001\%}_{-0.05\% \pm 0.001\%}

  2. Threshold: N_{disconnected}^{real} \simeq N_{measured}\!\,^{+1 \pm 1}_{-16 \pm 11} = N_{measured}\!\,^{+0.001\% \pm 0.001\%}_{-0.02\% \pm 0.01\%}

*) (+)-error means average not recognized disconnected bumps, (-)-error means average not recognized connected bumps

Test results with skewed bump bonding [25umx100um sensor] (bump bond connection maps):

Source scan: Figure_02_Source_Scan

Discrepancies:

Connected with crosstalk, disconnected with source: Figure_03_Source_Scan_vs_WrongConnCrosstalk

Disconnected with crosstalk, connected with source: Figure_03_Source_Scan_vs_WrongDiscCrosstalk

Connected with threshold, disconnected with source: Figure_03_Source_Scan_vs_WrongConnThr

Disconnected with threshold, connected with source: Figure_03_Source_Scan_vs_WrongDiscThr

Simulation:

Simulation

Edited by Michael Daas

Merge request reports