New HV scan with beam( tuneVertex_2022 trg + daq + tpx + itpc + btow + etow + esmd + tof + mtd + gmt + fcs + stgc + fst + l4):

22361030: HV 2900

22361031: HV 2850

22361032: HV 2850

22361033: HV 2940

22361034: HV 2940

22361035: HV 2940

22361036: HV 2960

22361037: HV 2960

22361038: HV 3000

22361039: HV 3000

22361040: HV 2920

22361041: HV 2920

Hit selection: Timebin ( -70 < tb < 20 )
Cluster selection:
- Clusters with >= 1 strips (All) : d361_ge1_summary.pdf
- Clusters with >= 2 strips : d361_ge2_summary.pdf
- Clusters with >= 3 strips : d361_ge3_summary.pdf
- Clusters with >= 4 strips : d361_ge4_summary.pdf
- Clusters with >= 5 strips : d361_ge5_summary.pdf
- Clusters with >= 6 strips : d361_ge6_summary.pdf

A few observations:
- The trends *make sense* now - using data for all HV settings from a single fill with consistent beam conditions.
- Selecting clusters with several strips hit (>=4) shows a trend of +0.1 % efficiency  per volt (about 20% difference between 2800V to 3000 V)
- ADC saturation increases with HV at the rate of 0.02% / V, from 4% at 2850 V to 7% at 3000 V
- Number of strips per real cluster increases with HV at 0.003 / V so you will have on average 0.6 more strips hit per cluster at 3000V compared to 2800 V.
 

NOTE: All of these measurements were taken with the stgc_n_sigma noise threshold of 15.