Survey strip performance in 24 octants

Input: 550k events from run 1305045 and 2/3 of 1305051. HV set 3,  default APV tune, 30-70 gas mix

Method:

• ignora all P-strips, work only with R-strips (the arcs)
• tag pulse seeds in strips using my pethod A
• select the highest pulse (maxAdc) in given octant
• accumulate # of pulses of that type per strip (1D plot on the bottom)
• accumulat correlation plot for all other strips for which integral of ADC pulse is not smaller than 1/3 of tagged pulse. The auto-correlation is excluded (2D plot)

In the process I discarded some of the noisy octants on event-by-event basis. Pedestals and status tables were computed in fly based on the firts 200 events in given daq file. I have not used DB peds.

Fig 1. Example of both plots for the best octant 1B.S . (PDF for all 24 octants is attachement A) . The magenta lines mark approximate location of HV segment boundaries. The green line mark location of 2 arc-spacers. Neon-blue marks APV ranges.

Inefficiency around spacer and HV segment boundaries are well indetifiable.

All 24 octants were inspected and the following *gross* problems were identified. (no attempt of fixing hot strips was made- wil come next).

Fig 2 Key features of  gross problems:

Table 1

Summary of observed issues:

1. In 3 quadrants: 1D, 2A, 4B  there is one HV segment w/o voltage.
2. quad 3B has nonuniform yield vs. R - possible gas flow problem
3. significant correlation between R strips served by APV 0 & 4 observed in many quadrants, often limitted to HV segment boundaries
4. inefficiency around spacers is of 4-5 R strips (about 4mm)
5. Masking of hot strips will equalize the plots
6. few times more data needed for relaible study of strip mapping 

 Reference:

Fig 3 : HV segments and spacer grid

Fig 4 Range of R-strips served by 10 APVs in a quadrant