Analysis of FermiLab test beam data taken in May of 2007 , based on run #34
(zom in)
- Detector layout
- 3 GEM detectors labeled A,B,C
- 2 plains per detector labeled U,V
- U plane readout by 3 APV chips labeled 0,1,2, total 192 channels
chip U0: chan0-31 not connected, chan32-63 connected to active strips
chip U1: chan64-127 connected to active strips
chip U2: chan128-191 connected to active strips
- V plane readout by 3 APV
chip U0: chan1-63 connected to active strips
chip U1: chan64-127 connected to active strips
chip U2: chan128-191 connected to active strips
- Strip naming scheme: DPnnn, (D=detector, P=plane, nnn=channel), e.g.: AU031
- Example of U-plane raw ADC spectra for connected & unconnected strip for the same chip. Pedestal peak (& ped fit range) is defined by threshold set at Y>15 & width>10 chan.
- Example of V-plane raw ADC spectra
- Example of correlation between the same channels, U-plane
- Correlation for V-plane.
- Distribution of pedestal position and width for all channels from the U-plane. Pedestal QA criteria were set as:
- ped in [500,700]
- sigma in [7,30]
PDF for all channels in diskB U-plane pages1-15, V-=plane:pages 26-50. Channels with failed QA ar marked in red.
- U-plane pedestals: position & width. Good 190 of 192.
- V-plane pedestals: position & width. Good 125 of 192.
- All 2x192 spectra gated with strip BV73 to have ADC<20. PDF for all channels in diskB
Result: channels in the same chip have also ADC=0, but other chips have ~normal spectra.
- Apply chip-based pedestal correction
Method: drop bad strips, drop rawADC<=0, calculate average per chip, subtract,add 600.
- U-plane pedestals: position & width.
- V-plane pedestals: position & width.
- Iteration2: Drop few channels with large sigma: U64, U128, V0, V62-64
List of masked strips
PDF for all channels in diskB After iteration 2
- U-plane pedestals: position & width. Good 190 of 192.
- V-plane pedestals: position & width. Good 125 of 192.
- Examples of improved spectra, very good case in the moddle of a chip. Raw spectra in blue, after iteration 2 in green
And not so good, at the edges of a chip.
- Now is time to correct for capacitor address. Below is distribution of used addresses for one detector. No correlation w/ ADC is observed.
- capID vs. ADC for strip U34
- capID vs. ADC for strip U60