First and Second Level Pedestal Corrections

  The pedestal value of the ADC output is an effective offset. It is defined by the integration of the excess signal from the gate pulse which allows current to pass into the ADC. In order to correct for this offset, the pedestal values are measured and recorded periodically during data acquisition for each ADC channel. This is done by using a clock as the trigger, which avoids the acceptance of data signals from the various detectors. Then, during the data replay, these pedestals are subtracted from their respective ADC channels, on an event by event basis.

Following the first (absolute) pedestal subtraction, a second pedestal subtraction algorithm is implemented to remove low energy common mode noise from the data. The second pedestal subtraction algorithm works as follows:

• Groups of ADC channels are defined according to observed correlations in their noise spectra.
• For each event, the algorithm selects the channel from each group with the lowest ADC value. The detector associated with this channel is assumed not to have registered light from an electromagnetic shower.
• A noise threshold is defined which is equal to the lowest ADC value in each group plus a predefined offset.
• The events in each group which fall below this noise threshold are averaged, and then subtracted from each channel's ADC value in the group.

Figure 4.1 depicts a typical pedestal spectrum, before and after the second pedestal subtraction algorithm. One can see that noise level decreases from 40 channels (1.6 MeV) to about 5 channels (0.2 MeV).

  
Figure 4.1: CsI Pedestal spectra before and after second pedestal subtraction algorithm has been implemented.

Figure 4.2 shows the 44 CsI pedestal spectra, after the first and second pedestal subtraction have been implemented.

  
Figure 4.2: Final CsI pedestal spectra after both first and second pedestal corrections, viewed from two different angles.

By inspection of this histogram, one can see that a summing threshold of  channels (0.2 MeV) can be chosen to separate noise from data. All CsI ADC values which are below are considered to be noise and are ignored, while all that are above are summed as data.

The first and second pedestal subtraction algorithms are also applied to the 64-element NaI array, after which one can safely select  channels (0.4 MeV). Figure 4.3 shows the 64 NaI pedestal spectra after the first and second pedestal subtractions. This discrepancy between and is due to the fact that the ADC gate width for the NaI detectors was longer (800 ns) than that of the CsI detectors (100 ns). Hence, more low energy noise was integrated by the ADC for the NaI channels.

  
Figure 4.3: Final NaI pedestal spectra after first and second pedestal corrections, viewed from two different angles.

Finally, one can see the root mean square values for the corrected CsI and NaI pedestal histograms in Fig. 4.4. The root mean square value is calculated as

where is the ADC value of the ith bin in the pedestal histogram, is the mean, and N is the total number of events.

  
Figure 4.4: Root mean square values of the corrected CsI and NaI ADC pedestal histograms.