Question of the Referees:
-----------
ADC resolution: Assuming a 12-bit dynamical range (instead of 13-bit),
what would be the resulting deterioration (e.g. in mass resolution, S/B)?
Would there be a noticeable impact on the physics results, and -if yes-
on which results?
-----------
The ADC range foreseen for the readout of the electromagnetic calorimeters of COMPASS is determined by the capabilities of the detector technique. The best electromagnetic shower reconstruction in a cellular calorimeter with a cell size of about one Moliere radius requires the analysis of a 5x5 matrix, i.e. 2 rings of cells around the cell with the maximal amplitude. This is essential for two shower separation where the shower shape deformation is used as an indication of the two overlapping signals. If a photon hits the centre of a calorimeter cell, the energy deposited in this cell is about 80% of the total and the energy deposited in each corner of the 5x5 matrix is about 0.1%. Hence a dynamic ADC range of 10 bits is desired for the appropriate shower reconstruction.
The maximum photon energy in the COMPASS ECALs depends on the beam momentum. It will be slightly below 300 GeV and 450 GeV for the first stage and the second stage of the experiment respectively. The minimum photon energy for which high precision measurement can be performed is defined by the light yield of the lead glass calorimeter, which is about 1 photo-electron per MeV of electromagnetic shower energy. If we assume that 20-30 photoelectrons is a minimal statistically significant signal which could be detected by the calorimeter cell, the energy down to 20-30 MeV can be measured in the corners of 5x5 matrix and photons down to 20-30 GeV can be analysed with a high precision. This requires an extra factor 10 in the dynamical range, so a 13 bit total ADC range is needed.
To summarize, an ADC readout with 13 bit range would let us make use of the
full potential of the electromagnetic lead glass calorimeters of COMPASS over
the whole range of photon energies.
The requirements of the various physics programs of COMPASS lead to a photon energy range from 0.5 GeV as lower limit to 200 GeV as upper limit. The current experiments on charm and central production - WA89 and WA102 - are using 12 bit ADC systems that are practically sufficient to cover the above photon energy range.
In conclusion, the choice between 13 bit or 12 bit dynamic ADC range
is not crucial for the performance of the electromagnetic calorimeters
of COMPASS. Practically all physics programs could be carried out with 12
bit dynamic range ADCs, but some improvements of the ECALs performance
with 13 bit range ADCs could be achieved. In particular a better
two-shower separation might be obtained over the wide energy
range. This would improve the acceptance for asymmetrical decays and
might be essential for high quality spin analysis.
The design of the ADC system, which is currently under development by the KEK and IHEP groups, is based on two parallel QVC channels with 10 bit linearity and different sensitivities. The difference between 12 bit and 13 bit dynamic range would only be a factor 4 or 8 between these sensitivities. The wider dynamic range does not affect significantly the overall ADC design and leads practically to the same cost per channel.
%