Correction Factor

Despite the fact that the overall effect tends to be neglible a correction factor should be applied to the data since the effect can be calculated. The above estimate is for the average case and a run by run correction would be preferable since rates were changing due to gain loss in the PMs over the whole run period.

The electronics dead-time can be calculated for each run using eq. . This number will be put into the scaler file and can be used to correct rates and yields in the same way the computer dead-time is used. The error of this correction is determined by the knowledge of the the proportionality factor k_el and the bar rates . In the following a short discussion how these values are determined and what there estimated error is.

A fit to the data taken during the test simulation gives a value . The main error associated with this value is the variation of gate length (depending on the overlap in the L/R coincidence) that occurred during the experiment. The gates varied typically between 50 and 60 ns, while the simulation was done at 60 ns gate width. Therefore we scale the measured k_el by 10% down and use with a 10% error assigned to it.

The bar rate can be calculated from the scaler values. These rates need to be corrected for beam-off periods. The fraction of beam-off period during a run can be calculated in the analyzer by setting a minimum threshold at the charge scaler value for each period between two scaler reads.

Another effect which is more complex and therefore harder to estimate is the following: The bar rate is not uniformly equal over the whole detector (the front planes counted higher then the planes further back). The same is true for the trigger rate: Protons are found predominantly in the top 5 bars of the first bar plane, while the neutron distribution peaks in the center of the detector. The impact of these non-uniformities on the electronics dead-time now depends on the details of the electronics setup, i.e. which detector channels are OR'ed together and the resulting correlation between trigger rate and bar rate in the part of the electronics which is causing the electronics dead-time. E.g. if one of the 10 lines (between Coinc 4516 and Disc PS706) has an above average trigger rate AND an above average bar rate then the actual electronics deadtime is higher than what is calculated from the average bar rate.
An estimate of this effect can be done by looking at the single rates of each photo tubes (available from asynchronous scaler readout), at the proton and neutron distribution (ntuple analysis) in the detector and at the exact electronics setup (from documentation). An exact evaluation would be rather complex and time consuming. Regarding the small size of electronics dead-time a rather crude estimate was done, leading to scaling factors of 1.02 and 1.18 for the neutron and the proton, respectively. The error assigned to this correction is 20%.

The following summarizes the correction procedure:

• Equation is used to calculate the electronics dead-time for both helicity states in each run.
• The value is used as proportionality factor for all runs.
• The bar rate is calculated from scaler values for each run.
• needs to be corrected for beam-off periods
• For the neutron needs to be scaled by a factor of 1.02.
• For the proton needs to be scaled by a factor of 1.18.
• The overall error assigned to the correction of electronics dead-time is 25%.
• The correction will be implemented in the analyzer and a number will be available in the scaler file for each run to correct yields for each helicity state.