A DLD hit generally consists of one MCP signal, followed by signals on each of the six anode channels. From these, time and position of the hit is reconstructed. The pulse assignment is done by slicing the digitizer trace saved by train into their corresponding pulse segments.
At the late boundary, this can unfortunately cause the anode signals to be cut off from its MCP signal and thus reconstruction fails. We noticed that there are essentially no hits within ~1/2 of the wire length at the end of every pulse. So far this has never been a problem, but due to the tight TOF window it became visible this time.
This MR changes the pulse slicing to use a "lookbehind" feature for anode channels, including a section corresponding to one wire length after after pulse for the purpose of reconstruction. Afterwards, any hits whose time is after the actual pulse window are removed, thus not creating any double hits.
How Has This Been Tested?
run = 38in_folder = '/gpfs/exfel/exp/SQS/202405/p006853/raw'calib_config_path = '/gpfs/exfel/exp/SQS/202405/p006853/usr/remi_proc_config_bottomions.yaml'
and on runs being taken right now.
Relevant Documents (optional)
Types of changes
New feature (non-breaking change which adds functionality)
