Skip to content

VD coldbox TP based trigger studies Dec 15 17, 2021

Jump to bottom
CharlieBatchelor edited this page Dec 20, 2021 · 32 revisions

Aim: to demonstrate TP generation and triggering based on TPs in the VD coldbox. Ideally to get horizontal muons (ie, those with a large number of channels hit).

Successful tests were run on 2021-12-16, in which triggers were generated based on TPs fed into the full trigger chain, which are subsequently processed by the horizontal muon trigger activity (TA) and trigger candidate (TC) makers. The TCs are then passed to a trigger decision (TD) maker, which decides the generation of a trigger.

Software Details

The working area /nfs/sw/dunedaq/dunedaq-v2.8.2-dqm-tpg-and-recording/work was used with packages checked out at these git hashes:

./detchannelmaps     dc49d73    * thea/pdsp_crate_fix
./dfmodules          4a8d263    * philiprodrigues/write-tpsets-via-serialization
./dqm                33f7a4d    * jmcarcell/new-fourier
./flxlibs            2b8298a    * patch/2.8.2
./minidaqapp         f38b44b    * philiprodrigues/vd-coldbox-tp-trigger
./readout            50aaa61    * philiprodrigues/use-channel-map
./trigger            be76fe1    * batchelor/horizontalMuon_2.8.2
./triggeralgs        6bca630    * batchelor/horizontalMuon_2.8.2

Configurations were generated for a range of n_channel_threshold values, as follows:

Good Runs

  • Run 12362 used channel threshold 10 and contains good events, with a small readout window
  • Run 12363 used channel threshold 10 and contains good events, with a larger readout window (30k timestamp ticks before and after, ie 1200 TPC samples)
  • Run 12385 used channel threshold 15 and contains a mix of empty fragments and good events, with large readout window.

Other runs had good triggers mixed with empty fragments due to some delay in the trigger chain that still needs to be investigated.

Trigger Chain

TP Generation

The TP generation was done with the software-based trigger primitive generation, as used in ProtoDUNE-SP I. This was based on a modified v2.8.2 version of the readout package. The main change was adding correct(?) channel mapping based on the VD coldbox channel map.

Various caveats:

  • Channels are split into collection and induction streams, based on the channel mapping in ProtoDUNE-SP. Then the TP finding is done on only the channels in the collection stream. As the VD coldbox channel mapping is quite different to ProtoDUNE-SP, this means that a subset of collection channels and a subset of induction channels are actually processed. An attempt was made to process all of the channels on the same thread, but np04-srv-030 (where the VD coldbox FELIX lives) was not able to keep up.
  • There appear to be some extremely noisy channels, which are not yet understood.

TA Generation

The horizontal muon TA generation monitors the incoming TPs over time with a configurable size window object, using the parameter: window_length. The basic idea is that a horizontally travelling muon will produce hits across many adjacent wire channels of the Coldbox APA, and when the number of channels hit exceeds the given configurable threshold, called n_channels_threshold, a TA is made and passed to the TC maker, the next level up in the hierarchy of trigger generation.

For Coldbox testing, the configuration was kept relaxed so that the adjacency checking functionality was not used. This means that the n_channels_threshold corresponds simply to the number of unique channels hit within the current window. That is, if the number of unique channels hit within the window exceeds the configured value of n_channels_threshold, a TA will be generated.

Coldbox Parameters

  • window_length - Was set to (400 frames x 25ns) = 10000 timestamp ticks.
  • n_channels_threshold - Triggers were successfully generated for values of 10 and 15.

TC Generation

Similarly, TC generation monitors the incoming TAs within a TC window object and issues a TC when a specified threshold such as adc_integral or n_channels_threhold is met. However, during the Coldbox tests, a trivial TC maker logic was used. This means that TCs are simply generated with a one-to-one correspondence with TAs.

The TC window size is determined by the start time of the earliest TA in the window and the end time of last TA in the TC window. For Coldbox testing, this simply corresponds to the start and end times of the single TA that is used to generate the TC. The window size of the TCs was therefore extended on both ends by 30000 ticks during testing to ensure all particle activity could be seen in the event display.

Clone this wiki locally