AIDA Event Builder - PNapi90/DESPEC-Analysis-Framework GitHub Wiki

The main task of the AIDA Event Builder (AEB) is the identification of possible beta decay events inside the AIDA MBS stream data dominated by roughly 90 % random noise. The identification of implantation events is already performed by the AIDA Unpacker. This task is done in three steps:

  • Forming of time strip clusters based on time differences (independent for x & y)
  • Forming of space strip cluster based on strip numbers (independent for x & y)
  • Forming of pixel clusters from space strip clusters for x & y

The x/y independent clustering steps (step 1 & 2) are performed since a direct pixel clustering would yield less good beta decay events (an asymmetry between the amount of fired x and y strips is very likely).

Time + space strip clusters

For each AIDA MBS stream data block (roughly 1000-7000 events), the events are sorted in respect to their x/y and z coordinates. For the setup of AIDA used in DESPEC, three layers of Si detectors are used with 128 strips in x and 128 strips in y direction each. This yields a total of six different x/y/z types of AIDA events. For each x/y/z, all respective events are compared using their measured times.


images/time_strip.png


A simple scenario for one x/y/z type is shown in the Figure above. The color of the fired strips emphasizes the measured time of the different strips. There are three clusters of "coincident" events. However, only the reddish and the blue clusters are also spatially adjacent. Only those cluster that have no space between their respective events are analyzed further.

XY Clustering

After the strips clusters are formed, x and y clusters are merged into XY clusters, if they are "coincident" (time difference in predefined range). An example for a "good" XY cluster is given in the Figure below (green).


images/Aida_good.png


Examples for bad clusters are shown in the other Figure (yellow). Those clusters are not considered further, since they are not adjacent.


images/bad.png


However, single strip clusters are already suppressed by the respective space clustering.

Remaining Events

Since it is also possible for beta decay events being split by the MBS stream, the latest events of each AIDA MBS stream data block are stored and added to the next data block.

Parallelization

Due to the large amount of data, the mentioned steps are calculated using parallelization. The technical details of this process can be found here.