Experimental setup

General setup

The experiment setup consists of three STT modules placed above each other (copper-coloured trapezoids in the picture), a trigger system, two scintillators with attached photomultipliers (black structures in the picture) and readout electronics. The scintillators are placed one above and one below the modules and in a way that the shortest straws and the whole width of the modules are covered completely. The readout electronics digitizes the analog signal from the straws and send it to the lab PC for a further analysis. There is a software developed by the group for analysis purposes of the experiment on the PC

styx.png
© STYX

Trigger system

STYX needs a trigger system for three reasons:

  • The STT modules do not offer a self triggering mode and it would have been relatively difficult to be implemented. Therefore a an external trigger is needed to initiate the storage of interesting events.
  • The front-end logic of the STT modules create a time-multiplexed output signal aligning the readout of six channels with a relative spacing of 200 ns. To disentangle single straw signals from this pulse train a reference signal with a fixed timing relative to the physical particle is needed.
  • Straw tubes are detectors of the drift time type. For a precise measurement of the spatial coordinate of a traversing particle a very stable time reference has to be used to convert measured times of arrival of drifting electrons into drift times.

These three requirements can be satisfied setting up a scintillator trigger for cosmic muons. A trigger system has been set up to trigger on cosmic muons traversing the full setup. It comprises two 40×60 cm plastic scintillator panels coupled via light-guides to photomultiplier tubes. Both panels are oriented in a way that they cover the length of the shortest straw and the whole width of the STT modules. One panel is mounted above the top module, the other below the bottom module. 

  • Logic
    For triggering a coincidence of both panels is used. The analog output of the PMTs is discriminated using rise time compensating discriminators resulting in a good timing stability and low time walk. Careful adjustment of the relative timing ensures that the rising edge of the coincidence signal is always determined by the bottom panel to reduce jitter. For all steps standard NIM blocks are used.
  • Adaption to readout
    A custom trigger converter box has been built to adapt logic standards. The box is fed by a standard TTL signal and fans out 16 identical LVDS signals which can be used for triggering or time references.

Software

The software package is a collection of classes structured in libraries and additional macros. It can be divided in four main parts: calibration, reconstruction, analysis, simulation and additional tools. There are four main parts: Geometry containing all information about the geometry of the experiment, Event Structure defines the structure of the event (provides information from hit up to reconstructed track), Task Objects (simulation, reconstruction, analysis) provides data access and storage of information into root files, Event Display allows visualizing of the events, simulated or real.

The working principle of the simulation is at the moment: 

Generate some tracks with random slope and random position according to cos(theta2) distribution
Calculate the distance of this track to each straw center
If the position is smaller than the straw radius, store the distance as the drift radius.
Remarks: It is also possible to simulate more than one track per even

There are two kinds of calibration impelmented in the software, calibration on the electronics level and t0 - calibration working with data. Drift time distribution in each straw has a same shape, but a different starting point. It is caused by delays of the signal due to different cable length and the implemented error while (de)-multiplexing of the straws. The main purpose of a calibration to sift drift time distribution of each straw so that the distribution would start at the same point. Beside that, the calibration decides if a straw is working properly.

Test pulse calibration
A pulse generator is used to generate a NIM -signal. The signal is then distributed to the Front-End electronics and can be readout with TRB treating it like a normal cosmic data. With the test pulses the electronic components of the setup can be tested, for example looking for broken cables or unfunctional chips.

As a pulse allways have the same shape and position in readout window it can be used to perform a calibration on electronics level. For this purpose the peak position of the pulse in each straw is individually calculated and its distance to a set point in the time window determined. The peak position is then moved to this set point and the difference is stored and can be used in the following calibrations.

t0 - calibration
This calibration uses an actual cosmic data gathered over a longer time, for example over night. A high number of events is needed to be able to gather enough information for every straw. For each straw drift times from different events are stored and then filled in a histogram so that a distribution based on many different drift times is created. For this distribution start and end points are calculated and the whole distribution is shifted to the zero point of the readout window.

Information about straw quality and drift time shift are applied to a new data sample and the user can perform next steps, reconstruction and analysis.

The reconstruction package takes the output of the DAQ raw data (or the simulation package) to build an event structure with the hit, segment and track objects as data members which are stored in a root file which can be used for the analysis and read by the event display program.

The main steps of the reconstruction are

  • Use a flexible base class (SAnalysis) to handle the input, output and looping
  • Create an output root file with a root tree
  • Call a constructor for the event (SEvent) which contains arrays for the hits (SHit), segments (SSegment) and tracks (STrack) plus additional datamembers for the number of the objects and ids.
  • Make a loop over all event entries
  • Interpret the raw data as drift times and assign them to the according hit objects
  • Use the tracking algorithms, to build the segments and tracks
  • Store the event in the tree

Tracking

The tracking package reconstructs the tracks in the event. It offers two different algorithms to reconstruct tracks. Both of which rely on segments fitted in only one module. To create a segment two different segment algorithm can be used, 4 and 10 neighbours algorithms. Both find a segment seed, a hit in the middle layer. If there are neighbouring straws fired, use the straw which has more surrounding hits. The 4 closest straws, two in upper and two in bottom layers, to the central straw are used to fit the segment. An algorithm using the 10 nearest neighbors is also available.


In the last step segment candidates are fitted to the hits. The software uses random slope and position and fits the best fit parameters to the group of hits. The second segment is fitted to the same group of hits but, from the other side of the hits, so that the different possibilities for incident angle of the track are covered. 
There are also two different track reconstruction algorithms. One algorithms finds tracks in the following way:

Find all combinations of segments in the upper and the lower module. Calculate a slope of the track candidate using the central position of the two segments. Save the track candidate if the slope differences are smaller than a given cut value.


A second available algorithm does: 

Use all found segments as track seeds.
Prolongate the segments to the other module and combine found hits into a track if more than 2 hits are found.
The algorithm can be executed in both directions (upper module first or lower modules first).