On the hardware side, we studied the response of L0 boards containing discrete components and an ASIC to produce the L0 trigger of the system. We found that, although the discrete component trigger meets the requirements of the L0 trigger, the response obtained with the ASIC trigger is much more satisfactory in terms of linearity of the attenuator, clipping and DT. Besides that, the ASIC trigger provides lower weight, consumption and cost. This is the option that is currently being fully characterized to be installed in the first LST prototype.
We found that for the LSTs, which are built to collect more light and have a lower energy threshold, the performance of the sum trigger is better at energies below 100 GeV. Among the sum trigger options under consideration, the best compromise between the cut given by the
clipping and the suppression of the after pulses is given by a clipping level = 8 phe, although
it is worth pointing out that there was no real optimization of the clipping value. We simply selected a few clipping values (6 phe, 8 phe) and the performance was checked among them. Moreover these clipping values were obtained before the change in geometry of the telescopes and QE of the PMTs. For the new configuration larger clipping values may even improve the performance. The MSTs are not designed to have good sensitivity at low energies (<100 GeV). Hence, the performance of majority and sum triggers have large uncertainties in this energy region because of the low number of triggers. The results for both triggers at higher energies are compatible within the errors. A comparison between the aggressive and safe scenarios shows an improvement of the collection area for the aggressive scenario at the lowest energies, although within the errors. The energy threshold is also slightly improved by the use of the aggressive
option. We also found that for both trigger options, a cut in the minimum number of phe of 25
phe in the events recorded does not affect the results obtained. It is therefore safe for further
simulations to apply this cut and eliminate events with less than 25 phe without changing the results.
Nevertheless, as we are only considering galactic NSB, the DTs used to reproduce the safe and aggressive trigger scenarios considered are high, the collection area at low energies decreases and the energy threshold stays at a level > 20 GeV in every case. If we considered lower extra- galactic NSB, those DTs would be reduced, and therefore the energy threshold of the telescope would be lower as well.
Regarding CTA requirements and LST goals, the gain in energy threshold is small when moving from the 7.5 kHz requirement to the 15 kHz goal. The gain of moving from single- telescope to stereo observations is ∼30 % in energy threshold.
We also evaluated the impact of having PMTs with pulse widths wider than the goal of 2.6 ns. In general, we find that the worsening of the collection area is between 5 - 8 % every 0.2 ns with respect to that achieved with a pulse width of 2.6 ns. To be on the safe side and keep the
worsening of the performance at a level. 20 %, the pulse width of the PMTs should be smaller
than 3.0 ns.
All these results should be confirmed and optimized when the telescope parameters are fixed and the final site for the telescopes decided. This will be done in the Production-III MC run, focused on the selection of the final telescope layout and the optimization of the system.
4
The Topo-trigger: A new stereo trigger for
lowering the energy threshold of IACTs
The purpose of the hardware presented in this chapter is to decrease the energy threshold of the MAGIC telescopes without significantly increasing the data acquisition rate. To achieve this purpose, we developed an additional level of trigger that relies on the location in both MAGIC cameras where the trigger is issued to rule out accidental events. This allows to decrease the DT, which results in a reduction of the energy threshold of the instrument. We simulated the Topo-trigger concept using the standard MAGIC MC and tested it with real telescope data. In this chapter we show the concept and results of these tests.
4.1
Limitations of the trigger system in the MAGIC telescope
The trigger system in the MAGIC telescope is hardware limited at several stages. Every time an L1 trigger is issued, the L1 trigger system is busy for 100 ns, not accepting any other trigger in this time. This L1 trigger dead time is given by:
Dead time = L1 rate × 100 ns
In order not to lose > 2% of the cosmic ray events, we cannot accept L1 trigger rates larger than 200 kHz.
The rate of stereo accidental triggers is given by:
where L3 Coincidence window = 180 ns. The maximum stereo rate the current DAQ can record is ∼3 kHz (Tescaro et al. 2013). The simulated stereo accidental trigger rate (open squares) and measured stereo accidental trigger rate (filled circles) of MAGIC are shown in Figure 4.1. We are currently working at the crossing point between the extrapolation of the stereo cosmic ray trigger rate and the stereo accidental trigger rate (shown as the crossing point between the two red lines). What we aim to do with the algorithm we are presenting in this work is to reduce the accidental stereo trigger rate to 10% of its value (open circles). We would then move to operate to the crossing point between the extrapolation of the stereo cosmic ray rate and the 10% of the accidental stereo trigger rate (the crossing point between the two blue lines).
Discriminator Threshold [phe]
3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 Rate [Hz] 2 10 3 10
Measured stereo rate
Simulated stereo accidental rate Simulated stereo accidental rate (x0.1) Fit to measured stereo accidental rate Fit to measured stereo cosmic ray rate Fit to simulated stereo accidental rate Fit to stereo accidental rate (x0.1)
Figure 4.1: Measured and simulated stereo trigger rate for the MAGIC telescopes. The crossing point of the red lines determines the current operation point in the MAGIC telescope. The crossing point of the blue lines marks the operation point where we would go by decreasing the accidental rate to 10% of its value and maintaining the same rate recorded.