A versatile digital camera trigger for telescopes in the Cherenkov Telescope Array

https://doi.org/10.1016/j.nima.2015.01.096Get rights and content

Abstract

This paper describes the concept of an FPGA-based digital camera trigger for imaging atmospheric Cherenkov telescopes, developed for the future Cherenkov Telescope Array (CTA). The proposed camera trigger is designed to select images initiated by the Cherenkov emission of extended air showers from very-high energy (VHE, E>20GeV) photons and charged particles while suppressing signatures from background light. The trigger comprises three stages. A first stage employs programmable discriminators to digitize the signals arriving from the camera channels (pixels). At the second stage, a grid of low-cost FPGAs is used to process the digitized signals for camera regions with 37 pixels. At the third stage, trigger conditions found independently in any of the overlapping 37-pixel regions are combined into a global camera trigger by few central FPGAs. Trigger prototype boards based on Xilinx FPGAs have been designed, built and tested and were shown to function properly. Using these components a full camera trigger with a power consumption and price per channel of about 0.5 W and 19, respectively, can be built. With the described design the camera trigger algorithm can take advantage of pixel information in both the space and the time domain allowing, for example, the creation of triggers sensitive to the time-gradient of a shower image; the time information could also be exploited to online adjust the time window of the acquisition system for pixel data. Combining the results of the parallel execution of different trigger algorithms (optimized, for example, for the lowest and highest energies, respectively) on each FPGA can result in a better response over all photons energies (as demonstrated by Monte Carlo simulation in this work).

Introduction

The proposed Cherenkov Telescope Array (CTA, [1]) is a large installation of Cherenkov telescopes of different sizes for the detection of very high-energy (VHE, E>20GeV) γ-rays. CTA will cover the energy range from few tens of GeV up to hundreds of TeV with a sensitivity at 1 TeV that is a factor of 10 better than achieved by the current-generation experiments H.E.S.S.,1 MAGIC,2 and VERITAS.3 It will also provide a good energy (about 10–15%) and angular resolution (on the arcmin scale) for reconstructed photons.

Currently considered array designs (see, for example, [2]) deploy a mixture of large-size telescopes (LSTs), medium-size telescopes (MSTs), and small-size telescopes (SSTs), with typical reflector diameters of 23 m, 12 m, and 4 m, respectively, on an area of roughly 1–10 km2 in order to ensure good performance over four orders of magnitude in photon energy. The enlarged instrumented area and telescope field of view (FOV) along with the wider energy range (when compared to current-generation experiments) imply a particular challenge for the trigger and data-acquisition systems which have to deal with a cosmic-ray-induced array trigger rate of O(10 kHz), typically an order of magnitude higher than in current installations.

CTA trigger designs typically comprise two trigger levels. At the telescope level, the proposed Cherenkov cameras (equipped with 1000–10,000 pixels with diameters corresponding to about 0.3°–0.1°) are expected to provide local camera triggers for γ-ray and cosmic-ray showers while efficiently suppressing the background. Typical γ-rays generate a Cherenkov light-flash of a few nanosecond duration in spatially neighbouring pixels, but at high energies (1TeV) and large impact distances of the shower with respect to the telescope the camera image acquires a substantial time-gradient [3] and can last several 10 ns. The background is dominated by the diffuse night-sky background (NSB) light from natural and artificial light sources, resulting in pixel count rates of O(100 MHz) at single photoelectron (p.e.) threshold, and by large-amplitude afterpulses mimicking Cherenkov signals in a pixel. Further suppression can be gained at the inter-telescope level where triggers can combine the information from spatially neighbouring telescopes or even from all telescopes in the array. Array-level or inter-telescope triggers typically require a coincidence of at least two telescopes in a time window of several 10 ns duration [4], [5] or make sure that the camera images are compatible with the origin from a γ-ray shower [6]. Such triggers reject, in particular, NSB triggers and events where a single muon from a hadronic shower hit a telescope and generated a camera trigger due to its Cherenkov emission.

At the telescope level, a versatile camera trigger is needed to select γ-rays over the full targeted energy range with good efficiency. Ideally, the trigger hardware should be applicable largely independent of the telescope type and the trigger should also provide guidance to the camera-readout system how much of the image should be kept (both in space and, in particular in the presence of a large time-gradient in the image, in time [7]). This paper describes the concept of a digital camera trigger based on Field Programmable Gate Arrays (FPGAs). Section 2 presents the design and a possible hardware implementation of the trigger. Section 3 discusses test results with prototype trigger boards. The impact of the proper choice of the camera trigger algorithms is illustrated in Section 4 using the results of a Monte Carlo simulation with the trigsim program (described in the appendix). Conclusions are presented in Section 5.

Section snippets

Camera trigger strategies

Camera trigger strategies employed in current-generation experiments [8], [9] are either based on the topological distribution of pixel hits (i.e. pixels with a signal above a discriminator threshold) or on the analogue sum of pixel signals. In the first approach (referred to as majority trigger) one requires at least Nmaj (for example 3) pixels above a certain threshold (a few p.e.) in a coincidence window of few nanosecond length. Such trigger designs differ in the definition of pixel groups

Trigger prototype boards and tests

Hardware implementations of the L0 and L1 stages were developed and studied in detail in the laboratory in order to show the validity and stability of the hardware concept. The development work was accompanied by extensive software tests using a VHDL Test Bench and the Xilinx ISE v13.4 software for design implementation and simulation. Particular attention has been paid to the accuracy of timing simulations the results of which have been verified by a comparison of FPGA signals (e.g. trigger

Simulated performance

In parallel to the design and test of trigger hardware, the performance of FPGA-based camera trigger algorithms has been investigated with the help of Monte Carlo simulations. A full comparison of the investigated trigger algorithms, the dependence of their performance on the NSB level, the telescope type, and the bandwidth of the used electronics are beyond the scope of this paper and are the subject of earlier [13], [14] and present work. The simulation results presented in the following are

Summary

The presented design of an FPGA-based digital trigger allows the triggering of cameras with an estimated latency between 80 and 300 ns, depending on complexity of the executed algorithm. The design grants the trigger algorithms access to overlapping camera regions that are large enough to contain a sizable fraction of a typical shower image. Each trigger algorithm can exploit the information for 37 pixels in its region in the space and time domain. Several trigger algorithms can be executed in

Acknowledgements

The authors would like to thank the members of the CTA consortium for stimulating discussions. We are grateful to the anonymous referee whose comments helped to improve the manuscript. Konrad Bernlöhr is acknowledged for help with the CTA simulation tools and for comments on Section 4. We also would like to thank Oscar Blanch Bigas for his valuable contributions to trigsim.

References (17)

  • B.S. Acharya

    Astroparticle Physics

    (2013)
  • K. Bernlöhr

    Astroparticle Physics

    (2013)
  • M. Heß

    Astroparticle Physics

    (1999)
  • S. Funk

    Astroparticle Physics

    (2004)
  • M. Meucci

    Nuclear Instruments and Methods in Physics Research A

    (2004)
  • K. Bernlöhr

    Astroparticle Physics

    (2008)
  • R. Brun et al.

    Nuclear Instruments and Methods in Physics Research A

    (1997)
  • G. Hermann

    American Institute of Physics Conference Series

    (2008)
There are more references available in the full text version of this article.

Cited by (16)

  • The NectarCAM timing system

    2023, Nuclear Instruments and Methods in Physics Research, Section A: Accelerators, Spectrometers, Detectors and Associated Equipment
  • A Trigger Interface Board for the Large and Medium Sized telescopes of the Cherenkov Telescope Array

    2022, Nuclear Instruments and Methods in Physics Research, Section A: Accelerators, Spectrometers, Detectors and Associated Equipment
  • Measurement of performance of the NectarCAM photodetectors

    2021, Nuclear Instruments and Methods in Physics Research, Section A: Accelerators, Spectrometers, Detectors and Associated Equipment
    Citation Excerpt :

    Each FEB amplifies, samples and digitizes the anode signals of FPMs (see Section 2) and transmits the resulting waveforms to the computer. The backends of the FEBs are connected to NectarCAM digital trigger board backplanes (DTB) that link the FEBs to the computer via ethernet cables and a hub (see [16]). The backplanes also transmit the external trigger signal produced by the flash LED pulser to the FEBs.

View all citing articles on Scopus
View full text