Speaker
Description
Scalable communication is needed for PEPS,a future gamma-ray array of 10
km2 with 55 solar-powered detector stations per km² at the Pierre Auger
Observatory. With the legacy access radio unavailable for PEPS, we propose a two-tier architecture: a 2.4 GHz FLRC fixed-slot first-hop layer from stations to a per-cell aggregation node, followed by 5 GHz backhaul to the DAQ. Separating access from backhaul allows aggregation-node placement near the cell center, reducing first-hop distances to 600–700 m with symmetric omnidirectional links. Preliminary 700 m LOS tests with 6-byte packets observed no loss at 650 kbps and <1% loss at 1.3 Mbps.
Summary (500 words)
PEPS aims to extend gamma-ray observations in the PeV energy range by deploying a dense array of water-Cherenkov detector stations at the Pierre Auger Observatory site. The communication system must be compatible with Auger station electronics while scaling from an initial 2 km² deployment to a possible 10 km² array, corresponding to approximately 110 to 550 stations at about 55 stations per km². The legacy Auger access radio is unavailable for the new readout system. The total data volume is modest: even 550 stations producing 1–2 kB/s each remain below a 10 Mbps-class backhaul load. The difficult part is not long-distance bandwidth, but low-power aggregation of many dense first-hop links. We propose a two-tier wireless readout cell architecture. The first tier uses Fast Long-Range Communication (FLRC), a packet mode of the SX1280 transceiver supporting Mbps-class operation in the 2.4 GHz ISM band, as a fixed-slot access layer between station-side bridge nodes and a local aggregation node. IThe second tier is 5 GHz commercial backhaul from the aggregation node to the DAQ. The 2.4 GHz band avoids coexistence with Auger’s existing 902–928 MHz communication system; sub-GHz alternatives such as Wi-Fi HaLow, while technically attractive, are disfavored by this constraint. This division keeps high-power long-distance networking hardware at the aggregation level instead of placing it at every station. Decoupling station-side access from long-distance backhaul also changes the geometry of the first hop. Instead of placing the first-hop receiver at a legacy tower or sector edge, the aggregation node can be placed according to radio geometry, for example near the center of a 1 km² cell. This reduces the worst-case first-hop distance to about 600–700 m and enables symmetric half-duplex links with omnidirectional antennas at both ends. For the test configuration, with approximately 19 dBm output power and 2 dBi antennas at both ends, the 700 m free-space path loss of about 97 dB leaves a link margin of roughly 20–30 dB depending on data rate, consistent with measured received powers around −70 dBm. The architecture adapts Auger-style deterministic access to a short-range cell geometry. Stations transmit in assigned time slots, while slot timing can be derived from the existing GPS time base already available at the detector stations. The station-side bridge translates the legacy UART data stream into scheduled FLRC packets; power characterization will follow after the firmware and duty-cycle strategy are finalized. Preliminary range validation was performed over a 700 m line-of-sight path using E28/SX1280-based FLRC modules configured around 19 dBm output power and 2 dBi omnidirectional antennas at both ends. At 650 kbps no packet loss was observed; at 1.3 Mbps packet loss was below 1%. Indoor non-line-of-sight tests through offices and corridor walls showed similar received signal levels and comparable packet-loss behavior. These results validate the first-hop distance scale and motivate longer tests with operational payload sizes. A two-station cell with one aggregation node and 5 GHz backhaul has been prepared for further testing. This work addresses a general instrumentation problem: scalable, low-power readout for dense autonomous detector arrays.