TAMBO Collaboration Meeting Summer 2026

.

As part of the development of the TAMBO-4 detector, the assembly of the panels provided by Harvard University was carried out. Data acquisition was performed using the RedPitaya platform, while data processing was conducted through a custom Python script. Initial tests were conducted under partial sealing conditions, which revealed a considerably high noise level. In response, a series of corrective actions were undertaken, including the complete hermetic sealing of all four panels, the optimization of the RedPitaya-based acquisition system, and the refinement of the Python processing script. Subsequent measurements demonstrated a significant reduction in background noise and a marked improvement in the overall quality of the acquired signals. Despite these enhancements, however, the signature of the minimum ionizing particle (MIP) peak remains elusive in our current analyses, indicating that further studies and refined detection strategies are required for its unambiguous identification.

In this talk, I will give an overview of the different hardware components of TAMBITO.

In this talk, I will present the Harvard group's work on scintillator detector R&D, with a particular focus on different design approaches to identifying the MIP peak.

The uDAQ system is currently under evaluation as a potential data acquisition solution for the TAMBO-4 detector prototype. This presentation details several technical challenges faced during initial testing and the procedures implemented to address them. First, the limited baudrate of the uDAQ-lite restricts large data transmission, which was solved by establishing a safety limit to stop operations upon overflow. Second, an anomalous peak observed near 3500 ADC was identified as a likely characteristic electronics artifact rather than a critical failure. Third, instances of binary misalignment and missing CPU triggers were managed by restricting data acquisition to smaller volumes. Finally, performance inconsistencies were observed where small panels ($20 \times 50\text{cm}$) failed while large panels ($185 \times 80\text{cm}$) remained operational. Given that the primary structural difference is the presence of an aluminium cage, this suggests the uDAQ system may be highly sensitive to the external environment.To conclude, we will present recent results acquired using the uDAQ system to demonstrate its current operational capabilities.

I will present preliminary laboratory measurements of the scintillation detectors using the uDAQs, along with some of the challenges encountered during these measurements.

The TAMBO project requires a communication system capable of reliably interconnecting a distributed network of detector nodes while satisfying demanding requirements in terms of data throughput, communication range, scalability, and energy consumption. The characteristics of the deployment environment and the expected number and distribution of nodes introduce additional challenges that must be considered when selecting and designing the communication architecture.

This presentation describes the ongoing development of the TAMBO communication system. First, the main communication challenges of the project and the resulting system requirements are presented. Based on these requirements, different wireless communication technologies are being evaluated within the proposed TAMBO communication architecture, with particular emphasis on conventional 2.4 GHz Wi-Fi and sub-GHz Wi-Fi HaLow for the main data communication network, together with LoRa as a complementary communication channel for system monitoring and low-rate telemetry.

Finally, preliminary experimental and analytical results obtained during the current development stage are presented. These results focus on three key performance indicators: achievable data throughput, communication coverage, and energy consumption. The preliminary findings provide an initial basis for comparing the proposed technologies and will guide the upcoming field tests and the final selection and integration of the TAMBO communication system.

Some possible methods for reducing required communications bandwidth and overall power consumption of scintillator panels will be described.

I will give a brief update on the muon+muon neutrino background implementation into TAMBOSim, although final event rates are still pending large simulation runs. I'll also discuss the TAMBOSim parallelization work I've been doing to speed up high-energy simulated event generation.

At 6.3 PeV, the Glashow resonance ($\bar{\nu}_e + e^- \to W^- \to \tau^-\bar{\nu}_\tau$) feeds taus into TAMBO's detection channel, making a nominally $\nu_\tau$-only telescope sensitive to $\bar{\nu}_e$. This opens the door to flavor-composition measurements at the PeV scale and to a neutrino-versus-antineutrino measurement. But is this measurement actually feasible in practice? We assess the realistic potential of isolating the $\bar{\nu}_e$ fraction from the dominant $\nu_\tau + \bar{\nu}_\tau$ background. Drawing on recent forecasting efforts, the underlying physics is well-established. Flavor-composition measurements could be viable if the PeV neutrino flux is high enough to yield tens of events. However, using it to distinguish neutrino production models like $pp$ vs. $p\gamma$ (the former producing more $\bar{\nu}_e$) remains out of reach. Feasibility hinges on a cascade of sensitive analytical choices: the assumed PeV flux, $\bar{\nu}_e$ acceptance modeling, shower-energy smearing, and the rigorous statistical treatment required near physical boundaries. Ultimately, we establish the specific analytical conditions required for TAMBO to successfully and honestly execute this measurement.

Earth-skimming tau neutrinos in the PeV–EeV range can be caught two ways. Radio antennas (GRAND) spread cheaply over thousands of square kilometers, as the rarest, most energetic events demand, and register the shower's electromagnetic component. Particle detectors packed across a canyon (TAMBO) cover far less ground, reach well below the radio threshold, and sample the full particle content. They would also share the roads, the crew and the sky. The question is whether one mountain will hold both. It will. We present Oroscope (github.com/mbustama/oroscope), an open tool that answers by searching real terrain. It sweeps an elevation map for buildable ground, then looks out from each surviving patch: right target, right distance and direction, enough rock behind it? It scores candidates and reports where each rejected one was lost. One engine serves GRAND and TAMBO (and leaves room to add other detection technologies): different numbers, same questions. Co-location is settled by the ground itself, before any physics: a hillside has one steepness; the two experiments want different steepness. What each needs to see is in no conflict: one hillside serves both views. We have identified candidate ground in the Peruvian Andes: along the Colca Valley in Arequipa, in the Callejón de Huaylas in Áncash, and around Cajatambo in Lima. Where both arrays have been searched at full resolution, 73–81% of the particle array's usable ground also serves the radio array. Their masks look nothing alike: one large radio site, tens of small particle sites on steep ground inside it. The share moves with array design; the particle array limits it, and the radio array pays almost nothing.

The high energies and long distances travelled by astrophysical neutrinos make them an appealing probe of new physics, as exemplified by the many ways in which IceCube has been used to constrain scenarios. However, the need to distinguish a point source signal from many background events can limit the power of these tests. In this talk, I will outline ways in which TAMBO's very low background rate could enable it to provide complementary tests of some models.

Musings on cosmic ray science we could (should) do with TAMBO/TAMBITO.

As a first step in characterizing the background, we studied proton-initiated showers on the TAMBO-4 prototype, whose initial deployment site is the PUCP campus in Lima. The simulation was carried out using TAMBOSim, and a brief discussion of the preliminary results is presented, contrasting them with the results obtained using only C8.

In this talk, Zigfried Hampel-Arias will present a brief career overview from cosmic ray physicist to remote sensing machine learning subject matter expert. First, he will present in depth on the HAWC Observatory, the resulting work from his PhD and postdoc tenures. Then he will describe the ongoing experimental results and validation from a career phase change to R&D in the interest of the intelligence community and US government interests, while continuing to contribute to efforts with external collaborators.

Overview of simulation including recent changes, fixed bugs, and (hopefully) a full set of figures for the paper

Extensive air-shower simulations from primary cosmic rays are traditionally performed using computationally intensive Monte Carlo methods, consuming significant computational resources in astroparticle physics. To address this challenge, we propose a continuous normalizing flow matching model for the TAMBO experiment to accelerate shower simulations. Trained on high-energy showers generated with CORSIKA, the model reproduces key observables, including energy spectra and time distributions. To improve computational efficiency, we implement several new features, including block-sparse attention to enable faster and more memory-efficient training, a DDPM solver to accelerate the sampling process, and additional architectural and training modifications to improve training stability. This work represents a first step toward enabling efficient end-to-end array layout optimization while substantially reducing the computational cost and time required for air-shower simulations.

tba

As the TAMBO collaboration advances toward site selection and infrastructure planning, decision-making must integrate socioeconomic feasibility and long-term societal impact alongside detector physics performance. This presentation outlines the key contributions from Universidad del Pacífico across two core dimensions: regional socioeconomic mapping and formal Cost-Benefit Analysis (CBA) for high-energy research infrastructures. First, we analyze Human Development Index (HDI) indicators across candidate sites in the Caylloma province (Arequipa, Peru). We present mapped indicator distributions, discuss site recommendations for TAMBITO and propose a modular scalability model prioritizing operational safety and accessible deployment. Second, we establish a formal CBA methodology tailored to large-scale research infrastructures. We categorize and evaluate both direct and indirect social benefits (such as human capital formation, scientific output, and technological spillovers) to construct a quantitative framework for forecasting TAMBO's Net Present Value (NPV). Through practical case examples, we provide the collaboration with a rigorous mechanism to support institutional alignment, regional policy dialogue, and long-term funding justification.

In this talk, we will present an overview of the measurements made during the site visit along with our experience making these measurements in non-laboratory conditions.