Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Status and performance results from NectarCAM -- a camera for CTA medium sized telescopes

T0 review · 4 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A modular Cherenkov camera calibration reaches 1.5% gain uniformity and 0.34 ns timing.

desk verdict Solid NectarCAM status report with first real performance numbers; the unvalidated shower-event claim is the main weakness, though the headline timing value does not depend on it. read the letter →

arxiv 1909.01969 v1 pith:FJUOOYE2 submitted 2019-09-04 astro-ph.IM astro-ph.HEhep-ex

Thomas Tavernier , Jean-François Glicenstein , François Brun This is my paper · ORCID
classification astro-ph.IMastro-ph.HEhep-ex
keywords NectarCAMCherenkovcameraPMTgaincalibrationsinglephotoelectronspectrumtimingtimeofmaximumimagingatmospherictelescopegamma-rayastronomy
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper describes how NectarCAM, a modular camera for the medium-sized telescopes of a future ground-based gamma-ray observatory, is calibrated and what its first performance numbers are. It reports that a two-component model of the single photoelectron response, with the component ratio fixed from measurements at higher voltages and extrapolated to the nominal operating voltage, brings the pixel-to-pixel gain dispersion to about $\sim 1.5\%$ after calibration. It further reports that timing offsets, traced to trigger-signal propagation delays and to the high-voltage dependence of photomultiplier transit time, can be corrected so that the spread of the mean time-of-maximum across the camera is about $0.34\ \mathrm{ns}$. With these calibrations, a one-month observation campaign on a prototype telescope structure recorded several hundred thousand atmospheric Cherenkov shower images. These numbers matter because uniform gain and timing across a camera are prerequisites for reconstructing gamma-ray shower images and for the array to reach its design sensitivity.

What carries the argument

The load-bearing object is the two-component single photoelectron spectrum used in the gain fit: a Gaussian peak plus a half-Gaussian centred at zero that accounts for the low-charge plateau seen at higher gains. Its component ratio is measured at high voltage and then frozen while fitting the multi-photoelectron spectrum at the nominal gain, which is what makes the absolute gain scale of every pixel consistent. The timing chain is carried by the maximum-likelihood time-of-maximum estimator, which slides a data-derived pulse-shape template against each waveform, together with two calibration maps: the trigger-accept delay per module and the HV-dependent transit-time correction per pixel. Together these convert raw ADC waveforms into photoelectron counts and arrival times with measured uniformity.

What would settle it

Take a NectarCAM module, illuminate it with a pulsed LED at an intensity below one photoelectron per pulse, record the charge spectrum at the nominal operating high voltage, and fit the two-component model with the component ratio left free; if the best-fit ratio differs from the frozen high-voltage value by more than the statistical uncertainty, the paper's gain calibration is biased by that difference.

Watch

Extended reading notes

Core claim

The central claim is that the NectarCAM camera, even in its partially equipped state, meets the uniformity requirements for imaging atmospheric Cherenkov astronomy. Using a double constrained Gaussian-Poisson model of the multi-photoelectron spectrum, in which the single photoelectron response is the sum of a Gaussian and a half-Gaussian low-charge component, the authors derive per-pixel gains and a gain-versus-high-voltage model; after calibration the relative gain dispersion is $\sigma_g/g_{\rm mean} \sim 0.015$. For timing, they estimate the time of maximum of each pulse by maximum likelihood against a pulse-shape template, then correct two identified sources of inter-pixel delay: the propagation of the trigger-accept signal to different modules, and the dependence of the photomultiplier transit time on the high voltage. After these corrections the standard deviation of the mean time-of-maximum across the camera is $0.34\ \mathrm{ns}$, improved from $0.43\ \mathrm{ns}$ after the trigger calibration alone. The paper also shows that shower images recorded under real night sky, including the time-of-maximum pattern along the shower-image axis, are usable for a standard shower analysis.

Load-bearing premise

The load-bearing assumption is that the two-component single photoelectron model, with its component ratio fixed from fits at higher high voltages, remains valid at the nominal high voltage; if the shape of the single photoelectron response changes with voltage in a way not captured by that extrapolation, the derived gains and all photoelectron counts would be biased.

Editorial extensions

If this is right

  • If the $1.5\%$ gain dispersion holds for the full 1855-pixel camera, charge reconstruction in photoelectrons will be limited by Poisson fluctuations and night-sky noise rather than by pixel-to-pixel gain scatter.
  • If the $0.34\ \mathrm{ns}$ timing spread holds, timing can be used as a cleaning cut to reject non-coincident pixels and to reconstruct the angular velocity of the shower image along the shower ellipse axis.
  • The sliding 16 ns integration window aligned on the pulse maximum captures the full signal of energetic showers while keeping the noise window short, so the calibrated gain directly improves the signal-to-noise ratio.
  • The success of the night-sky campaign, despite non-ideal observing conditions, indicates that a partially equipped camera can already be commissioned and produce a large sample of Cherenkov events for performance studies.
  • The calibration procedure—gain-versus-HV model, trigger-delay map, and transit-time correction—is modular and can be repeated for each of the 265 modules during full-camera integration.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A natural next test, not reported here, is to measure the single photoelectron spectrum at the nominal high voltage and check whether the component ratio frozen from high-voltage fits is actually constant; if it drifts, the absolute gain scale would shift by a measurable amount.
  • If the timing precision survives full-camera integration, time-ordered image cleaning could outperform amplitude-only tailcuts under high night-sky background, a comparison the paper leaves to future Monte Carlo studies.
  • The same double-Gaussian calibration could be exported to other photomultiplier-based Cherenkov cameras, since the low-charge plateau is a general PMT effect rather than a NectarCAM-specific failure.
  • The angular-velocity parameter fitted along the shower axis in shower data might serve as an independent discriminator between gamma-ray showers and cosmic-ray background, but the paper does not quantify its separation power.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. This conference proceedings paper reports the status and calibration results of NectarCAM, a camera for the medium-sized telescopes of the Cherenkov Telescope Array. It describes the modular hardware, the dark-room test bench at CEA Paris-Saclay, and a sky-observation campaign on the MST prototype in Adlershof. The main quantitative claims are a relative PMT gain dispersion of about 1.5% after calibration using a double-constrained Gaussian-Poisson model (Section 2), a timing dispersion σTOM of about 0.43 ns after trigger-accept delay calibration and about 0.34 ns after a high-voltage-dependent time-transfer correction (Section 3.2), and the statement that the Adlershof campaign yielded 'several hundred thousand of exploitable images from atmospheric Cherenkov events' (Sections 3.3 and 4). The paper also illustrates a basic shower analysis with timing information, including a Hillas fit and a fit of the angular velocity along the shower axis.

Significance. If the headline performance numbers are reliable, the paper provides a useful status update for a CTA camera component and demonstrates that the NectarCAM architecture can achieve sub-nanosecond timing and a few-percent gain uniformity in real operating conditions. The calibration equations (Eqs. 2.1, 2.2, and 3.1) are clearly written and follow standard practice for imaging atmospheric Cherenkov telescopes. The paper also credibly reports on a real-sky engineering campaign, which is an essential integration milestone. However, the load-bearing parts of the central claims are the extrapolation of the single-photoelectron model from high to nominal high voltage and the classification of the Adlershof events as atmospheric Cherenkov showers; both are asserted rather than demonstrated. The absence of uncertainties on the two headline numbers further limits the strength of the performance claims.

major comments (4)
  1. [Section 2, Eq. (2.2) and paragraph after Fig. 2] The paper claims that the two-component single photoelectron model is 'stable with the high voltage used and can be extrapolated to the nominal gain of the NectarCAM PMTs,' but no data, fit parameters, or statistical comparisons are given for the multiple high-voltage values. Since the double-constrained Gaussian shape is used to set the gain g for every pixel, a high-voltage-dependent change in the width or the ratio of the two components would directly bias all photoelectron counts in the shower analysis. Please show the fitted component parameters at several high-voltage points and a quantitative closure test, or explicitly restrict the calibration claim to the measured voltage range and estimate the resulting systematic uncertainty on the gain.
  2. [Section 3.2, Figs. 5 and 6] The headline timing results σTOM ≈ 0.43 ns and σTOM ≈ 0.34 ns are quoted without uncertainties, without the number of pixels or events used, and without a statement about whether the TOM distributions are consistent with a single Gaussian. Figure 5 and the projections in Figure 6 appear to show structured, multi-peaked distributions, so the reported standard deviation may not capture the actual spread or long tails. Please report the statistical and systematic errors on these widths, the fitted functional form, and the impact of the HV-correction fit shown in Figure 6.
  3. [Section 3.3 and Section 4] The statement that the Adlershof campaign produced 'several hundred thousand of exploitable images from atmospheric Cherenkov events' is not substantiated by the analysis presented. Figure 7 shows a single event with a Hillas ellipse, but there is no Monte Carlo comparison, no event-rate comparison with the expected cosmic-ray shower rate at the site, no explicit rejection of night-sky background, local lights, muons, or electronic noise, and no definition of 'exploitable.' Because the timing performance and the photoelectron counts in the shower-data section inherit this classification, the claimed sub-nanosecond timing under real-sky conditions is based on events of unproven atmospheric origin. Please provide quantitative event-selection criteria and validation, or soften the claim to 'candidate atmospheric Cherenkov events' and clearly state the caveat.
  4. [Section 3.1, Eq. (3.1)] The pulse shape fs(t) used for the TOM likelihood is described as 'extracted from real data with a charge of few tens of photo-electrons and normalised to the value of a single p.e.,' but the extraction method, the number of events, the choice of representative pixels, and the sensitivity of the TOM estimate to the shape choice are not specified. Since the timing calibration and the shower timing reconstruction both rely on this template, please document how fs(t) is built and provide a systematic check of the resulting TOM bias.
minor comments (5)
  1. [Section 2, first paragraph] Eq. (2.1) uses the symbol σ2ses with inconsistent spacing and the sentence defining the single photoelectron variance is grammatically incomplete; please rephrase and ensure all symbols are defined immediately after the equation.
  2. [Section 3.3, equation for Q] The summation limits are written as 'i = 10' and 'i = −6' in the display, which is confusing because the index is relative to the waveform maximum; please clarify that the window is from −6 to +10 time slots around imax.
  3. [Section 4, first sentence] 'NecatarCAM' is a typo for 'NectarCAM' in the conclusion.
  4. [Figure 4 and Figure 6] The figures lack axis labels in reduced form, and no uncertainties/error bars are shown on the histograms; please add labels and indicate the statistical precision.
  5. [Section 3.3, baseline subtraction] The description of the baseline subtraction ('floating average of each 60 samples of the readout window') is ambiguous; please specify whether the average is computed per event, per pixel, and over which time interval, and how the 60 samples relate to the readout window length.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the reported calibration and timing results are empirical instrument characterizations, not predictions that reduce to their own inputs.

full rationale

The paper's central results are instrument calibration metrics: PMT gains are fitted from LED multi-photoelectron spectra and reported as a dispersion, and timing offsets are calibrated with pulsed LED data and reported as residual spreads. Neither result is a 'prediction' that is equivalent to its input by construction. The relative gain dispersion of ~1.5% is the statistical spread of independently fitted pixel gains, not a quantity forced by the fit model. The TOM calibration similarly reports the spread of mean times after applying adjustable delays and an HV-dependent correction; this is a residual measurement, not a circular prediction. The two-component single-photoelectron model is extrapolated from higher high-voltage data to nominal high voltage using fixed ratios; this is an assumption that could bias the gains, but it is an external constraint from separate measurements, not a redefinition of the target result. The classification of Adlershof events as 'atmospheric Cherenkov events' is asserted without detailed validation (no Monte Carlo comparison or background rejection), which is an evidence gap about correctness, not a circular derivation. Self-citations in the paper refer to hardware design and trigger electronics and are not load-bearing for the measured performance values. Therefore no circular step is present, and the score is 0.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The central performance claims rest on a chain of calibration fits. The fit parameters are not published as values, and the two-component model's extrapolation to nominal voltage is asserted rather than demonstrated. No new physical entities are introduced.

free parameters (6)
  • PMT gain g per pixel = not specified (distribution in Fig. 4)
    Extracted from the multi-photoelectron spectrum fit; converts ADC counts to photoelectrons.
  • Light intensity lambda = not specified
    Mean number of photoelectrons per pulse, fitted in the Poisson-Gaussian model.
  • Single p.e. width sigma_ses = not specified
    Variance of the single photoelectron response, fitted and constrained via ratios at higher HV.
  • Pedestal mean x0 = not specified
    Pedestal offset fitted from the charge distribution.
  • Two-Gaussian component ratios = fixed from high-HV studies
    Ratio of standard deviations and sum between the two Gaussians is measured at higher voltage and fixed at nominal HV.
  • Per-module L1A trigger delays = adjusted per module, not tabulated
    Tuned to minimize the spread of mean time-of-maximum across the camera.
assumptions (4)
  • domain assumption The measured charge distribution is a Poisson-weighted convolution of a single photoelectron response with a Gaussian pedestal.
    Assumed in Eqs. 2.1 and 2.2 without independent verification for this PMT type.
  • ad hoc to paper The single photoelectron response can be described by a double constrained Gaussian model with parameters stable under high voltage.
    Introduced in Section 2; the paper states it is based on 'studies performed at larger values of high voltage' but does not provide supporting data or a reference for the extrapolation.
  • domain assumption The LED pulser produces a Poisson light source and the dark room flatness is sufficient for camera-scale calibration.
    Needed for the calibration runs; flatness was checked with an x,y table, but quantitative flatness values are not reported.
  • domain assumption The pulse shape template for time-of-maximum fitting is extracted from data and is representative across pixels and voltages.
    Assumed in Section 3.1; the template is not described in detail and its stability is not characterized.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Status and performance results from NectarCAM -- a camera for CTA medium sized telescopes." pith.science (2026). https://pith.science/paper/FJUOOYE2

@misc{pith2026190901969,
  author       = {Pith},
  title        = {Pith review of: Status and performance results from NectarCAM -- a camera for CTA medium sized telescopes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FJUOOYE2}},
  note         = {Machine review of arXiv:1909.01969}
}
read the original abstract

The Cherenkov Telescope Array (CTA) will be the first ground-based observatory for gamma-ray astronomy. With more than a hundred of 4th generation of Imaging Atmospheric Cherenkov Telescopes (IACTs) distributed in two large arrays, CTA will reach unprecedented sensitivity, angular resolution, and spectral coverage. Three classes of IACTs -- 40 Medium-Sized Telescopes (MSTs), 8 Large-Sized Telescopes (LSTs) and 70 Small-Sized Telescopes (SSTs) -- are required to cover the full CTA energy range (20 GeV to 300 TeV). NectarCAM is a Cherenkov camera which is designed to equip medium sized telescopes of CTA, covering the central energy range from 100 GeV to 30 TeV, with a field of view of 8 degrees. It is based on a modular design with data channels using the NECTAr chip, which is equipped with both GHz sampling Switched Capacitor Array and 12-bit Analog to Digital Converter (ADC). The camera will comprise 265 modules, each consisting of 7 photomultiplier Tubes (PMTs) and a Front-End Board performing the data capture, sending the data over the Ethernet after the trigger decision at rates up to 10 kHz. This contribution provides an overview of the status of the first NectarCAM camera currently under integration in CEA Paris-Saclay (France). Furthermore, we will discuss the calibration strategies and present performance results from the CEA Paris-Saclay test bench and from the first data taken under a real sky on the prototype of medium sized telescope (MST) structure in Adlershof (Germany).

Figures

Figures reproduced from arXiv: 1909.01969 by the authors.

Figure 1
Figure 1. The NectarCAM camera in the MST structure at Berlin-Adlershof. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Single p.e. spectrum using the double constrained Gaussian model. This shape is the one [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Pixel multi-electron spectrum. The charge is obtained by using an integration window [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Gain distributions of the camera PMT’s. figure (a) shows the 1D distribution and figure [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Distribution of the mean TOM in the camera before and after calibration. Each module [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: 2D distribution of the mean time of arrival in each pixel as a function the PMT high [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 7
Figure 7. Figure 7: Top left panel : photo-electron image of a [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. A Systematic Assessment of Data Volume Reduction for IACTs

    astro-ph.IM 2024-11 conditional novelty 5.0 of 10

    A DBSCAN time-based pixel selection retains more low-charge shower pixels than Tailcuts at matched data-reduction factors in CTAO simulations, with mixed noise-rejection performance.

Reference graph

Works this paper leans on

4 extracted references · 4 canonical work pages · cited by 1 Pith paper

  1. [1]

    J-F.Glicenstein, M.Shayduk et al., NectarCAM, a camera for the medium sized telescopes of the Cherenkov telescope array in proceedings of Heidelberg International Symposium on High Energy Gamma Ray Astronomy (Gamma 2016), 2016 [arXiv:1508.06555]

    work page Pith review arXiv 2016

  2. [2]

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

    U. Schwanke et al., A versatile digital camera trigger for telescopes in the Cherenkov Telescope Array, NIM A 782 (215), 92. [arXiv:1502.05506]

    work page Pith review arXiv

  3. [3]

    Peñil P., Tejedor L. et al. (2017). A trigger interface board to manage trigger and timing signals in cta large-sized telescope and medium-sized telescope cameras. In Proceedings of 35th ICRC, PoS(ICRC2017) 808, Busan, South Korea. [arXiv:1709.04692]

    work page Pith review arXiv 2017

  4. [4]

    Champion C., Punch M. et al. (2017). Ticks: A flexible white-rabbit based time-stamping board . In Proceedings of ICALEPCS 2017 . Barcelona (Spain).[arXiv:1710.07128] 7

    work page Pith review arXiv 2017

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.