Pith. sign in

REVIEW 2 major objections 5 minor 15 references

Repurposing acquisition devices into trigger-based timing synchronization of breakdown events during MITICA high voltage holding experiments

T0 review · 2 major / 5 minor · reviewed 2026-07-10 · grok-4.5

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

Pith's one-line read Repurposed RedPitaya boards can reconstruct breakdown timing to 8 ns by recording trigger bit-fields and back-projecting static fiber and logic delays.

desk verdict Solid engineering methods paper: asynchronous THub timing from commodity RedPitayas works and delivers usable BD lags; residual physics interpretation is incomplete but does not sink the method. read the letter →

arxiv 2607.08501 v1 pith:T725OBS3 submitted 2026-07-09 physics.comp-ph

Andrea Rigoni Garola , Luca Lotto , Gabriele Manduchi , Tommaso Patton , Raffaele Ghiraldelli , Cesare Taliercio , Marco Boldrin , Luca Trevisan
show 3 more authors
Paolo Cinetto Paolo Barbato Giuseppe Chitarin
This is my paper · ORCID
classification physics.comp-ph
keywords breakdowntimingasynchronoussynchronizationRedPitayatriggerhubopticalfiberdelayMITICAtransientrecordervoltageholding
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

MITICA must hold 1 MV across vacuum gaps without damaging external components, so its voltage-holding tests need precise relative timing of the impulsive current discharges that mark breakdowns. Commercial high-precision clock networks were unavailable under the plant's insulation and supply constraints, so the authors turned ordinary low-cost digitizers into timing hubs that both distribute optical triggers and capture every trigger-out as a multi-channel waveform. Offline, each recorded edge is shifted by a once-measured static delay (fiber time-of-flight plus device logic) so that all probes share a common relative time base at the native 8 ns sample period. The method recovered reproducible lags of roughly 600 ns between the two transmission-line Rogowski coils and about 1.3 µs between the vessel-floor plates and the line sensors, giving a practical way to locate future arcs.

What carries the argument

The Timing Hub (THub): a RedPitaya whose ADC bus is rerouted to sample six LVDS trigger inputs as a bit-field transient; rising-edge sample counts yield raw arrival times that are corrected by the back-projection of Eq. (1) using once-measured fiber and logic offsets.

What would settle it

Repeat the round-trip and external-trigger delay measurements of Fig. 8 at several points during a long campaign; any drift larger than a few samples would leave residual offsets that destroy the reported 600 ns and 1.3 µs correlations.

Watch

Extended reading notes

Core claim

An asynchronous architecture that records the logical envelope of all trigger lines on a single RedPitaya and then subtracts the calibrated static delays of fiber and internal logic reconstructs true relative event times across a sparse, tree-connected set of electrically isolated digitizers to the 8 ns resolution of the 125 Msps clock, without requiring a distributed phase-locked time base.

Load-bearing premise

The total delay of every optical fiber and every device's internal trigger logic stays fixed for the whole multi-month campaign and can therefore be captured by a single static calibration.

Editorial extensions

If this is right

  • Any plant that already has optical trigger fibers can add relative timing at 8 ns resolution by inserting a few THubs and performing a one-time delay calibration.
  • Observed inter-probe lags become a diagnostic for discriminating in-vessel versus out-of-vessel breakdowns once the core-snubber contribution is fully modeled.
  • Tree topologies of THubs can be reconfigured after the shot; the reconstruction algorithm simply accumulates the parent-to-child delay path.
  • The same firmware pattern works for any transient recorder that can emit a trigger-out and sample multiple digital lines at its native rate.

Reading between the lines

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

  • The 1.3 µs GP00-to-Rogowski lag, if confirmed as core-snubber delay, supplies an independent estimate of the effective permeability of the FINEMET cores under real BD currents.
  • Because absolute time is supplied only by NTP on the root hub, the method can be upgraded to GPS or White-Rabbit absolute stamps without changing the relative-delay layer.
  • The unsupervised CAE-plus-GMM classifier used to reject false triggers is itself a reusable pre-filter for any sparse, high-rate BD archive.
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

2 major / 5 minor

Summary. The manuscript presents an asynchronous relative-timing architecture for synchronizing sparse, electrically insulated digitizers during MITICA high-voltage holding tests. RedPitaya boards are repurposed as Timing Hubs (THubs) that both multiplex optical triggers and record the multi-channel trigger bit-field as a 125 Msps transient. Offline back-projection of measured static delays (fiber TOF, electro-optic conversion, device logic Δ_tr/Δ_th, and THub processing t_p) via Eqs. (1)–(3) reconstructs event times across a tree of hubs and probes. The method is applied to three 2024–2025 campaigns; after unsupervised CAE+GMM cleaning of false triggers, cross-correlation of Class-A waveforms yields reproducible inter-probe lags (~600 ns HVD–TL1, few-ns GP00 internal, ~1.3 µs GP00–Rogowski). Absolute time is anchored by NTP on the root hub.

Significance. If the static-delay reconstruction holds, the work supplies a practical, low-cost alternative to PTP/White-Rabbit for transient localization in harsh HV environments where commercial synchronized digitizers are unavailable or electrically unsuitable. The approach re-uses an existing optical trigger fabric, is self-calibratable in principle, and is demonstrated on a full-scale ITER-relevant facility. Explicit strengths include the open delay equations, the multi-THub tree correction, the documented firmware features required for the method, and the honest discussion of residual physics discrepancies. These make the paper useful both for the fusion-diagnostics community and for other sparse, high-isolation acquisition networks.

major comments (2)
  1. [Sec. II-D, Eq. (1)–(3)] Sec. II-D and Eq. (1): The reconstruction t_true,i = t_s,i - (t_dl,i + t_p) is load-bearing and rests on the claim that every component of t_dl,i (fiber TOF, electro-optic stages, FPGA logic, and especially the threshold term Δ_th of Eq. (3)) remains constant to a few samples over the multi-month campaign. The paper reports a single manual calibration (450 ns / 140 ns fibers, Δ_tr = 100 ns, t_p = 80 ns) and never re-measures or bounds drift. A short stability check (repeat round-trip measurements at campaign start/mid/end, or a bound on temperature/voltage dependence of Δ_th) is needed to underwrite the 8 ns claim; without it the offline back-projection remains an unquantified engineering risk.
  2. [Sec. III-B, Figs. 12–14] Sec. III-B / Fig. 12–14: After filtering, only 156 Class-A pulses remain, ME10 data are excluded, and the GP00–Rogowski lag (~1.3 µs) substantially exceeds pure TL propagation (~270 ns). The core-snubber hypothesis is plausible but untested against a circuit model or additional probes. The central claim that the architecture enables BD localization therefore rests on a small, partially incomplete dataset whose largest lag is still under investigation. Either enlarge the validated set or qualify the localization claim more carefully (e.g., “inter-probe lag measurement” rather than full localization).
minor comments (5)
  1. [Sec. II-B, Fig. 4] Fig. 4 caption and surrounding text: the rerouting of the ADC bus to LVDS/TMDS33 ports is clear in principle, but a one-line statement of the resulting bit-to-port mapping (or a small table) would help readers reproduce the firmware change.
  2. [Sec. II-D, Eq. (3)] Eq. (3): Δ_d = 85 ns is stated for the RedPitaya; indicate whether this value was measured on the same boards used in the campaign or taken from a generic characterization.
  3. [Sec. III-A] Sec. III-A: the CAE loss weights (α=5, β=5, W=64) and the choice of four GMM classes are free hyper-parameters. A brief sensitivity note (or a statement that the Class-A selection is robust under modest changes) would strengthen confidence that the delay histograms are not classification artifacts.
  4. [Front matter / Refs.] Typographical: “Ghirardelli” vs “Ghiraldelli” in the author list/affiliations; “FINEMET TF-1Hmaterial” missing space; arXiv date and IEEE submission year (2026) should be consistent with journal style once accepted.
  5. [Sec. III, Fig. 9] Fig. 9 and the false-trigger discussion: a short quantitative statement of the fraction of events removed by the classifier would help the reader judge residual contamination risk in the 156-pulse set.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: independent static delay calibrations feed a linear back-projection whose outputs are measured inter-probe lags, not forced predictions.

full rationale

The paper's core reconstruction (Eqs. 1-3) subtracts measured fiber TOF, electro-optic, and logic latencies (450 ns / 140 ns round-trips, Δtr = 100 ns, tp = 80 ns, Δd = 85 ns) obtained by separate external-trigger or round-trip tests (Sec. II-D, Fig. 8) from the THub bit-field timestamps; those constants are never fitted to the Class-A BD waveforms whose relative lags (600 ns HVD-TL1, ~1.3 µs GP00-Rogowski) are later reported. The multi-THub tree correction is a pure rigid shift by the same measured parent-child delays. The 1-D Inception autoencoder and GMM clustering (Sec. III-A) serve only as an offline data-cleaning filter whose hyperparameters do not enter the timing equations. Self-citations ([6], [10], [12], [13]) supply supporting firmware or prior diagnostic experience but are not invoked as uniqueness theorems or load-bearing premises that close the derivation. The residual physics discrepancy with pure TL propagation is openly left for future modeling; nothing reduces by construction to its own inputs.

Assumptions & free parameters 6 free parameters · 4 assumptions · 1 invented entities

The central claim rests on a small set of domain assumptions about short-term clock stability and static delays, plus a handful of measured (not free-fitted) calibration numbers and standard engineering components. No new physical entities are postulated; the THub is a firmware reconfiguration of an existing commercial board.

free parameters (6)
  • fiber delay THub1–THub2 = 450 ns
    Set to 450 ns for the 100 m optical link after manual round-trip measurement; enters every multi-hub correction.
  • fiber delay THub2–devices = 140 ns
    Set to 140 ns for the 30 m links; used for Rogowski-coil channels.
  • Δtr (external-trigger device delay) = 100 ns
    Measured 100 ns; subtracted in Eq. (2) for externally triggered channels.
  • tp (THub internal processing delay) = 80 ns
    Measured 80 ns; common offset removed from all channels in Eq. (1).
  • Δd (RedPitaya threshold activation time) = 85 ns
    Measured 85 ns; appears in the sampling-frequency-dependent threshold delay of Eq. (3).
  • CAE loss weights α, β and window W = α=5, β=5, W=64
    Hand-chosen (α=5, β=5, W=64) for the spike-weighted reconstruction loss used only in the unsupervised cleaning stage, not in the timing equations themselves.
assumptions (4)
  • domain assumption Clock drift among independent RedPitaya oscillators is negligible over a single ~100 µs transient window.
    Stated in Sec. II-B as the justification for asynchronous operation; without it the sample-count differences would not map to physical time.
  • domain assumption All trigger-propagation delays (fiber TOF, electro-optic conversion, FPGA logic, threshold decision) are constant over the multi-month campaign and fully captured by the static calibration measurements of Fig. 8.
    Explicit premise of Sec. II-D; if any component drifts by more than a few samples the back-projection of Eq. (1) becomes biased.
  • ad hoc to paper A THub never re-propagates a trigger out the same port that received it, and the parent-port bit is excluded from the recorded bit-field.
    Required for loop-free tree topologies (Sec. II-C); enforced by configuration but not independently verified in the published data.
  • domain assumption The 1D Inception-style autoencoder + GMM pipeline correctly isolates true Class-A breakdown waveforms from false electrostatic-discharge triggers.
    Used to reduce 3100 pulses to the 156-event analysis set; misclassification would bias the reported delay histograms.
invented entities (1)
  • THub (Timing Hub) independent evidence
    purpose: Repurposed RedPitaya that simultaneously multiplexes optical triggers and records their bit-field evolution as a transient for offline delay reconstruction.
    Central hardware abstraction of the paper; it is a firmware and cabling configuration of a commercial board rather than a new physical object, and its behavior is fully described by the listed axioms and measured delays.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Repurposing acquisition devices into trigger-based timing synchronization of breakdown events during MITICA high voltage holding experiments." pith.science (2026). https://pith.science/paper/T725OBS3

@misc{pith2026260708501,
  author       = {Pith},
  title        = {Pith review of: Repurposing acquisition devices into trigger-based timing synchronization of breakdown events during MITICA high voltage holding experiments},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/T725OBS3}},
  note         = {Machine review of arXiv:2607.08501}
}
read the original abstract

A critical requirement for MITICA -- a full-scale prototype of the heating Neutral Beam Injectors hosted at the Consorzio RFX Neutral Beam Test Facility for the ITER experiment -- is the capability to withstand a continuous voltage of 1MV across the vacuum gaps insulating the beam source from the grounded vessel. To validate such feature, a dedicated voltage-holding test campaign was conducted throughout 2024 and 2025 using a full-scale mock-up of the beam source. The tests also involved an accurate characterization of the associated breakdown events: vacuum dielectric failures which result in rapid potential drops and generate strong current discharges. This contribution will present a relative time reconstruction architecture based on cost-effective, embedded RedPitaya (Zynq-7000 FPGA) devices repurposed as timing hubs. These nodes function as configurable trigger multiplexers while simultaneously recording trigger signals as transients to facilitate the offline reconstruction of event sequences. The method allows self-calibration through measuring the static intrinsic delays of the optical fibers and internal logics, generating delay offsets to synchronize acquired waveforms across a sparse, connected-graph topology of both acquisition devices and hubs themselves.

Figures

Figures reproduced from arXiv: 2607.08501 by the authors.

Figure 1
Figure 1. Sketch of the MITICA plant: on the left side a view of the beam source vessel (BSV) showing the source mock-up and the acceleration grid; the marks [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Graphical representation of the segments acquired in transient-recorder [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Synchronous method (a): all the devices that are triggered, or that collect triggers, share a synchronized clock (either internal or external). Examples include NTP, PTP, and White Rabbit. Asynchronous (b): the trigger is collected with respect to a single time reference and subsequently corrected for the signal TOF along the connecting components. The time reference is then kept aligned with absolute time. absolute… view at source ↗
Figures from the paper (10 more)
Figure 5
Figure 5. Figure 5: Scheme of the trigger connections for the asynchronous setup. [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]
Figure 6
Figure 6. Figure 6: Sketch of relative time reconstruction. The first timeline shows [PITH_FULL_IMAGE:figures/full_fig_p004_6.png]
Figure 8
Figure 8. Figure 8: Trigger calibration measurements that need to be performed: [PITH_FULL_IMAGE:figures/full_fig_p005_8.png]
Figure 7
Figure 7. Figure 7: Scheme of multiple THubs connected in a tree topology, where all [PITH_FULL_IMAGE:figures/full_fig_p005_7.png]
Figure 9
Figure 9. Figure 9: Example of a false BD event recorded by the RedPitaya boards in [PITH_FULL_IMAGE:figures/full_fig_p006_9.png]
Figure 10
Figure 10. Figure 10: 1D convolutional Inception-style autoencoder architecture: three [PITH_FULL_IMAGE:figures/full_fig_p006_10.png]
Figure 11
Figure 11. Figure 11: Event classes identified by the UMAP classifier. [PITH_FULL_IMAGE:figures/full_fig_p007_11.png]
Figure 12
Figure 12. Figure 12: Histograms of the time delay between the HVD and TL1 Rogowski [PITH_FULL_IMAGE:figures/full_fig_p008_12.png]
Figure 13
Figure 13. Figure 13: Example of resynchronized signals for the event 2024072308. (a) [PITH_FULL_IMAGE:figures/full_fig_p008_13.png]
Figure 14
Figure 14. Figure 14: Histogram of the time delay between the GP00 and Rogowski [PITH_FULL_IMAGE:figures/full_fig_p009_14.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

15 extracted references · 15 canonical work pages

  1. [1]

    The prima test facility: Spider and mitica test-beds for iter neutral beam injectors,

    V . Toigo, R. Piovanet al., “The prima test facility: Spider and mitica test-beds for iter neutral beam injectors,”New Journal of Physics, vol. 19, no. 8, p. 085004, aug 2017. [Online]. Available: https://doi.org/10.1088/1367-2630/aa78e8

    work page doi:10.1088/1367-2630/aa78e8 2017

  2. [2]

    Design of electrodes for high voltage tests in mitica,

    D. Aprile, G. Bertonet al., “Design of electrodes for high voltage tests in mitica,” in2023 30th International Symposium on Discharges and Electrical Insulation in Vacuum (ISDEIV), 2023, pp. 521–523

    work page 2023

  3. [3]

    A strategy to identify breakdown location in mitica test facility: results of high voltage test campaign,

    L. Zanotto, M. Boldrinet al., “A strategy to identify breakdown location in mitica test facility: results of high voltage test campaign,”Fusion Engineering and Design, vol. 187, p. 113381, 2023. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S0920379622003714

    work page 2023

  4. [4]

    Electrical diagnostics for high voltage tests in mitica,

    T. Patton, D. Aprileet al., “Electrical diagnostics for high voltage tests in mitica,”Fusion Engineering and Design, vol. 192, p. 113602, 2023. [Online]. Available: https://www.sciencedirect.com/science/article/pii/ S0920379623001862

    work page 2023

  5. [5]

    Modelling activity in support of mitica high voltage system protections,

    M. Dan, M. Boldrinet al., “Modelling activity in support of mitica high voltage system protections,”Fusion Engineering and Design, vol. 190, p. 113517, 2023. [Online]. Available: https: //www.sciencedirect.com/science/article/pii/S0920379623001011

    work page 2023

  6. [6]

    A zynq-based flexible adc architecture combining real-time data streaming and transient recording,

    A. R. Garola, G. Manduchiet al., “A zynq-based flexible adc architecture combining real-time data streaming and transient recording,”IEEE Transactions on Nuclear Science, vol. 68, pp. 245–249, 2018. [Online]. Available: https://api.semanticscholar.org/CorpusID:119222356

    work page 2018

  7. [7]

    Ieee standard for a pre- cision clock synchronization protocol for networked measurement and control systems,

    IEEE Instrument and Measurement Society, “Ieee standard for a pre- cision clock synchronization protocol for networked measurement and control systems,” 2008

    work page 2008

  8. [8]

    White rabbit - sensor/actuator protocol for the cern lhc particle accelerator,

    P. Loschmidt, G. Gadereret al., “White rabbit - sensor/actuator protocol for the cern lhc particle accelerator,” 11 2009, pp. 781 – 786

    work page 2009

Show all 15 references
  1. [9]

    Red pitaya stemlab development board,

    Red Pitaya, “Red pitaya stemlab development board,” https://redpitaya. com/, 2026

    2026

  2. [10]

    A mixed approach for clock synchro- nization in distributed data acquisition systems,

    G. Manduchi, A. Rigoniet al., “A mixed approach for clock synchro- nization in distributed data acquisition systems,”Sensors, vol. 24, no. 18, p. 6155, 2024

    2024

  3. [11]

    Going deeper with convolutions,

    C. Szegedy, W. Liuet al., “Going deeper with convolutions,” in2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015, pp. 1–9

    2015

  4. [12]

    Diagnostic data integration using deep neural networks for real-time plasma analysis,

    A. R. Garola, R. Cavazzanaet al., “Diagnostic data integration using deep neural networks for real-time plasma analysis,”IEEE Transactions on Nuclear Science, vol. 68, no. 8, pp. 2165–2172, 2021

    2021

  5. [13]

    Data reconstruction using varia- tional autoencoders and error analysis compared to b-spline interpola- tion,

    L. Orlandi, A. Rigoni Garolaet al., “Data reconstruction using varia- tional autoencoders and error analysis compared to b-spline interpola- tion,”Plasma Physics and Controlled Fusion, vol. 68, no. 2, p. 025028, 2026

    2026

  6. [14]

    Umap: Uniform manifold approximation and projection,

    L. McInnes, J. Healyet al., “Umap: Uniform manifold approximation and projection,”Journal of Open Source Software, vol. 3, no. 29, p. 861, 2018. [Online]. Available: https://doi.org/10.21105/joss.00861

    2018 doi

  7. [15]

    FINEMET® Nanocrystalline Cores: Common Mode Choke Cores,

    Proterial, Ltd. (Formerly Hitachi Metals, Ltd.), “FINEMET® Nanocrystalline Cores: Common Mode Choke Cores,” Proterial America, Ltd., Tech. Rep., 2026, official technical data and B- H curve specifications for FINEMET cores. [Online]. Available: https://www.proterial.com/e/prod...

    2026

Pith tools

Reviewed July 10, 2026 · model on record in the stance chip above.