{"id":"b6b20bc5-3c34-4f80-a1bc-6b9afdfcff92","arxiv_id":"1908.08009","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Network time synchronization over Ethernet can synchronize multiple Pixie-Net radiation detector readout modules well enough to reconstruct coincidence events for radioxenon detection, down to sub-nanosecond precision.","lead":"A detector electronics team replaced dedicated clock cables with network time synchronization (IEEE 1588 PTP and synchronous Ethernet) for a radioxenon detector readout system, achieving measured timing precision from about 250 ns down to below 1 ns depending on network hardware. The work supports coincidence measurements with much less wiring in compact gas-detection systems.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central claim depends on the Dell 2216 switch's latency stability, which is not tested under final three-module operation, real network load, or extended runtime; a load-stress timing test is required.","rationale":"After reading the manuscript, the central claim indeed hinges on the Dell switch's empirical latency stability. The NetGear result demonstrates that PTP over a non-PTP switch is not intrinsically reliable; the Dell result shows one switch model worked in a short, low-load test. Scaling to the three-module system introduces more network traffic (data streaming from all modules), additional PTP slaves, and silicon detectors with different signal characteristics. None of these were tested. The paper acknowledges the need to test switch models but does not provide the long-term or load-stress data necessary to validate the deployment claim. A 72-hour pulser test with simultaneous data streaming directly measures the offset and jitter stability under the relevant conditions. If the offset remains within ~100 ns and FWHM below 500 ns, the concern is resolved; if not, the central claim is not supported. Secondary issues, such as the unquantified 3.3% ROI difference and the confusing 20 ps/130 ns timing statement, are real but less load-bearing; they affect precision of reporting, not the core feasibility argument. The reader's weakest assumption matches this assessment, so no verdict change is needed beyond the existing CONDITIONAL.","tokens_in":11039,"tokens_out":14758,"duration_ms":141900,"concrete_test":"Run a 72-hour timing test with the final three Pixie-Net modules and the Dell PowerConnect 2216 switch: feed a common pulser signal to one input on each module, enable PTP synchronization, and continuously stream list-mode data from all modules to a network drive at the design rate. Use LinuxPTP to record the DP83640 clock offset and compute the ΔT distribution between module pairs in 10-minute bins. The central claim is falsified if any 10-minute mean offset moves by more than 100 ns or any FWHM exceeds 500 ns, because that would make the 0.32 µs reconstruction window unable to capture all true coincidences without accepting background under real operating load.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim, stated as \"good PTP synchronization is sufficient (even with a non-PTP switch), and allow to reliably reconstruct true coincidence events, without any loss, nor by adding fortuitous coincidences,\" rests on a single empirical result: the Dell PowerConnect 2216 switch produced a 254 ns FWHM time accuracy and an ROI count within 3.3% of the shared-clock control in a 2-hour, two-module, NaI(Tl)-only test (Section III.B, Table II). This is not a protocol guarantee: the same PTP setup through a NetGear GS108 switch lost over 89% of true coincidences and introduced fortuitous ones (Fig. 12). The paper's own conclusion warns that \"switch characteristics significantly affects performance, it is important to select and test a suitable model,\" but the reported test does not cover the final deployment conditions: three Pixie-Net modules, silicon detectors (not yet manufactured at time of writing), sustained list-mode data streaming over the same switch, or operation beyond two hours. The Dell switch's latency stability under real network load and over extended periods is uncharacterized. If its offset drifts or its jitter grows, the fixed 0.32 µs reconstruction window will lose true coincidences or add background, invalidating the headline claim. This empirical stability assumption is the load-bearing point.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports an implementation of IEEE 1588 precision time protocol (PTP) and synchronous Ethernet (SyncE) on XIA Pixie-Net readout modules, with the goal of replacing dedicated clock and trigger cabling in a prototype radioxenon detection system at CEA. The authors characterize timing resolution for various network switches and synchronization methods using LaBr3 detectors and a pulse generator, and then perform 22Na coincidence measurements with two NaI(Tl) detectors. They report that PTP through a non-PTP Dell PowerConnect 2216 switch achieves 254 ns FWHM time accuracy and that the 511 keV coincidence ROI count differs by only 3.3% from a shared-clock control, leading to the conclusion that PTP synchronization is sufficient to reconstruct true coincidences without loss or added fortuitous coincidences. They also compare against a Geant4 simulation and report good agreement.","tokens_in":11278,"tokens_out":4731,"duration_ms":49484,"significance":"If the central engineering claim holds, the paper provides a useful practical result: for detector systems requiring a few hundred nanoseconds of synchronization, network-based PTP with an appropriately selected non-PTP switch can replace a dedicated clock distribution tree, reducing cabling and power consumption. The main strengths are the direct comparison against a shared-clock baseline, the use of an independent Geant4 simulation as a control, and the systematic survey of multiple switches and synchronization methods. The significance is moderate, however, because the headline conclusion rests on the empirical stability of a single switch model and on a two-module, NaI(Tl)-only test, so the generality of the claim is not yet established.","major_comments":[{"comment":"The central claim that PTP through a non-PTP switch allows reliable coincidence reconstruction 'without any loss, nor by adding fortuitous coincidences' is supported only by a single two-hour measurement with two Pixie-Net modules and two NaI(Tl) detectors using the Dell PowerConnect 2216 switch (Fig. 13, Table II). The same protocol through a NetGear GS108 switch loses more than 89% of true coincidences and adds fortuitous ones, so the reported behavior is an empirical property of one switch, not a property of PTP. The final CEA system will use three modules, silicon detectors, and sustained list-mode streaming, none of which is exercised in this test. Please either add a load-stress stability test in the final topology, or explicitly limit the validation claim to the tested configuration and state the qualification tests planned before deployment.","section":"Section III.B and Conclusions"},{"comment":"The text states that the shared-clock reference time resolution is ~20 ps, 'measured by the time stamp differences of every coincident pulse pair,' but Section II.C states that time stamps have 8 ns precision and Table II lists 129 ns for the shared-clock NaI(Tl) configuration. These statements are mutually inconsistent. If the 20 ps value was obtained from a constant-fraction waveform analysis rather than from time stamps, that must be stated explicitly; as written, the reference resolution is not reproducible.","section":"Section III.B and Section II.C"},{"comment":"The reported 3.3% difference between the Dell PTP ROI count (35,440) and the shared-clock ROI count (36,640) is quoted without uncertainty. From Poisson counting statistics alone, the difference of about 1,200 counts is roughly 4.5 sigma, so it is not statistical evidence of 'no loss.' Please report statistical and systematic uncertainties, and quantify the expected survival of true coincidences within the optimized 320 ns window given the measured 254 ns FWHM time distribution.","section":"Table II and Section III.B"},{"comment":"The 320 ns coincidence window was optimized post hoc for the Dell switch data, and the same data are then used to assess whether coincidences were lost or added. This creates a circularity in the claim of 'no fortuitous coincidences.' Reporting the ROI count as a function of window width (Fig. 14) is helpful, but the plateau should be quantified with uncertainties and, ideally, an independent background or random-coincidence estimate should be used to show that the chosen window lies in a stable region rather than in a locally favorable fluctuation.","section":"Fig. 13 and Fig. 14"}],"minor_comments":[{"comment":"The abstract states timing precisions 'between 300 ns and 200 ps,' while Section III.A reports ~250 ns for the best non-PTP switch; please make the numbers consistent or define whether 300 ns refers to a different measurement.","section":"Abstract and Section III.A"},{"comment":"The sentence 'All-PTP networks have better time resolution that non-PTP networks' contains a typo: 'that' should be 'than.'","section":"Section III.A"},{"comment":"The terms 'time resolution,' 'time accuracy,' and 'timing precision' are used interchangeably; a short definition of each would avoid ambiguity, especially in Table II where the first row is labeled 'Time resolution (ns).'","section":"Section III.B"},{"comment":"The switch identifiers [A] through [I] are defined only after the numbered reference list and are not integrated into the main reference list; consider citing them where they first appear in Section II.C or moving the definitions into the text.","section":"References"},{"comment":"The timing resolution values in Figs. 8-10 are presented as single points without error bars or a statement about the number of repetitions; adding this information would improve confidence in the comparisons across switches.","section":"Figures 8-10"}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid engineering/application note, and the control measurements against a shared clock and Geant4 simulation are appropriate. The main risk is that the headline conclusion is stated more strongly than the single-switch, two-module evidence supports; the editor may wish to ask the authors either to add a load-stress test for the final three-module system or to visibly narrow the scope of the validation claim. No citation or novelty concerns on my side."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nShort version: this is a solid engineering paper that demonstrates something genuinely useful — you can synchronize the ADC clocks in multiple Pixie-Net modules over ordinary Ethernet well enough to do real coincidence counting, without dedicated clock cables. The best result, a 254 ns FWHM time accuracy over a non-PTP switch with an ROI count within 3.3% of a shared-clock baseline, is believable and nicely anchored by comparison to Geant4.\n\nWhat is actually new: the Pixie-Net integration, driving the ADCs from the DP83640 PTP/SyncE-generated clock, and the systematic comparison across eight switches. That is the kind of practical data people building compact DAQ systems will want. The shared-clock and simulation cross-checks are the right way to validate the claim.\n\nSoft spots, in ascending order:\n\n1. Numbers often lack uncertainties — e.g., time resolutions in Fig. 10 and the 3.3% ROI difference. It is an engineering paper, but for such measurements a few error bars or standard deviations would help.\n\n2. The text says the shared-clock reference time resolution is ~20 ps, while earlier it says time stamps have 8 ns precision and that sub-stamp values come from constant fraction analysis. The paper should explain what quantity is actually being reported in the 20 ps number, otherwise it looks inconsistent.\n\n3. The larger issue: the \"reliably reconstruct true coincidences without loss\" conclusion rests on one non-PTP switch (Dell 2216) in a two-hour, two-module NaI(Tl)-only test. The same protocol through a NetGear GS108 lost almost all coincidences. The authors do honestly say switch choice matters, and the Dell result is encouraging, but the claimed reliability for the final three-module system with silicon detectors and sustained list-mode streaming is an extrapolation. A load test with the actual data traffic, or an explicit statement that this is a single-switch empirical result pending further testing, would be needed to fully support the conclusion.\n\nThese are fixable. The central technical approach holds up, and the paper deserves serious referee time. The caveats are typical for a prototype DAQ paper, not signs of a broken method.\n\nMy recommendation: send it to review, with a referee asked to push for uncertainty reporting and for the authors to either provide a network-load stress test or soften the general claim.","headline":"Solid engineering demonstration that PTP over commodity Ethernet can replace shared-clock cabling for compact DAQ, with a switch-dependent caveat that needs more testing.","tokens_in":11782,"tokens_out":2695,"would_cite":true,"duration_ms":26090,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Network time synchronization over ordinary Ethernet can replace dedicated clock cables in multi-module radiation detector readouts without losing true coincidence events.","keywords":["IEEE 1588 precision time protocol","precision time protocol (PTP)","synchronous Ethernet","network time synchronization","coincidence detection","radioxenon monitoring","digital data acquisition","radiation detector readout"],"falsifier":"Run the same 22Na coincidence measurement for an extended period (for example 24 hours) in the final three-module configuration while streaming data through the same non-PTP switch under realistic network load; if the time-difference FWHM grows beyond roughly 320 ns, or the 511 keV region-of-interest count deviates from the shared-clock control by more than about 5%, the sufficiency claim would be disproved.","tokens_in":10849,"feed_emoji":"⏱️","tokens_out":8442,"duration_ms":75904,"temperature":0.7,"pith_summary":"This paper reports that a multi-module radiation detector readout can be time-synchronized over the same Ethernet network that carries the data, eliminating dedicated clock and trigger cables. The authors implemented IEEE 1588 precision time protocol (PTP) and synchronous Ethernet (SyncE) in the readout electronics, and used them to synchronize three modules for a radioactive-gas detector prototype. With a non-PTP network switch that happened to have a small, stable time offset, measured time accuracy was 254 ns FWHM and the recovered 511 keV coincidence count differed from a shared-clock control by only 3.3%. The implication is that ordinary network hardware can meet the several-hundred-nanosecond coincidence windows needed for low-activity radioxenon monitoring, while PTP switches and SyncE provide progressively finer timing for more demanding setups.","feed_headline":"Ordinary switch keeps coincidence counts within 3.3% of shared clock","feed_subtitle":"IEEE 1588 over Ethernet synchronized three detector readout modules well enough to rebuild true coincidences without loss.","key_machinery":"The load-bearing mechanism is an Ethernet PHY that adds hardware timestamping of PTP packets and can output a locally generated but system-wide synchronized clock. That clock directly drives the ADCs and the FPGA pulse-processing logic, so event time stamps and digitized waveforms live in the same synchronized time base. An open-source PTP stack computes network delays and adjusts the local clock frequency, and synchronous Ethernet mode derives the local clock from the upstream link. What makes or breaks the application is the stability of the switch's internal delay: one non-PTP switch kept a small, nearly constant offset, while another showed a large and fluctuating offset. The detector-specific measurement method, constant-fraction timing on captured waveforms, provides the reference for judging these offsets.","core_discovery":"The central claim is that IEEE 1588 PTP, with hardware timestamping integrated into the data-acquisition clock path, is sufficient for reconstructing multi-detector coincidence events even when the network passes through a switch that does not support PTP. Timing precision of the synchronization, measured by the FWHM of the time-difference distribution between two modules, was about 6-10 ns on all-PTP networks, about 190-1000 ps with synchronous Ethernet and a pulser, and about 250 ns with one non-PTP switch. Using that switch for 22Na coincidence measurements with NaI(Tl) detectors, the total count in the 511 keV coincidence region of interest differed from the shared-clock setup by 3.3%, while a different non-PTP switch with a large, unstable offset lost most true coincidences and added fortuitous ones. The result is presented as validating network time synchronization as a replacement for shared clocks in this compact detector system, with the caveat that switch characteristics dominate performance and must be selected and tested.","pith_inferences":["A practical qualification rule for any candidate switch is to measure its time-offset drift over hours before trusting it; the deciding property here was offset stability, not the switch's PTP label.","Because the timing characterization used two modules and the detector system uses three, scaling to more synchronized modules could make switch-latency asymmetries accumulate; a multi-node timing test would show whether the 254 ns figure is preserved as the system grows.","The paper notes the silicon pixel detectors were still being manufactured when the coincidence tests were done, so the full electron/X-ray coincidence mode remains untested; the same synchronization should hold, but the optimal coincidence window and region-of-interest structure may shift once those detectors are installed.","The sub-nanosecond results with synchronous Ethernet suggest that network synchronization could reach into time-of-flight territory if dedicated low-jitter synchronization hardware is used, since the practical limit here was the detector response, not the network."],"forward_implications":["Dedicated clock and trigger distribution cabling can be omitted from multi-module detector systems whenever the coincidence window is several hundred nanoseconds or wider.","A detector system's network infrastructure can be chosen to match its timing needs: cheap non-PTP switches for coarse coincidence windows, PTP switches for about 10 ns resolution, and synchronous Ethernet for sub-nanosecond resolution.","PTP gives synchronized acquisition start at an absolute date and time with no extra cabling, whereas synchronous Ethernet synchronizes frequency only and still needs a start/stop signal.","With timestamped data buffered locally, centralized software-trigger decisions can be made over the network with round-trip times near 250 µs, enabling quasi-real-time global event selection without hard-wired trigger logic."],"supporting_citations":[{"why":"The non-PTP switch whose stable ~60 ns offset and 254 ns FWHM timing made full coincidence reconstruction possible.","marker":"[A]"},{"why":"The non-PTP switch whose large, unstable offset lost most true coincidences and added fortuitous ones, providing the negative control.","marker":"[C]"},{"why":"Supplies the constant-fraction waveform timing algorithm used for sub-8 ns time-difference measurements.","marker":"[6]"},{"why":"Defines the IEEE 1588 precision time protocol that the synchronization implementation is built on.","marker":"[9]"},{"why":"Documents the Ethernet PHY whose hardware timestamping and synchronized clock output connect PTP to the ADC/FPGA clocking.","marker":"[10]"},{"why":"Identifies the readout module in which the synchronization firmware is integrated.","marker":"[12]"},{"why":"Provides the Monte Carlo simulation that generated the control coincidence spectrum used as the comparison standard.","marker":"[26]"},{"why":"Supplies the open-source PTP stack that computes clock delays and adjusts the local clock frequency.","marker":"[30]"}],"fun_headline_variants":["Ethernet PTP syncs detectors, kills clock cable mess","IEEE 1588 network sync keeps true coincidences at 3.3%","PTP switch quality sets detector sync precision","Detector time alignment over Ethernet, down to 200 ps","Network sync replaces shared clock for detector readout"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central result depends on the tested non-PTP switch keeping its small, nearly constant time offset under real operating conditions; if that stability does not hold in the final three-module system, under extended operation, or under real network load, the claim that PTP over a non-PTP network is sufficient would no longer stand.","fun_headline_variants_meta":{"raw":{"variants":["Ethernet PTP syncs detectors, kills clock cable mess","IEEE 1588 network sync keeps true coincidences at 3.3%","PTP switch quality sets detector sync precision","Detector time alignment over Ethernet, down to 200 ps","Network sync replaces shared clock for detector readout"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000991,"raw_usage":{"total_tokens":4255,"prompt_tokens":1056,"completion_tokens":3199,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":672,"completion_tokens_details":{"reasoning_tokens":3115}},"tokens_in":672,"tokens_out":3199,"duration_ms":28411,"temperature":1.0,"reasoning_tokens":3115,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:51:18.889000+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same 22Na coincidence measurement for an extended period (for example 24 hours) in the final three-module configuration while streaming data through the same non-PTP switch under realistic network load; if the time-difference FWHM grows beyond roughly 320 ns, or the 511 keV region-of-interest count deviates from the shared-clock control by more than about 5%, the sufficiency claim would be disproved.","supporting_citations":[{"cited_title":"Development of 500 MHz Multi -Channel Readout Electronics for Fast Radiation Detectors","cited_arxiv_id":null,"evidence_quote":"Supplies the constant-fraction waveform timing algorithm used for sub-8 ns time-difference measurements."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the IEEE 1588 precision time protocol that the synchronization implementation is built on."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Documents the Ethernet PHY whose hardware timestamping and synchronized clock output connect PTP to the ADC/FPGA clocking."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Identifies the readout module in which the synchronization firmware is integrated."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the open-source PTP stack that computes clock delays and adjusts the local clock frequency."}],"review_version":1}