Development of an imager with high time resolution optical photon counter

A. Sato; K. Hashiyama; M. Hasebe; M. Maeshiro; M. Miyahara; M. Shoji; R. Honda; R. Sato; T. Nakamori; T. Sato

REVIEW 3 major objections 5 minor 10 references

Development of an imager with high time resolution optical photon counter

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read The IMONY imager's new single-board readout timestamps all 64 pixels to 100 ns, with every channel passing initial dark-count and 500 Hz laser checks.

desk verdict A compact 64-channel photon-counting imager with real engineering value; the timing claim is believable but not yet measured end-to-end. read the letter →

arxiv 2506.07442 v1 pith:UOHJUO2Q submitted 2025-06-09 astro-ph.IM physics.ins-det

A. Sato , T. Nakamori , M. Shoji , T. Sato , K. Hashiyama , M. Hasebe , M. Maeshiro , R. Sato

show 2 more authors

R. Honda M. Miyahara

This is my paper

classification astro-ph.IMphysics.ins-det

keywords photoncountingimagerGeiger-modeavalanchephotodiodeMPPCFGATIASICGNSStimestampingCrabpulsarfasttransientsfront-endelectronics

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

Astrophysical transients such as Crab giant radio pulses last from nanoseconds to microseconds, and catching any coincident optical emission needs a photon detector with comparable time resolution. IMONY is such a detector: a 64-pixel Geiger-mode avalanche photodiode imager that counts single photons. This paper presents the upgraded readout, where four 16-channel FGATI ASICs and one FPGA replace four separate boards, and where every detected photon receives a timestamp with $100\,\mathrm{ns}$ resolution from a GNSS-locked clock. The initial laboratory evaluation shows that all 64 channels produce dark counts when shielded from light and that all 64 channels detect $500\,\mathrm{Hz}$ periodic picosecond laser pulses, with Fourier peaks at the expected frequency. The paper's claim is that these results confirm the pulse detection and time-stamping functions of all 64 channels work correctly, clearing the way for telescope observations of fast optical phenomena.

What carries the argument

The mechanism that carries the argument is the matched frontend-to-timestamp chain: a monolithic 64-pixel Geiger-mode avalanche photodiode array, whose pixel sizes are 75, 100, 150, and $200\,\mu\mathrm{m}$, is read out channel by channel into four FGATI ASICs, each a 16-channel analog chip with a fast transimpedance amplifier and comparator that converts each photoelectron into an LVDS pulse. A single FPGA samples the 64 LVDS lines every $5\,\mathrm{ns}$, generates a data packet only when a hit occurs, and records time via two counters: a pulse-per-second counter from the GNSS receiver and a $10\,\mathrm{MHz}$ counter that resets on each PPS, giving $100\,\mathrm{ns}$ timestamp resolution. The FPGA also controls high voltage and chip thresholds over SPI and streams the packets to a PC over Ethernet using SiTCP, replacing the previous four-board cabling with board-to-board connectors.

What would settle it

Trigger the $405\,\mathrm{nm}$ picosecond laser from a timing generator whose trigger edge is independently recorded with sub-10 ns accuracy, then histogram the difference between each IMONY timestamp and the trigger time; a spread wider than $100\,\mathrm{ns}$ (or a bias that drifts with the GNSS satellite constellation) would falsify the claimed timestamp resolution.

Watch

Extended reading notes

Core claim

The central claim is that a single integrated readout board for the IMONY imager gives all 64 pixels independent single-photon detection and time-stamping, with a timestamp resolution of $100\,\mathrm{ns}$ determined by the GNSS module's $10\,\mathrm{MHz}$ reference clock. Each pixel's analog pulse is amplified and converted to an LVDS hit pulse inside an FGATI ASIC, and the FPGA records the hit pixel together with a time read from 1 PPS and $10\,\mathrm{MHz}$ counters. Under light-shielded conditions every channel produced dark counts, and under uniform illumination by a $405\,\mathrm{nm}$ picosecond laser triggered at $500\,\mathrm{Hz}$, every channel showed a Fourier peak at $500\,\mathrm{Hz}$ and its first harmonic. The paper takes these observations as evidence that the sensor signals are processed and recorded correctly by the system, which is a necessary step before the imager can be used for simultaneous optical and radio observations of the Crab pulsar.

Load-bearing premise

The load-bearing premise is that the GNSS receiver's $10\,\mathrm{MHz}$ clock and PPS signal, together with the FPGA's counting logic, actually deliver the manufacturer-specified timing accuracy; the paper takes the $100\,\mathrm{ns}$ timestamp resolution from those specifications rather than from an end-to-end measurement.

Editorial extensions

If this is right

The instrument can, in principle, correlate each optical photon with a radio event at sub-millisecond precision, which is exactly the scale of Crab giant radio pulses.
The single-board, cable-free construction removes the unstable wiring that previously interrupted observations, shortening telescope setup time.
Because each channel is timestamped independently and simultaneously, the 8x8 array yields high-rate light curves for all pixels without channel dropout.
A GNSS-based time base means multiple IMONY units, on different telescopes or with different color filters, can be synchronized to a common absolute clock.
The compact stack could be mounted on small portable telescopes for occultation timing, where a fast, timestamped light curve is essential.

Reading between the lines

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

The paper establishes operation, not accuracy: the $100\,\mathrm{ns}$ claim rests on the GNSS module's manufacturer-specified jitter and the $10\,\mathrm{MHz}$ clock, so a measured end-to-end residual versus a known fast pulser would be the natural next experiment.
If the timing chain holds at the $100\,\mathrm{ns}$ level, the same FPGA design could scale to larger sensors or multiple synchronized modules, yielding a multi-color fast imager with each color channel on the same absolute time base.
The $5\,\mathrm{ns}$ FPGA polling interval with $100\,\mathrm{ns}$ timestamps implies a timing quantization structure that a Poisson light source could characterize; the paper does not report this.
The dark-count map shown uses preliminary thresholds, so optimizing per-channel thresholds could reduce noise and improve effective single-photon efficiency on sky.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Quick take: this is a clean, honest engineering report. The genuinely new piece is the single-board integration of the IMONY readout: four FGATI ASICs plus one FPGA on a 10x20 cm board stack, with GNSS time tagging and Ethernet readout. That integration is real and useful, and the paper shows all 64 channels produce dark counts and all 64 see the 500 Hz laser peaks. The work is a legitimate extension of the group's prior IMONY line, not a new physics result.

What it does well: the description of the board and signal flow is clear, the evaluation is sensible for a first lab check, and the authors are appropriately careful not to claim more than the data show in most places. The dark-count map and the 500 Hz Fourier peaks demonstrate that the front-end, FPGA, and readout chain work for all channels.

The soft spot is the timing claim. The 100 ns timestamp resolution is quoted from the 10 MHz counter quantization and the GNSS module's manufacturer-specified jitter. There is no end-to-end timing measurement on this new board, no comparison of the recorded photon times against a knowable external event with better than 100 ns accuracy. The 500 Hz laser test only proves that the system detects periodic light at a 2 ms period; it would pass with millisecond-level timing errors. The paper itself says quantitative evaluation is ongoing, yet the summary says the system 'can perform precise light curve measurements.' That sentence overstates what is shown. This is a minor blemish for an engineering report, not a fatal flaw, but it should be fixed before publication.

The other weakness is reproducibility: no firmware, schematics, or data are provided. For an instrument paper that is common, but it does limit how much an independent group could replicate.

Who this is for: people doing fast optical photometry, Crab GRP counterparts, occultations, multi-color fast work on small telescopes. It deserves a serious referee: the integration is real, the risk to the stated science case is clearly identified, and a good referee can push for an end-to-end timing measurement and a more measured summary. My recommendation: send it to review. The central engineering claim looks solid; the timing claim needs to be either measured or plainly labeled as a design goal.

Referee Report

3 major / 5 minor

Summary. The manuscript describes the development of a new integrated readout system for IMONY, a 64-pixel optical photon-counting imager based on Geiger-mode avalanche photodiodes. The new hardware replaces four separate 16-channel readout boards with a single three-board stack containing four FGATI ASICs, one FPGA, and a GNSS receiver, with Ethernet readout via an optical SFP module. The authors report initial laboratory tests: all 64 channels produce dark counts and all channels detect a picosecond laser pulsed at 500 Hz, showing Fourier peaks at 500 Hz and 1 kHz. They claim a timestamp resolution of 100 ns for each detected photon, determined by a 10 MHz reference clock from the GNSS module. The paper concludes that the system can perform precise light curve measurements and outlines future observational plans, including Crab pulsar giant-radio-pulse studies and occultation observations.

Significance. If the stated timing performance is confirmed, the integrated IMONY system would be a compact, 64-channel single-photon imager useful for sub-millisecond multi-wavelength transient studies, particularly optical follow-up of Crab giant radio pulses. The engineering integration is a genuine step forward over the previous four-board system, and the paper provides a clear block diagram, photographs, and reproducible functional tests. The dark-count map and the 500 Hz laser test provide credible evidence that all 64 channels are operational for pulse detection and readout. However, the central quantitative claim of 100 ns timestamping accuracy is not established by the reported measurements; the evaluation verifies counting and coarse periodicity but not the absolute or relative timing accuracy of individual photon timestamps. The manuscript's own text states that quantitative evaluation is ongoing, which conflicts with the summary claim of precise light-curve capability.

major comments (3)

[Section 5, Summary and future work] The concluding statement that the system 'can perform precise light curve measurements' is not supported by the reported evaluation. Section 4 demonstrates only that all 64 channels produce dark counts and that all channels show Fourier peaks at 500 Hz and its harmonic for a periodically triggered laser; these results verify event detection and coarse periodicity, not timing precision. Section 3 explicitly states that 'quantitative evaluation of the new integrated system is currently ongoing and will be reported in future work.' For the stated science goal of correlating optical photons with Crab giant radio pulses, the absolute timestamp accuracy is the load-bearing quantity, so either an end-to-end timing calibration must be presented or the summary claim must be qualified accordingly.
[Section 3, New system] The 100 ns timestamp resolution is inferred from the 10 MHz counter quantization and from the manufacturer's clock-jitter specification of the FURUNO GF-8803 GNSS receiver, but no measurement of the total timestamp error of the new readout chain is given. The error budget should include the FGATI propagation delay, comparator threshold-walk from pulse-height variation, FPGA sampling of asynchronous LVDS hits, and the phase alignment between the PPS signal and the 10 MHz counter. The 500 Hz laser test in Section 4 is insensitive to these effects because a periodic 2 ms signal would still produce a 500 Hz peak even with millisecond-level timing errors. I recommend adding a timing-calibration measurement, for example electrical test pulses injected at known delays relative to PPS or laser pulses detected simultaneously by a fast reference photodiode, and reporting per-channel timestamp residuals (mean, rms, and outlier fraction) rather than only the quantization step.
[Section 4, Evaluation] The sentence 'the pulse detection and time-stamping functions of all 64 channels are operating correctly' overstates what the dark-count map in Figure 6 shows. A count-rate map verifies that every channel produces pulses and that those pulses are read out, but it does not exercise the time-stamping path: the dark-count test could pass with all timestamps equal to a constant or with random timing. The timing function is separately claimed on the basis of the 500 Hz Fourier peaks, which again test periodicity, not the accuracy of individual timestamps. The functional distinction between detection, counting, and accurate time-tagging should be stated explicitly, and the evidence for each should be separated.

minor comments (5)

[Section 4, Figure 6 caption] The temperature range '25−6◦C' is ambiguous and should be written as a proper range, for example 25–26 °C, if that is what is meant.
[Section 3, New system] The sentence 'The data generation algorithm remains unchanged from the current version' should presumably refer to the previous version of the system, not the current one.
[Section 4, Figure 7] Only one example Fourier power spectrum is shown; please quantify the peak significance or report the peak amplitude or signal-to-noise ratio for each of the 64 channels to substantiate the claim that all channels detected the periodic pulses.
[Section 4, Evaluation] The threshold settings and sensor high voltage used for the dark-count and laser tests are not reported; giving these values would improve reproducibility.
[Abstract and Section 2] There are several missing spaces between words in the rendered text (for example 'andgiantradiopulsesfromtheCrabpulsar' in the abstract); these should be corrected in the published version.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the new board's performance is checked against an external laser and dark counts, and the 100 ns timing is an engineering assumption, not a derived prediction.

full rationale

The paper makes no claim that would reduce to its own inputs. The central laboratory claims—dark counts on all 64 channels and 500 Hz laser peaks detected on every channel—are evaluated against an external clock generator and light source, not against the system's own timestamping assumptions. The 100 ns timestamp resolution is stated as determined by the 10 MHz GNSS reference clock and manufacturer-specified jitter, which is an engineering specification cited from the vendor, not a fitted parameter or a derived prediction. References to the authors' prior work ([3,4,7]) describe the previous IMONY system and are used for context or for previously measured timing jitter of the old readout; they are not used to establish the new board's functionality. The paper explicitly notes that quantitative evaluation of the new integrated system is ongoing, which is a limitation in validation, not a circular step. The Fourier peak at 500 Hz and the dark-count map are independent empirical checks that would fail if the readout did not work. No equation is defined in terms of a target result, no fitted input is renamed as a prediction, and no load-bearing uniqueness or ansatz is imported via self-citation. Thus the derivation chain is self-contained and the circularity score is 0.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

No new physical entities or fitted model constants appear. The instrument's performance depends on operational settings and on the assumed behavior of a commercial sensor, an ASIC, and a GNSS module, all treated as black boxes.

free parameters (2)

FGATI comparator threshold per channel = not reported (preliminary, unoptimized)
Thresholds are adjusted by hand via SPI and directly determine which pulses are counted; the values used for Figures 5 through 7 are not given.
Sensor bias high voltage = not reported
The high voltage is adjusted to set Geiger operation and dark count rate; the operating point is not specified.

assumptions (3)

domain assumption Geiger-mode APD pixels produce one clear electrical pulse per detected photon.
Section 2 describes the sensor as delivering a few-nanosecond pulse with high signal-to-noise ratio; this underpins the single-photon counting claim and is not re-validated in this paper.
domain assumption FGATI ASIC amplifies and discriminates signals as specified.
Section 3 relies on the KEK ASIC [9] for all analog processing; no standalone ASIC calibration is presented.
domain assumption The FURUNO GF-8803 GNSS receiver supplies a 10 MHz clock and PPS with tens-of-nanoseconds jitter per manufacturer specification.
Section 3 sets the 100 ns timestamp resolution from this specification; it is not measured end-to-end in this paper.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Development of an imager with high time resolution optical photon counter." pith.science (2026). https://pith.science/paper/UOHJUO2Q

@misc{pith2026250607442,
  author       = {Pith},
  title        = {Pith review of: Development of an imager with high time resolution optical photon counter},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UOHJUO2Q}},
  note         = {Machine review of arXiv:2506.07442}
}

abstract

Astrophysical transient phenomena on sub-millisecond timescales, such as fast radio bursts and giant radio pulses from the Crab pulsar, have been primarily observed in radio wavebands. To investigate their origins, a photon detector with high sensitivity and high time resolution is required also in other wavelengths. Recently, we developed the Imager of MPPC-based Optical photoN counter from Yamagata (IMONY), an observation system utilizing a Geiger-mode avalanche photodiode (GAPD) as a sensor. The sensor consists of 64 pixels, each comprising a GAPD and a quenching resistor, with pixel sizes of 75, 100, 150, and 200\,$\mu$m. Each pixel signal is read out independently, enabling single-photon detection. After successfully observing the Crab pulsar using two Japanese telescopes, we upgraded the readout boards to achieve a more compact and stable system. The new system incorporates an analog application-specific integrated circuit (ASIC) developed at KEK for multi-purpose fast readout for silicon photomultipliers. This ASIC features a fast transimpedance amplifier and a comparator, independently processing 16 channels. A Global Navigation Satellite System receiver and a Field Programmable Gate Array (FPGA) provide timestamps for each detected photon with a resolution of 100 ns. The FPGA transmits the acquired data to a PC via Ethernet. This paper presents the details of the new system and the results of its initial evaluation.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

10 extracted references · 9 canonical work pages

[1]

J. M. Cordes, et al.,The Brightest Pulses in the Universe: Multifrequency Observations of the Crab Pulsar’s Giant Pulses,APJ612(2004) 375. [arXiv:astro-ph/0304495]

arXiv 2004
[2]

M. J. Strader, et al.,Excess Optical Enhancement Observed with ARCONS for Early Crab Giant Pulses,APJL779(2013) L12. [arXiv:1309.3270]

work page Pith review arXiv 2013
[3]

Development of an optical photon-counting imager with a monolithic Geiger APD array

T. Nakamori, et al.,Development of an optical photon-counting imager with a monolithic Geiger Avalanche Photodiode array,PASJ73(2021) 66. [arXiv:2010.11907]

work page Pith review arXiv 2021
[4]

Nakamori, et al.,Optical photon-counting observation of the Crab pulsar with Kanata telescope using prototype IMONY,PASJ(2025), in press [DOI:10.1093/pasj/psaf009]

T. Nakamori, et al.,Optical photon-counting observation of the Crab pulsar with Kanata telescope using prototype IMONY,PASJ(2025), in press [DOI:10.1093/pasj/psaf009]

work page doi:10.1093/pasj/psaf009 2025
[5]

Hashiyama, et al.,Imager of MPPC-based optical photon counter from Yamagata,Proc

K. Hashiyama, et al.,Imager of MPPC-based optical photon counter from Yamagata,Proc. SPIE 13103(2024) 1310314

work page 2024
[6]

Uchida,Hardware-Based TCP Processor for Gigabit Ethernet,IEEE Transactions on Nuclear Science55(2008) 1631

T. Uchida,Hardware-Based TCP Processor for Gigabit Ethernet,IEEE Transactions on Nuclear Science55(2008) 1631

work page 2008
[7]

M. Hasebe, et al.,Calibrating the photometric performance of a high-time-resolution photon-counting imager for optical astronomy, in the proceedings of the6th International Workshop on New Photon-Detectors (PD24), Vancouver, BC, Canada, November 19-21 (2024)

work page 2024
[8]

Hashiyama,Study of Pulsed Emission of the Crab Pulsar with a High Time Resolution Optical Photon Counting Imager, Ph.D

K. Hashiyama,Study of Pulsed Emission of the Crab Pulsar with a High Time Resolution Optical Photon Counting Imager, Ph.D. Thesis for the University of Tokyo (2025)

work page 2025

Show all 10 references

[9]

M. J. Kholili, et al.,A low-power and high-gain frontend for GHz application using trans-impedance amplifier for fast particle detection,JINST18(2023), P11010

2023
[10]

Inome, et al.,Quality Control of High-Speed Photon Detectors,EPJ Web of Conferences210 (2019), 05012

Y. Inome, et al.,Quality Control of High-Speed Photon Detectors,EPJ Web of Conferences210 (2019), 05012. – 6 –

2019

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

J. M. Cordes, et al.,The Brightest Pulses in the Universe: Multifrequency Observations of the Crab Pulsar’s Giant Pulses,APJ612(2004) 375. [arXiv:astro-ph/0304495]

arXiv 2004

[2] [2]

M. J. Strader, et al.,Excess Optical Enhancement Observed with ARCONS for Early Crab Giant Pulses,APJL779(2013) L12. [arXiv:1309.3270]

work page Pith review arXiv 2013

[3] [3]

Development of an optical photon-counting imager with a monolithic Geiger APD array

T. Nakamori, et al.,Development of an optical photon-counting imager with a monolithic Geiger Avalanche Photodiode array,PASJ73(2021) 66. [arXiv:2010.11907]

work page Pith review arXiv 2021

[4] [4]

Nakamori, et al.,Optical photon-counting observation of the Crab pulsar with Kanata telescope using prototype IMONY,PASJ(2025), in press [DOI:10.1093/pasj/psaf009]

T. Nakamori, et al.,Optical photon-counting observation of the Crab pulsar with Kanata telescope using prototype IMONY,PASJ(2025), in press [DOI:10.1093/pasj/psaf009]

work page doi:10.1093/pasj/psaf009 2025

[5] [5]

Hashiyama, et al.,Imager of MPPC-based optical photon counter from Yamagata,Proc

K. Hashiyama, et al.,Imager of MPPC-based optical photon counter from Yamagata,Proc. SPIE 13103(2024) 1310314

work page 2024

[6] [6]

Uchida,Hardware-Based TCP Processor for Gigabit Ethernet,IEEE Transactions on Nuclear Science55(2008) 1631

T. Uchida,Hardware-Based TCP Processor for Gigabit Ethernet,IEEE Transactions on Nuclear Science55(2008) 1631

work page 2008

[7] [7]

M. Hasebe, et al.,Calibrating the photometric performance of a high-time-resolution photon-counting imager for optical astronomy, in the proceedings of the6th International Workshop on New Photon-Detectors (PD24), Vancouver, BC, Canada, November 19-21 (2024)

work page 2024

[8] [8]

Hashiyama,Study of Pulsed Emission of the Crab Pulsar with a High Time Resolution Optical Photon Counting Imager, Ph.D

K. Hashiyama,Study of Pulsed Emission of the Crab Pulsar with a High Time Resolution Optical Photon Counting Imager, Ph.D. Thesis for the University of Tokyo (2025)

work page 2025