{"id":"03c87c13-131d-48ca-9857-a02056e12bc7","arxiv_id":"2601.23106","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A 10-channel frequency-domain multiplexed TES readout achieves 156 ksps sampling and about 3 kHz of stable feedback bandwidth, meeting a key speed requirement for CUPID-style cryogenic calorimeters.","lead":"This paper demonstrates a frequency-multiplexed readout for superconducting transition-edge sensors that samples ten channels at 156 kHz, about a thousand times faster than the same hardware used in cosmic microwave background telescopes. The readout is aimed at future neutrinoless double-beta decay experiments, which need fast detectors to separate rare signal events from background pileup.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"3 kHz stable feedback bandwidth is demonstrated only on the dummy-resistor payload (Fig. 6); no equivalent measurement on the TES array in transition, so the abstract's 'real TES-based system' claim is unsupported.","rationale":"The paper's strongest claim is the 3 kHz stable feedback bandwidth, which is the quantitative requirement for CUPID-style pileup rejection. The evidence for this is the noise roll-off in Fig. 6, taken on a dummy load. The transferability to TESs is not trivial because the DAN loop gain depends on the SQUID input impedance, which is affected by the TES's electrothermal response and the parasitic line impedance. The reader's weakest_assumption identifies exactly this gap. I concur: the central claim is plausible but unverified for real TESs. The additional limitations (6/9 TESs transitioning, only 2 channels with light pulses, no energy resolution) further support a CONDITIONAL verdict. No internal inconsistency invalidates the design; the concern is about missing evidence for the headline claim. A direct bandwidth measurement with TESs in transition would settle it. Therefore I recommend keeping the verdict UNCHANGED (CONDITIONAL).","tokens_in":11958,"tokens_out":8752,"duration_ms":96112,"concrete_test":"Repeat the noise-spectrum and/or DAN network-analysis measurement with the 9-TES array biased in transition at 12 mK, with the same firmware and DAN gain settings used for the dummy-payload run. Extract the -3 dB roll-off frequency of the nuller-line noise (or of the closed-loop transfer function) for each transitioning channel. If the roll-off is below 3 kHz — or if the loop becomes unstable at the nominal gain — the headline bandwidth claim does not hold for real TESs.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that the readout delivers a stable 3 kHz feedback bandwidth in a real TES-based system is not supported by the presented data. In §5, the noise spectra showing a 3 kHz roll-off (Fig. 6) are taken with the 'dummy payload' of 0.5 Ω SMD resistors, not with the TES array. The only TES measurements are a DAN network analysis, load curves, and a light-pulse test on two of six transitioning TESs (Figs. 7–8); no closed-loop bandwidth or pulse-rise-time measurement is reported with TESs in transition. The paper itself notes loop-gain nonuniformity unique to high-bandwidth readout (§4.2) and a measured 50–75 mΩ parasitic line impedance that 'increases the rise times' (§5). Both effects alter the frequency-dependent loop gain relative to a resistive dummy, so the 3 kHz safe bandwidth and the implied ~120 μs rise-time capability may not transfer. The abstract overstates the evidence.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a 10-channel frequency-domain multiplexed (fMUX) readout for transition-edge sensors (TESs), based on the McGill/Berkeley ICEboard digital electronics. The authors describe new high-bandwidth firmware that samples detectors at 156 ksps, three orders of magnitude faster than the CMB-oriented SPT-3G system, and a redesigned LC resonator board. They report a DAN (digital active nulling) feedback loop with a claimed stable bandwidth of about 3 kHz, supported by noise spectra taken with a dummy payload of 0.5 Ω SMD resistors (Fig. 6). With an actual 9-TES array, they observe 6 TESs transitioning, measure load curves with parasitic line impedances of 50–75 mΩ, and demonstrate coincident optical light pulses on two TES channels (Fig. 8). The abstract concludes that the system demonstrates a stable feedback bandwidth of 3 kHz in a real TES-based system.","tokens_in":12240,"tokens_out":3754,"duration_ms":39601,"significance":"If the central bandwidth claim is sound, this work represents a meaningful step toward using multiplexed TES readout in next-generation cryogenic calorimeters such as CUPID. The system leverages mature, deployed hardware, achieves 10× multiplexing with a kHz-scale signal bandwidth, and identifies practical issues (resonator frequency shifts, parasitic impedance, loop-gain nonuniformity) that are directly relevant to the target application. The reported sampling rate increase and the explicit design for CUPID-like physics requirements are valuable contributions. However, the headline result—a stable 3 kHz feedback bandwidth in a real TES-based system—is currently supported only by a dummy-resistor measurement, and the manuscript does not provide a TES-in-transition bandwidth or pulse-shape measurement that would directly substantiate the abstract's claim. The strengths are the detailed hardware description, the open acknowledgment of observed nonidealities, and the clear statement of the system's scalability limits.","major_comments":[{"comment":"The claim 'demonstrates a stable feedback bandwidth of 3 kHz in a real TES-based system' is not supported by the presented data. The 3 kHz noise roll-off in Fig. 6 is measured with a 'dummy payload' of 0.5 Ω SMD resistors (stated at the start of §5), not with TESs biased in transition. The only TES measurements shown—network analysis, load curves, and the two-channel light-pulse time streams in Figs. 7–8—do not include a closed-loop bandwidth or pulse-rise-time measurement. The paper itself identifies effects that could change the loop gain with real TESs: loop-gain nonuniformity unique to high-bandwidth readout (§4.2) and a measured 50–75 mΩ parasitic line impedance that 'increases the rise times' (§5). These could reduce the safe bandwidth below 3 kHz when the TES array is operating. Please provide a TES-based bandwidth measurement (e.g., a recorded pulse with measured rise time, or a","section":"§5, Fig. 6, Abstract"},{"comment":"The statement 'the highest attained safe DAN readout bandwidth is around 3 kHz' (end of §4.2) is presented without an explicit measurement or derivation. Figure 6 shows the noise roll-off, but the correspondence between that roll-off and the DAN loop bandwidth is asserted rather than demonstrated. Since the decimation chain (CIC /64 and FIR /2) also shapes the passband, please indicate how the 3 kHz value was extracted (e.g., loop-gain measurement, step response, or fit to the noise roll-off) and confirm that the observed 3 kHz roll-off is not an artifact of the CIC/FIR filters. This is important because the central result hinges on this attribution.","section":"§4.2, §5"}],"minor_comments":[{"comment":"The phrase 'in a real TES-based system' overstates the evidence for the 3 kHz bandwidth, because the bandwidth measurement in §5 is performed with dummy resistors. Consider rephrasing to 'demonstrates a stable feedback bandwidth of 3 kHz with a resistive test payload' or adding a TES-based measurement.","section":"Abstract"},{"comment":"The figure caption should state explicitly that the noise spectra are taken with the 0.5 Ω SMD dummy payload, not with a TES array in transition. This will prevent misinterpretation.","section":"Figure 6 caption"},{"comment":"The sentence 'This allowed us to perform the basic multiplexing and noise characterization [12] of the system first without the complications related to operating TES detectors' is good, but the following sentence 'We have also measured the readout noise of the system with DAN enabled' should also mention that this measurement still uses the dummy payload.","section":"§5, first paragraph"},{"comment":"The stability criterion τ_TES > 5.8·τ_e is quoted from Ref. [20]. The derivation or a brief justification would improve self-containedness, but this is not essential for the present result.","section":"§3.2, Eq. (3.1)"},{"comment":"Only 6 of 9 TESs transitioned and light pulses are shown on only 2. This is noted in the text, but a brief discussion of why the other 3 TESs did not transition (e.g., contact resistance, wiring, or TES variability) would be useful for assessing system yield.","section":"§5, TES array"},{"comment":"The phrase 'three orders of magnitude faster' in the abstract is approximate (156 ksps vs. 153 sps is a factor of ~1020). The text is acceptable, but consider using 'three orders of magnitude' as an order-of-magnitude, not exact, statement.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"This is a solid engineering paper with a clear target application, but the abstract overstates the evidence for the headline 3 kHz bandwidth. The authors likely have the capability to add a TES-in-transition bandwidth/pulse measurement; if they do, the paper would be a strong contribution. If not, the abstract and conclusions must be softened. The self-citation to the companion noise paper (Ref. [12]) is appropriate, and the use of existing literature for stability/crosstalk formulas is standard. No concerns about novelty or scope for JINST."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a real engineering step forward, and the paper is worth refereeing. But the headline claim is not yet backed by the data: the 3 kHz loop bandwidth is measured with a resistive dummy payload, not with the TES array in transition. The abstract’s phrase “real TES-based system” goes beyond what the figures show.\n\nWhat’s new: they take the McGill/Berkeley digital fMUX used for CMB and rearchitect it for much higher bandwidth. Removing the polyphase filter banks cuts loop latency from ~14 μs to ~2.2 μs, the new decimation path gives 156.25 ksps per channel, and the DAN loop is the same scheme but at a higher rate. The cold hardware is also reworked: 10 LC resonators at 1–5 MHz, 4 μH inductors, with spacing chosen to keep leakage crosstalk below ~0.4%. All of this is described clearly and the firmware resource numbers are plausible.\n\nWhat works: the resonator network analysis with a dummy payload shows all ten resonances and a 3 kHz roll-off in the DAN noise spectra, consistent with the loop model. The light-pulse measurement on two TESs shows the full chain works end to end. The parasitic impedance from the long NbTi cabling is identified and quantified (50–75 mΩ), which is useful for the CUPID design. The paper is honest about known noise sources and refers to a companion noise paper.\n\nWhere it’s soft: the 3 kHz bandwidth is a dummy-payload result. With real TESs in transition you have a different dynamic impedance, plus the loop-gain nonuniformity the authors themselves flag in §4.2 and the extra rise time from the parasitic impedance. The only TES data are a network analysis, load curves, and light pulses on two of six transitioning devices. No closed-loop bandwidth or pulse rise-time measurement is reported on the TES array, so the claim that this meets the ~120 μs rise-time requirement is inferred, not demonstrated. The abstract overstates the evidence. The 6/9 transitioning is unexplained, and no data or firmware are released, though the latter is not unusual for this kind of paper.\n\nWho this is for: anyone working on CUPID-1T, future tonne-scale cryogenic calorimeters, or high-bandwidth fMUX for TESs. It will be most useful as a reference for the firmware architecture and the resonator design.\n\nRecommendation: send it to peer review, but require the authors to either soften the abstract to say the 3 kHz bandwidth is demonstrated with a dummy payload, or add a closed-loop bandwidth/pulse measurement with the TES array. With that, the work is a solid contribution.","headline":"A credible high-bandwidth fMUX demonstration for CUPID, but the 3 kHz stable bandwidth is shown on dummy resistors, not on the actual TES array; the abstract overstates the evidence.","tokens_in":12863,"tokens_out":3058,"would_cite":true,"duration_ms":29939,"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":"A frequency-domain multiplexed readout brings kHz-scale signal bandwidth to cryogenic TES detectors, enabling faster pulse discrimination for neutrinoless double beta decay searches.","keywords":["transition-edge sensors","frequency-domain multiplexing","cryogenic calorimeters","neutrinoless double beta decay","digital active nulling","SQUID readout","FPGA firmware","resonator multiplexing"],"falsifier":"Inject a small-signal sine or step into the SQUID summing junction while a real TES array is biased in the transition, and sweep frequency; if the DAN closed-loop bandwidth is not approximately 3 kHz across channels, or if antinulling appears at the Nyquist band edges with the real detectors, the central claim fails.","tokens_in":11885,"feed_emoji":"⚡","tokens_out":3017,"duration_ms":35182,"temperature":0.7,"pith_summary":"This paper demonstrates a frequency-domain multiplexed (fMUX) readout for transition-edge sensors (TESs) that raises the signal bandwidth into the kHz range, three orders of magnitude faster than the same electronics used in cosmic microwave background telescopes. The goal is to enable TES-based cryogenic calorimeters for neutrinoless double beta decay searches, where faster pulses would reject the dominating two-neutrino pileup background. The system combines 10 superconducting LC resonators, a SQUID amplifier, and FPGA-based digital active nulling to sample at 156 kHz with a stable feedback bandwidth of about 3 kHz, matching the roughly 120 microsecond rise time of the detectors. The paper also reports initial operation with an array of nine TESs, including detection of optical light pulses, and identifies parasitic impedance and loop-gain nonuniformity as current limitations.","feed_headline":"Multiplexed TES readout hits 3 kHz feedback bandwidth","feed_subtitle":"Faster THz-scale sampling rejects two-neutrino pileup in next-generation cryogenic calorimeters.","key_machinery":"The load-bearing mechanism is the Digital Active Nulling (DAN) feedback loop, implemented in FPGA firmware, which dynamically cancels the current through the SQUID input coil and encodes all science signal on a separate nuller line. The loop gain, set by the product of a programmable digital gain and the external loop gain, determines the signal bandwidth; the loop latency and loop-gain nonuniformity set the stability ceiling. The cold front end consists of 10 superconducting LC resonators in the 1–5 MHz range, a DC-SQUID array, and a resonator board mounted on the still stage, with the TESs on the mixing chamber. The new firmware increases the output data rate to 156 ksps by using wide chan","core_discovery":"The central claim is that a digital frequency-domain multiplexing readout can be adapted from low-bandwidth cosmology applications to the high-bandwidth needs of cryogenic calorimeters. By removing the polyphase filter banks used in the CMB firmware, reducing the multiplexing factor to 10, and lowering the loop latency to about 2.2 microseconds, the system achieves a stable Digital Active Nulling (DAN) feedback bandwidth of roughly 3 kHz. This bandwidth is demonstrated in noise spectra taken with a dummy resistor payload, where the white-noise roll-off at 3 kHz matches the expected detector rise time of about 120 microseconds. The paper argues that this meets the physics requirement for reje","pith_inferences":["If the loop-gain nonuniformity were compensated with a frequency-dependent DAN controller or a higher-order control loop, the stable bandwidth could likely extend beyond 3 kHz, enabling even faster TES signals or sharper pulse timing.","The same readout architecture could be adapted to other rare-event searches requiring fast, low-radioactivity multiplexed detectors, such as dark matter direct detection or coherent elastic neutrino-nucleus scattering.","A direct validation of the 3 kHz bandwidth claim would be a pulse rise-time measurement on all transitioning TESs; the paper demonstrates light pulses on only two devices, so a full-array timing test remains an open check.","The measured series parasitic impedance of 50–75 milliohms on the TES lines, which slows pulse rise times, suggests that improved cabling or connector design would directly enhance pileup rejection capabilities."],"forward_implications":["TES-based cryogenic calorimeters can reject two-neutrino double beta decay pileup, which is expected to contribute up to half of the background in the region of interest for tonne-scale searches.","The reduced multiplexing factor of 10 keeps the thermal load low and limits the number of channels lost on hardware or wiring failure, matching the architectural constraints of the CUPID cryostat.","The demonstrated 156 ksps sampling rate opens the door to pulse-shape discrimination techniques in rare-event searches.","The firmware uses only about 4% of FPGA LUTs and 2.7% of DSP blocks, so the design can scale to 8 modules (80 channels) once the Ethernet data offload bottleneck is removed.","The observed resonance frequency shift of about 4.5% from parasitic inductance is small enough to maintain crosstalk below 0.4% for most channel pairs, keeping the multiplexing scheme viable."],"fun_headline_variants":["Multiplexed TES readout reaches 3 kHz feedback","High-bandwidth TES readout for neutrinoless double beta","Fast digital fMUX enables 3 kHz TES feedback","TES readout 1000x faster with frequency-domain multiplexing","Stable 3 kHz bandwidth in multiplexed TES readout"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The 3 kHz stable feedback bandwidth is demonstrated with dummy resistor loads, and the paper assumes the same loop behavior carries over to the real transition-edge sensors despite observed loop-gain nonuniformity.","fun_headline_variants_meta":{"raw":{"variants":["Multiplexed TES readout reaches 3 kHz feedback","High-bandwidth TES readout for neutrinoless double beta","Fast digital fMUX enables 3 kHz TES feedback","TES readout 1000x faster with frequency-domain multiplexing","Stable 3 kHz bandwidth in multiplexed TES readout"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001446,"raw_usage":{"total_tokens":5715,"prompt_tokens":851,"completion_tokens":4864,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":595,"completion_tokens_details":{"reasoning_tokens":4777}},"tokens_in":595,"tokens_out":4864,"duration_ms":41258,"temperature":1.0,"reasoning_tokens":4777,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T06:11:29.877017+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Inject a small-signal sine or step into the SQUID summing junction while a real TES array is biased in the transition, and sweep frequency; if the DAN closed-loop bandwidth is not approximately 3 kHz across channels, or if antinulling appears at the Nyquist band edges with the real detectors, the central claim fails.","supporting_citations":[],"review_version":1}