{"id":"27303bd6-20d4-42ee-96cc-a305bc92149d","arxiv_id":"2506.07710","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"This letter reports a 40.68-MHz active rectifier with 200-ns-settling delay compensation and a phase-based TX-side load monitoring scheme that cuts radiated power in implant WPT.","lead":"This paper describes a wireless power receiver for medical implants that aims to minimize tissue heating. It introduces a fast-settling active rectifier and a transmitter-side monitoring scheme that reduces radiated power by roughly 10 percent in measurements.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The TX-side PRE optimization rests on an unvalidated single-calibration phase setpoint: measured ΔΦ vs R_RX is shown for only one link condition, so the phase-to-R_RX mapping may not survive coupling or detuning changes.","rationale":"I read the paper in good faith and agree with the reader's identification of the weakest assumption. The rectifier itself is measured, and the 10% PRE improvement under one link condition is a real data point. But the proposed TX-side load monitoring is the part of the system that would justify the closed-loop PRE claim, and its correctness depends on a startup-calibrated phase setpoint remaining valid as the link changes. The paper does not show that ΔΦ_OPT is independent of coupling and detuning; in fact, the standard coupled-resonator analysis says the reflected impedance depends on both M and R_RX, so the phase is not a pure load readout unless special conditions hold. Since Fig. 7(e) shows only a single measured condition, the central load-monitoring claim is under-validated. I do not think this warrants rejection: the concept is plausible and the measured improvement is promising, but the missing validation across link variations is exactly the kind of condition that should be attached to acceptance. The reader's CONDITIONAL verdict already captures this, so I recommend no change. I would note that the simulated 200-ns settling number and the VCR inconsistency are real but presentation-level issues; they do not threaten the core rectifier operation the way the phase-calibration premise threatens the PRE/load-monitoring contribution.","tokens_in":6234,"tokens_out":4918,"duration_ms":65081,"concrete_test":"At a fixed R_RX sweep over the stated 120 Ω–1 kΩ range, measure ΔΦ versus R_RX at three coil separations (e.g., 8 mm, 12 mm, and 16 mm) and with the RX tank detuned by ±5% using the tunable capacitor. For each condition, determine ΔΦ_OPT by sweeping V_TX and locating the minimum V_TRANS amplitude or minimum PRAD/POUT. Then apply the nominal 12-mm ΔΦ_OPT as the closed-loop setpoint at the other link conditions and measure the resulting PRAD/POUT. If the optimal phase shifts by more than the phase-measurement resolution (e.g., >1°) or if the fixed setpoint deviates by more than 10% from the condition-specific PRE optimum, the single-calibration premise fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central system-level claim—that a one-time startup calibration of ΔΦ_OPT enables TX-side load monitoring and closed-loop PRE optimization—depends on ΔΦ being a single-valued, link-invariant proxy for R_RX. This is not established. In the coupled-resonator model, the TX-side phase depends on the reflected impedance ω²M²/Z_RX, which is a function of both the RX load R_RX and the mutual inductance M, and it also shifts with TX/RX detuning. The paper reports ΔΦ versus R_RX for only one measured link condition (12 mm separation, nominal tuning; Fig. 7(e)) and provides no evidence that either (i) ΔΦ_OPT is invariant across coupling and detuning or (ii) tracking a fixed ΔΦ_OPT reproduces the PRE optimum when the link changes. The 10% radiated-power improvement in Fig. 7(d) is likewise measured at a single link condition. If ΔΦ_OPT drifts with M or with tank detuning, the closed-loop regulator will hold a non-optimal R_RX and the claimed radiated-power benefit can shrink, vanish, or reverse. This is a correctness risk in the proposed load-monitoring concept itself, not merely a presentation issue. The simulated 200-ns settling claim and the VCR 93.9%/93.8% discrepancy are secondary; they concern labeling and consistency, whereas the phase-calibration premise is load-bearing for the PRE contribution.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The letter presents a 40.68 MHz wireless power transfer receiver for biomedical implants, with two main contributions. First, it describes a 40 nm CMOS active rectifier using a sample-and-accumulate voltage-domain ON/OFF delay compensation that is claimed to settle in a worst case of 200 ns, achieving a measured VCR of 93.9% and a post-layout simulated PCE of 90.1% in a 0.19 mm^2 area with on-chip resonant and filter capacitors. Second, it proposes a power radiated efficiency (PRE) optimization scheme in which a single startup calibration of the TX-side phase difference ΔΦ is used for TX-side load monitoring and closed-loop power regulation, with a measured ~10% radiated-power improvement compared with PTE optimization. The measured results include steady-state transients, a load-step response, VCR versus load, PRE/PTE versus load, and phase versus load, while the 200 ns settling and the 90.1% PCE are identified in the body as simulated.","tokens_in":6463,"tokens_out":5024,"duration_ms":60111,"significance":"If the claims are confirmed, the fast-settling delay-compensation technique would be a useful step forward for active rectifiers in implantable WPT, since it avoids a slow feedback loop and is compatible with on-chip integration of the resonant and filter capacitors. The PRE concept has genuine safety relevance because it targets radiated power into tissue rather than conventional PTE. The authors deserve credit for reporting measured VCR and load-step behavior, for disclosing that PCE could not be measured, and for providing a comparison table with state-of-the-art designs. The significance is tempered by two gaps: the headline 200 ns settling time is only simulated, and the closed-loop PRE regulator relies on an unvalidated assumption that a single startup phase calibration remains optimal as the link changes. The measured 10% PRE improvement is encouraging, but it is reported for only one link condition.","major_comments":[{"comment":"The closed-loop PRE regulator is built on the premise that the TX-side phase difference ΔΦ is proportional to R_RX and that a single startup calibration value ΔΦ_OPT remains optimal during operation. The paper provides no derivation of this mapping and shows only one measured link condition (12 mm separation, nominal tuning) in Fig. 7(e). In the coupled-resonator model the TX-side phase depends on the reflected impedance, which is a function of both the RX load and the mutual inductance, and it also shifts with tank detuning. Therefore a fixed ΔΦ_OPT may not correspond to the PRE optimum when the link changes, and the regulator could hold a non-optimal R_RX, causing the claimed radiated-power benefit to shrink or reverse. This is a load-bearing issue for the TX-side load-monitoring contribution. Please provide either an analytical derivation of the phase-load relationship including coupling and detuning dependence, or measurements at multiple separations and detuning conditions, or a calibration/update procedure that maintains validity; and report the PRE improvement across a range of link conditions.","section":"Section II and Fig. 7(e)"},{"comment":"The abstract and Table I present '200 ns' and 'fastest reported' settling as a headline result, but in Section III the value is explicitly described as 'simulated worst-case settling <200 ns.' The load-step measurement in Fig. 7(b) is limited by the output pole formed by the on-chip filter capacitor and the load (RC = 150 ns), so it cannot directly confirm the compensation-loop settling time. Please label the 200 ns value as simulated in the abstract and Table I, and provide either a direct measurement that isolates the compensation loop from the output filter or a simulation setup that explicitly separates the two, so that the reader can assess the claimed settling advantage.","section":"Section III, Section IV, Fig. 7(b)"}],"minor_comments":[{"comment":"The VCR values are inconsistent: the abstract states a measured VCR of 93.9%, Fig. 7(c) shows a peak of about 93.9% at a 700 Ω load, while Table I lists a peak VCR of 93.8% and a VCR of 92.0% at RL = 500 Ω. Please reconcile these numbers and specify the operating point used for the abstract claim.","section":"Abstract, Table I, Fig. 7(c)"},{"comment":"The conclusion states that VCR and PCE are comparable to the state of the art, but the 90.1% PCE is a post-layout simulated result and the text states that PCE could not be measured. Please qualify the PCE comparison explicitly as simulated in both the conclusion and the table.","section":"Conclusion and Table I"},{"comment":"The analytical 30% improvement in Fig. 1(b) differs from the measured ~10% improvement in Fig. 7(d); the text explains that additional link losses shift the curves, but it would be clearer to state the loss assumptions in the analytical model and the exact conditions of the measurement.","section":"Fig. 1(b) and Section IV"},{"comment":"The caption of Fig. 3(b) appears to label the sampling and compensation capacitors in reverse relative to the text: the text says C_S samples the error voltage and C_C stores compensation, while the caption reads 'Sampling capacitor CC Compensation capacitor CS.' Please correct the labeling or clarify the notation.","section":"Fig. 3"},{"comment":"Reference [12] lists IEEE J. Solid-State Circuits volume 68 for a 1990 publication; the JSSC volume number appears to be incorrect. Please verify the citation details.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript fits the scope of IEEE Solid-State Circuits Letters. The active-rectifier contribution is largely sound, but the PRE load-monitoring contribution depends on a one-time phase calibration whose validity across coupling and detuning changes is not demonstrated. Since this premise is load-bearing for the system-level claim, I cannot recommend acceptance without additional analysis or measurements. The PCE and settling-time labeling issues are secondary but should be corrected in revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. The active-rectifier delay-compensation scheme is the real contribution: a direct voltage-domain sample-and-accumulate loop replaces the usual slow feedback amplifier, and the fabricated 40-nm chip shows a measured VCR of 93.9%, a clean load-step transient, and a plausible 200-ns settling. That part is worth a serious look. The system-level PRE contribution is a good idea, but the evidence is thin: the TX-side phase sensing is validated at one link condition only (12 mm separation, nominal tuning), and the paper doesn't address whether the ΔΦ calibration survives coupling or detuning changes.\n\nWhat's new: the direct voltage-domain compensation without a slow feedback loop is distinct from the cited adaptive schemes, and integrating the resonance capacitor on-chip at 40.68 MHz is practical. The PRE definition is simple and the 10% measured radiated-power reduction at the demonstrated condition is honest, though the 30% analytical improvement is idealized.\n\nSoft spots. First, the headline 200-ns settling is simulated, but the abstract states it as measured fact. The PCE 90.1% is post-layout simulated, and the paper only says so in Section IV, not consistently in the abstract or Table I. Second, the VCR is inconsistent: 93.9% in the text versus 93.8% in Table I. Minor, but sloppy. Third, and more important: the TX-side load monitoring relies on a single startup calibration of ΔΦ_OPT, and Fig. 7(e) shows the phase-versus-load curve for one link condition. The reflected impedance seen by the TX is a function of ω²M²/Z_RX, so the ΔΦ-to-R_RX mapping is not necessarily invariant to M or detuning. If the implant moves, the calibrated setpoint may no longer correspond to optimal PRE, and the closed-loop regulator would hold a non-optimal R_RX. The paper needs either a sensitivity analysis across coupling and detuning or a clear statement that the system assumes a fixed link.\n\nBottom line: the rectifier contribution is solid and deserves refereeing. The PRE claim is under-validated but not fatal; it is a proof-of-concept at one operating point. I would send this to peer review, with a request to fix the simulated-versus-measured labeling, reconcile the VCR numbers, and either add coupling/detuning data or explicitly scope the PRE claim. A serious editor should not desk-reject it.","headline":"The rectifier delay-compensation work is genuinely new and measured; the TX-side PRE load sensing is an interesting idea that is validated at a single link condition, and the abstract oversells simulated results as measured.","tokens_in":7059,"tokens_out":2976,"would_cite":true,"duration_ms":35416,"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 40.68-MHz active rectifier with sample-and-accumulate delay compensation settles in under 200 ns, while a transmitter-side phase-only monitor reduces radiated power by about 10 percent in measurement.","keywords":["active rectifier","delay compensation","wireless power transfer","biomedical implants","power radiated efficiency","sample-and-accumulate","transmitter-side load monitoring","40.68 MHz"],"falsifier":"Hold the receiver tank load fixed and sweep the coil separation or lateral misalignment while recording the transmitter-side phase difference; if the phase changes by more than a few degrees at the same load, the single-startup-calibration assumption fails and the TX-side monitor would require recalibration or the PRE optimum would drift.","tokens_in":5985,"feed_emoji":"⚡","tokens_out":7332,"duration_ms":82523,"temperature":0.7,"pith_summary":"This paper tries to establish that a wireless-power receiver for implants can settle fast and stay safe at the same time. Its active rectifier replaces the slow feedback amplifier used in prior delay compensation with a direct sample-and-accumulate voltage correction, reaching a worst-case settling of about 200 ns and then correcting within a single 25-ns cycle during load or input steps. On the same 40-nm CMOS chip it reports a measured voltage conversion ratio of 93.9% and a post-layout simulated power conversion efficiency of 90.1%, with the resonance and filter capacitors integrated on chip. The paper further claims that optimizing the link for power radiated efficiency instead of power transfer efficiency reduces radiated power by about 10% in measurement, using only a transmitter-side phase measurement to regulate the link. A sympathetic reader would care because lower radiated power means less tissue heating in implants, and the fast settling keeps efficiency up during ASK downlink communication.","feed_headline":"200-ns rectifier settling trims implant radiated power by 10 percent","feed_subtitle":"Sample-and-accumulate delay compensation plus a TX-phase-only link monitor keep efficiency high without back-telemetry.","key_machinery":"The load-bearing circuit is the sample-and-accumulate voltage-domain delay compensation: a sampling capacitor captures the error voltage $\\Delta V_S$ at the switching instant, identical current sources plus a zero-crossing detector transfer that charge onto a compensation capacitor $C_C$, and the accumulated voltage $V_C = \\sum V_S$ directly offsets the comparator, eliminating the intentionally slow feedback loop. A delay line shifts the OFF-trigger point away from the AC-input trough where the derivative of the input voltage is near zero. The link-level mechanism is the power-radiated-efficiency (PRE) optimization rule: during startup the implant draws constant power, the transmitter sweeps $V_{\\mathrm{TRANS}}$, the minimum received amplitude marks peak PRE, and the corresponding phase $\\Delta\\Phi_{\\mathrm{OPT}}$ between $V_{\\mathrm{TX}}$ and $I_{\\mathrm{TX}}$ becomes the setpoint that the transmitter power amplifier regulates against during normal operation.","core_discovery":"The paper's central claim is that the two main obstacles to practical 40.68-MHz implant power delivery, slow adaptive delay compensation and receiver-side power optimization, can be removed with one rectifier design and one link-control rule. In the rectifier, the on/off comparator delay is compensated by sampling the error voltage at each switching instant and accumulating it directly onto a compensation capacitor, so the compensation voltage is added in series to the comparator input as the new offset. Because no low-bandwidth feedback loop is needed, the compensation settles in a worst case of about 200 ns and continuously re-corrects within a single 25-ns cycle over load or input variations, leaving a residual delay below 200 ps. At the link level, the paper claims that holding the transmitter-side phase angle between $V_{\\mathrm{TX}}$ and $I_{\\mathrm{TX}}$ at a value recorded during a one-time startup calibration maintains the receiver tank load at the power-radiated-efficiency optimum, and that this reduces radiated power by about 10% in measurement compared with optimizing for power transfer efficiency.","pith_inferences":["Beyond the paper, the one-time startup calibration could be re-run periodically without hardware changes, turning the single setpoint into a slow background monitor that would absorb coil movement or detuning; the paper does not test this re-calibration scenario.","Beyond the paper, because the compensation relies on matched sampling and compensation capacitors, capacitor mismatch likely sets a floor on residual switching delay; measuring settling across process corners would reveal whether the 200-ns worst case is set by matching or by the charge-transfer logic.","Beyond the paper, the measured 10% PRE benefit is link-loss dependent, so an extension would sweep coil separation or misalignment and record how the PRE-versus-PTE gap changes to map where the single-calibration assumption starts to fail."],"forward_implications":["Implant rectifiers can track fast load steps and ASK downlink data without the microsecond-scale settling reported for earlier adaptive delay-compensated designs, so efficiency stays high during communication.","With the switched-capacitor power converter in the architecture, any output load can be transformed to the PRE-optimal receiver tank load, extending the efficiency optimum across the whole load range.","The transmitter-side phase measurement closes the control loop without implant-to-transmitter back-telemetry, simplifying the implanted electronics and data link.","Integrating the resonance and filter capacitors on chip at 40.68 MHz reduces the implant to a coil plus the chip, a path to smaller implants at a power density of 118 mW/mm².","Radiated power, the quantity that sets tissue-heating limits, is lower under PRE optimization than under PTE optimization; the paper measures about 10% improvement including real link losses."],"supporting_citations":[{"why":"provides the prior 40.68-MHz cycle-based on/off-delay-compensated rectifier whose settling time this work is compared against","marker":"[8]"},{"why":"establishes the adaptive on/off delay compensation with a low-bandwidth feedback loop that limits response speed","marker":"[2]"},{"why":"supplies the wireless phase-shift control concept that the TX-side load monitoring is similar to","marker":"[7]"},{"why":"provides the auto-zeroed inverter comparator used in the zero-crossing detectors and rectifier comparators","marker":"[12]"},{"why":"is the active rectifier used as a comparison in the table for VCR and settling time","marker":"[5]"},{"why":"represents the PTE-oriented PMU optimization that the PRE approach is intended to replace","marker":"[9]"}],"fun_headline_variants":["200ns rectifier settling trims implant radiated power 10%","Implant rectifier: 200ns settling, 10% less radiation","200ns settling rectifier cuts implant radiated power","200ns rectifier: 10% less implant radiation","Implant power: 200ns rectifier, 10% lower radiation"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that the transmitter-side phase setpoint recorded once at startup remains the correct operating point as the coil link changes, since the phase-to-receiver-load mapping is treated as independent of coupling strength and detuning; Fig. 7(e) shows only one measured link condition.","fun_headline_variants_meta":{"raw":{"variants":["200ns rectifier settling trims implant radiated power 10%","Implant rectifier: 200ns settling, 10% less radiation","200ns settling rectifier cuts implant radiated power","200ns rectifier: 10% less implant radiation","Implant power: 200ns rectifier, 10% lower radiation"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001133,"raw_usage":{"total_tokens":4704,"prompt_tokens":940,"completion_tokens":3764,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":556,"completion_tokens_details":{"reasoning_tokens":3673}},"tokens_in":556,"tokens_out":3764,"duration_ms":38815,"temperature":1.0,"reasoning_tokens":3673,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T05:27:18.085303+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Hold the receiver tank load fixed and sweep the coil separation or lateral misalignment while recording the transmitter-side phase difference; if the phase changes by more than a few degrees at the same load, the single-startup-calibration assumption fails and the TX-side monitor would require recalibration or the PRE optimum would drift.","supporting_citations":[{"cited_title":"A 40.68-MHz Active Rectifier With Cycle-Based On-/Off-Delay Compensation for High-Current Biomedical Implants,","cited_arxiv_id":null,"evidence_quote":"provides the prior 40.68-MHz cycle-based on/off-delay-compensated rectifier whose settling time this work is compared against"},{"cited_title":"Adaptive On/Off Delay- Compensated Active Rectifiers for Wireless Power Transfer Systems,","cited_arxiv_id":null,"evidence_quote":"establishes the adaptive on/off delay compensation with a low-bandwidth feedback loop that limits response speed"},{"cited_title":"A 6.78-MHz Wireless Power Transfer System With Inherent Wireless Phase Shift Control Without Feedback Data Sensing Coil,","cited_arxiv_id":null,"evidence_quote":"supplies the wireless phase-shift control concept that the TX-side load monitoring is similar to"},{"cited_title":"New CMOS high-speed, high- accuracy auto-zero comparator design based on symmetric cross-coupled concepts,","cited_arxiv_id":null,"evidence_quote":"provides the auto-zeroed inverter comparator used in the zero-crossing detectors and rectifier comparators"},{"cited_title":"A 13.56 MHz CMOS Active Rectifier With Switched-Offset and Compensated Biasing for Biomedical Wireless Power Transfer Systems,","cited_arxiv_id":null,"evidence_quote":"is the active rectifier used as a comparison in the table for VCR and settling time"},{"cited_title":"A 13.56 MHz Wireless Power Transfer System With Reconfigurable Resonant Regulating Rectifier and Wireless Power Control for Implantable Medical Devices,","cited_arxiv_id":null,"evidence_quote":"represents the PTE-oriented PMU optimization that the PRE approach is intended to replace"}],"review_version":1}