High-power Test and System Integration of Direct RF Sampling Based LLRF Control and Monitoring System for S-Band Accelerating Structures

Ankur Dhar; Chao Liu; Emilio Nanni; Emma Snively; Joseph Olszewski; Michael Hoganson; Sophia Morton; Thi Le; Wei-Hou Tan

REVIEW 2 major objections 5 minor 1 cited by

High-power Test and System Integration of Direct RF Sampling Based LLRF Control and Monitoring System for S-Band Accelerating Structures

T0 review · 2 major / 5 minor · reviewed 2026-07-13 · grok-4.5

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

Pith's one-line read Direct RF sampling LLRF with a custom solid-state amplifier meets the open-loop pulse stability needed for S-band accelerator drive systems.

desk verdict Solid open-loop S-band integration data for RFSoC LLRF + custom SSA; useful progress report, not a closed-loop demonstration. read the letter →

arxiv 2604.01698 v2 pith:QITQYNHZ submitted 2026-04-02 physics.acc-ph astro-ph.IM

Chao Liu , Ankur Dhar , Michael Hoganson , Emilio Nanni , Joseph Olszewski , Emma Snively , Sophia Morton , Thi Le

show 1 more author

Wei-Hou Tan

This is my paper · ORCID

classification physics.acc-phastro-ph.IM

keywords LLRFdirectRFsamplingRFSoCS-bandacceleratingstructuresolid-stateamplifierpulsestabilityparticleacceleratorcontrol

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

This paper shows that a next-generation low-level RF controller built on an RF system-on-chip can sample and generate S-band pulses directly, without the analog mixers of conventional systems, and can be integrated with a custom solid-state amplifier that reaches the kilowatt level required to drive a klystron. Stage-by-stage characterization of power levels and pulse stability yields open-loop amplitude fluctuations as low as 0.067 percent and phase fluctuations near 0.05 degrees for microsecond pulses, levels the authors state are better than the needs of the targeted test facilities and generally adequate for linear accelerators. The work matters because the same highly configurable digital front-end, high data throughput, and on-chip compute resources also make the platform suitable for collecting consistent high-resolution RF datasets and for edge deployment of real-time control algorithms. Completing this integration path is presented as the necessary step before the system can serve as the full drive for existing S-band stations and be extended to other frequency bands.

What carries the argument

The NG-LLRF chassis (direct RF sampling and synthesis on an RFSoC evaluation board at multi-GSPS rates) paired with the custom solid-state amplifier that supplies the intermediate drive power; their sequential integration and loop-back measurement of pulse-top and pulse-to-pulse statistics carry the performance argument.

What would settle it

Closed-loop high-power tests that drive the klystron show amplitude fluctuations remaining above 0.1 percent or phase fluctuations above 0.05 degrees after any mismatch mitigation, or the oscillations prove irreducible without degrading the reported numbers.

Watch

Extended reading notes

Core claim

When an RFSoC-based next-generation LLRF chassis is integrated with a custom solid-state amplifier and characterized in open-loop S-band loopback, the resulting RF pulses exhibit amplitude and phase stability that meets or exceeds the requirements of the targeted linear-accelerator applications and falls within the range generally regarded as sufficient for most linacs.

Load-bearing premise

The measured open-loop stability will still hold once the system drives a real klystron under closed-loop conditions, which requires that the observed pulse-top oscillations from amplifier-load mismatch can be fixed or tolerated.

Editorial extensions

If this is right

Once facility triggering is synchronized, the same NG-LLRF plus amplifier chain can replace the existing drive of an S-band klystron.
High-resolution, consistently timestamped RF waveforms become available as training data for large autonomous control models.
The platform’s wider RF bandwidth and on-chip resources support edge deployment of real-time AI/ML feedback algorithms.
Successful S-band integration supplies a concrete path for deploying the same architecture at other frequency bands.

Reading between the lines

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

Eliminating the pulse-top oscillations attributed to load mismatch is likely the remaining practical gate before open-loop numbers translate to operational closed-loop performance.
Direct sampling can shrink the analog component count and calibration overhead of multi-station RF systems relative to traditional heterodyne LLRF.
The high streaming throughput may allow one chassis to monitor many RF stations simultaneously, lowering per-channel cost at large facilities.

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

This is a clean engineering progress note: they take their existing RFSoC NG-LLRF platform (already shown at C-band), retune it for 2856 MHz, build a custom ~1 kW SSA, and document the staged open-loop integration with concrete power and stability numbers.

What is new is the SSA input–output curve (to 61 dBm), the 2 µs / 5 µs pulse-top captures, and the tabulated pulse-to-pulse amplitude/phase fluctuations over 120 shots at two drive levels. The longer, higher-power cases land at ~0.067 % amp and ~0.048° phase—right at or under the 0.1 % / 0.05° rule-of-thumb they cite. The measurement chain is straightforward (direct sampling, digital down-conversion, known attenuations) and the figures/tables support the numbers they report. Self-citation is heavy but expected; it supplies the platform pedigree rather than circular claims.

Soft spots are exactly the ones the authors already flag and the reader correctly isolates. Everything is open-loop SSA loopback; the pulse-top oscillations (visible in Figs. 4–5) are left as “mismatch/reflection, needs investigation.” Shorter-pulse entries sit above the thresholds. The abstract’s “considerably better than requirements” is a touch ahead of the mixed data, but the body is more careful and lists the remaining steps (trigger integration, closed loop, klystron drive). No derivation errors, no invented entities, no circular modeling.

This is for people already working on LLRF upgrades or AI/ML-ready waveform capture at S-band stations. It is not a technology breakthrough and does not yet prove full high-power closed-loop performance, but it is honest, reproducible-looking characterization that a serious referee can evaluate. I would send it to review; the community needs these intermediate integration papers if the RFSoC path is going to mature.

Referee Report

2 major / 5 minor

Summary. The manuscript reports stage-by-stage integration and open-loop characterization of an RFSoC-based next-generation LLRF (NG-LLRF) system with a custom solid-state amplifier (SSA) for S-band accelerating structures, aimed at NLCTA and related SLAC test facilities. Direct RF sampling/synthesis on a ZCU208-based chassis is used without analog mixers. The authors measure SSA power transfer (reaching ~1 kW / 60 dBm at ~4.9 dBm input, peak 61.12 dBm), capture pulse-top amplitude and phase for 2 µs and 5 µs pulses at two drive levels, and quantify pulse-to-pulse fluctuations over 120 consecutive pulses at 120 Hz. Reported open-loop stability reaches ~0.067% amplitude and ~0.048° phase for the longer, higher-power case; the work is positioned as an essential step toward closed-loop klystron drive and AI/ML-ready control.

Significance. Direct-sampling LLRF platforms are of practical interest for future linacs because they remove analog mixers, increase bandwidth and data throughput, and provide on-chip compute resources useful for edge AI/ML. The paper supplies concrete hardware measurements—power curves, time-domain pulse tops, and tabulated standard deviations—rather than purely conceptual claims. If the open-loop numbers hold under closed-loop and full high-power klystron conditions, the platform would be a credible candidate for NLCTA upgrades and similar S-band stations. Strengths include quantitative tables over 120 pulses, clear stage-by-stage integration narrative, and explicit acknowledgment that closed-loop and klystron drive remain future work.

major comments (2)

[Abstract / Tables 1–2] Abstract and Introduction claim the platform “delivered considerably better performance than the requirements of the targeted applications” and is “considerably better than the requirements of LCLS.” Tables 1–2 show that only the 5 µs, higher-power open-loop cases meet or approach the authors’ own cited sufficiency thresholds (~0.1% amp / ~0.05° phase); the 2 µs entries (0.139–0.190% amp, 0.072–0.085° phase) sit above those thresholds. The body correctly flags open-loop status and remaining work, but the abstract phrasing overstates the present data. Align the abstract claim with the mixed tabulated results or restrict it to the longer-pulse cases that actually meet the thresholds.
[SSA and NG-LLRF / Figs. 4–5] Figures 4–5 show clear oscillations on both amplitude and phase pulse tops, attributed to load–SSA mismatch and reflection. The text states these “need to be further investigated and resolved” for stringent flatness applications, yet the stability numbers in Tables 1–2 are extracted from the same pulse tops. Because the central claim is that the integrated NG-LLRF+SSA chain already delivers accelerator-relevant stability, the manuscript should either (a) quantify the oscillation amplitude/period and show that the reported std-devs remain valid after filtering or windowing, or (b) demonstrate that the oscillations are outside the bandwidth of interest for the intended control loops. Leaving them as a qualitative caveat weakens the load-bearing stability claim for the full high-power path.

minor comments (5)

[Throughout] Numerous missing spaces and concatenated words appear throughout (e.g., “HighprecisionLow-levelRF”, “directwithouttheanalogmixers”, “TheNG-LLRFcanbeclocked”). A thorough copy-edit pass is needed for readability.
[Custom Solid State Amplifier / Fig. 3] Figure 3 caption and axis labels would benefit from explicit units on both axes and a note of the pulse width/duty cycle used for the power sweep, so the transfer curve can be reproduced.
[SSA and NG-LLRF] The text states that the SSA monitoring port has “an integrated 30 dB attenuation” and an additional 10 dB external attenuator is used for loop-back; a short table or sentence listing the full attenuation chain and any calibration uncertainty would strengthen the absolute power claims.
[Introduction / Conclusion] References [13–14, 17] are prior work by the same group; a sentence clarifying what is new in the present S-band integration relative to the earlier C-band high-power test would help readers assess novelty.
[Tables 1–2] Table 1 and Table 2 captions repeat “at 120 Hz in 1s”; the body already states 120 consecutive pulses. Minor redundancy can be removed.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: open-loop SSA/NG-LLRF stability numbers are direct waveform statistics, not forced by fit or self-citation.

full rationale

This is an experimental integration and characterization paper. The load-bearing numerical claims (SSA P_in/P_out curve reaching ~1 kW, pulse-top traces in Figs. 4–5, and the amplitude/phase fluctuation percentages and degrees in Tables 1–2) are computed directly from RFSoC-captured baseband waveforms of 120 consecutive pulses; they are not obtained by fitting a model parameter and then re-labeling a related quantity as a prediction, nor by any algebraic identity that reduces output to input by construction. Self-citations ([13], [14], [17], etc.) supply background on the RFSoC platform and prior C-band/loop-back results, but the present SSA integration data stand independently of those citations and are externally falsifiable by repeating the open-loop measurement. No uniqueness theorem, ansatz, or definitional loop appears. Consequently the derivation chain (measurement procedure o tabulated std-devs) contains no circular step.

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

Experimental hardware paper. Claims rest on measured RF power and waveform statistics under explicitly chosen drive settings rather than free parameters fitted to a theoretical model or newly postulated physical entities. Domain assumptions are standard RF-engineering practice (digital down-conversion fidelity, coupler attenuation knowledge, trigger timing). No invented particles, forces, or conserved quantities appear.

free parameters (2)

DAC amplitude set-points (2000 / 4000) = 2000 and 4000 (arbitrary DAC units)
Hand-chosen digital drive levels that map to the two reported input powers (-13.05 dBm / -7.15 dBm); the stability numbers are reported at these specific points and are therefore conditioned on them.
SSA monitoring-port attenuation = 30 dB + 10 dB
Stated as an integrated 30 dB plus an external 10 dB; absolute power and amplitude scales depend on the accuracy of these fixed attenuations, which are taken as known rather than re-measured in the paper.

assumptions (3)

domain assumption Digital down-mixing of the directly sampled 2856 MHz waveform inside the RFSoC produces baseband amplitude and phase that faithfully represent the RF pulse envelope for the purpose of stability statistics.
Invoked throughout the loopback analysis (Figs. 4–7 and Tables 1–2); no independent analog reference measurement is shown to quantify residual digital-mixing error.
domain assumption The commercial SSA module and custom power-supply boards behave as a linear (or at least repeatable) amplifier over the tested input range once the enable trigger is applied >5 µs early.
Underpins the power-curve and pulse-top measurements; saturation and gain compression are noted but not modeled.
domain assumption Standard deviation of the per-pulse average amplitude (normalized) and of the per-pulse average phase over 120 consecutive shots is an adequate metric of ‘pulse-to-pulse fluctuation’ for comparison with the 0.1 % / 0.05° rule-of-thumb.
Used to populate Tables 1–2 and to claim sufficiency relative to linear-accelerator requirements [16].

how reviews work

0 comments

Cite this review

Pith. "Pith review of High-power Test and System Integration of Direct RF Sampling Based LLRF Control and Monitoring System for S-Band Accelerating Structures." pith.science (2026). https://pith.science/paper/QITQYNHZ

@misc{pith2026260401698,
  author       = {Pith},
  title        = {Pith review of: High-power Test and System Integration of Direct RF Sampling Based LLRF Control and Monitoring System for S-Band Accelerating Structures},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QITQYNHZ}},
  note         = {Machine review of arXiv:2604.01698}
}

read the original abstract

High precision Low-level RF (LLRF) control and monitoring systems for future particle accelerators will be a significant technical challenge as the requirements in performance, flexibility and affordability become increasingly stringent. We have developed an RF system-on-chip (RFSoC) based next-generation LLRF (NG-LLRF) for S-band accelerating structures, which samples and synthesizes the RF pulses directly without the analog mixers used for traditional LLRF systems. The platform delivered considerably better performance than the requirements of the targeted applications, such as the upgrades for Next Linear Collider Test Accelerator (NLCTA) and test facilities at SLAC. As part of the upgrade program, we also developed a custom solid-state amplifier (SSA) to deliver RF pulses at the desired power level of the klystron. Integration of the LLRF with the SSA and the high-power test facility could be challenging. The power levels and RF pulse stability at each stage of the high-power RF drive system must be optimized to deliver the desired RF performance. In this paper, the integration procedure and the test and characterization results at each stage of integration will be summarized, analyzed and discussed. This integration is an essential step for the full deployment of the NG-LLRF system to test facilities and accelerators in different frequency bands.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Low Level RF and Timing System Design for the Cool Copper Collider
physics.acc-ph 2026-06 unverdicted novelty 5.0 of 10

Design and prototype testing of an RFSoC-based LLRF system with direct sampling for stabilizing RF fields in the C3 collider cavities.

Reference graph

Works this paper leans on

17 extracted references · 3 linked inside Pith · cited by 1 Pith paper

[1]

Characterizing the performance of high-speed data converters for rfsoc-based radio astronomy receivers

C. Liu, M. E. Jones, and A. C. Taylor, “Characterizing the performance of high-speed data converters for rfsoc-based radio astronomy receivers”,Monthly Notices of the Royal Astronomical Society, vol. 501, no. 4, pp. 5096–5104, 2021

2021
[2]

Development of general-purpose digital backend for single-dish centimetre and millimetre-wave telescopes

C. Liu, M. E. Jones, and A. C. Taylor, “Development of general-purpose digital backend for single-dish centimetre and millimetre-wave telescopes”, in2022 3rd URSI Atlantic and Asia Pacific Radio Science Meeting (AT-AP-RASC), pp. 1–4, 2022

2022
[3]

Advanced rfsoc readout for space- based superconducting sensor arrays

S. W. Hendersonet al., “Advanced rfsoc readout for space- based superconducting sensor arrays”, inMillimeter, Submil- limeter, and Far-Infrared Detectors and Instrumentation for Astronomy XI, vol. 12190, pp. 339–355, 2022

2022
[4]

A born again 32-meter radio telescope for mexico

S. E. Kurtzet al., “A born again 32-meter radio telescope for mexico”, inGround-based and Airborne Telescopes IX, vol. 12182, pp. 929–937, 2022

2022
[5]

Evaluatingdirectrfsam- plingperformanceforrfsoc-basedradio-frequencyastronomy receivers

C.Liu,L.Ruckman,andR.Herbst,“Evaluatingdirectrfsam- plingperformanceforrfsoc-basedradio-frequencyastronomy receivers”,arXiv preprint arXiv:2309.08067, 2023

arXiv 2023
[6]

Developmentofrfsoc-baseddirectsam- pling highly multiplexed microwave squid readout for cmb andsubmillimetersurveys

C. Liu, Z. Ahmed, S. W. Henderson, R. Herbst, L. Ruckman, andT.Satterthwaite,“Developmentofrfsoc-baseddirectsam- pling highly multiplexed microwave squid readout for cmb andsubmillimetersurveys”,inMillimeter,Submillimeter,and Far-Infrared Detectors and Instrumentation for Astronomy XII, vol. 13102, pp. 422–429, 2024

2024
[7]

Llrf system considerations for a compact,commercialc-bandacceleratorusingtheamdxilinx rf-soc

J. Einstein-Curtiset al., “Llrf system considerations for a compact,commercialc-bandacceleratorusingtheamdxilinx rf-soc”,arXiv preprint arXiv:2310.04209, 2023

arXiv 2023
[8]

Direct rf sampling based llrf control system for c-band linear accelerator

C. L. et al., “Direct rf sampling based llrf control system for c-band linear accelerator”, inProc. IPAC’24, Nashville, TN, pp. 25–28, May 2024. doi:10.18429/JACoW-IPAC2024-MOCN2

doi:10.18429/jacow-ipac2024-mocn2 2024

Show all 17 references

[9]

A compact low-level rf control system for ad- vanced concept compact electron linear accelerator

C. Liuet al., “A compact low-level rf control system for ad- vanced concept compact electron linear accelerator”,Review of Scientific Instruments, vol. 96, no. 9, 2025

2025
[10]

Highpowerarbitraryrfpulseshapingtestswith ng-llrf and cool copper collider prototype structure

C.Liuetal.,“Highpowerarbitraryrfpulseshapingtestswith ng-llrf and cool copper collider prototype structure”,arXiv preprint arXiv:2511.08410, 2025

2025
[11]

High precision rf pulse shaping with direct rf sampling for future linear accelerators

C. Liuet al., “High precision rf pulse shaping with direct rf sampling for future linear accelerators”, inProc. IPAC’25, Taipei, Taiwan, Jun. 2025, pp. 3237–3240. doi:10.18429/JACoW-IPAC2025-THPS140

2025 doi
[12]

Next gen- eration llrf control platform for compact c-band linear ac- celerator

C. Liu, R. Herbst, L. Ruckman, and E. Nanni, “Next gen- eration llrf control platform for compact c-band linear ac- celerator”, inEPJ Web of Conferences, vol. 315, p. 02008, 2024

2024
[13]

High-powertestofac-bandlinearaccelerating structurewithanrfsoc-basedllrfsystem

C.Liuetal.,“High-powertestofac-bandlinearaccelerating structurewithanrfsoc-basedllrfsystem”,ReviewofScientific Instruments, vol. 96, no. 4, 2025

2025
[14]

Nextgenerationllrfcontrolandmonitoringsys- tem for s-band linear accelerators

C.Liuetal.,“Nextgenerationllrfcontrolandmonitoringsys- tem for s-band linear accelerators”, inProc. IPAC’25, Taipei, Taiwan, Jun. 2025, pp. 3241–3244. doi:10.18429/JACoW-IPAC2025-THPS141

2025 doi
[15]

Rfsoczcu208evaluationkit

AMD,“Rfsoczcu208evaluationkit”,PublishedonAMDwe- biste https://www.amd.com/en/products/adaptive- socs-and- fpgas/evaluation- boards/zcu208.html, (Accessed: 24 may 2026),

2026
[16]

Esppu input: c3 within the

M. B. Andorfet al., “Esppu input: c3 within the" linear col- lider vision"”,arXiv preprint arXiv:2503.20829, 2025

2025 arXiv
[17]

Next generation direct rf sampling llrf control and monitoring system for linear accelerators

C. Liuet al., “Next generation direct rf sampling llrf control and monitoring system for linear accelerators”, in Proc. 20th Int. Conf. Accel. Large Exp. Phys. Control Syst. (ICALEPCS’25), Chicago, IL, USA, Sep. 2025, pp. 199–204. doi:10.18429/JACoW-ICALEPCS2025-MODR005

2025 doi

Pith tools

Reviewed July 13, 2026 · model on record in the stance chip above.

[1] [1]

Characterizing the performance of high-speed data converters for rfsoc-based radio astronomy receivers

C. Liu, M. E. Jones, and A. C. Taylor, “Characterizing the performance of high-speed data converters for rfsoc-based radio astronomy receivers”,Monthly Notices of the Royal Astronomical Society, vol. 501, no. 4, pp. 5096–5104, 2021

2021

[2] [2]

Development of general-purpose digital backend for single-dish centimetre and millimetre-wave telescopes

C. Liu, M. E. Jones, and A. C. Taylor, “Development of general-purpose digital backend for single-dish centimetre and millimetre-wave telescopes”, in2022 3rd URSI Atlantic and Asia Pacific Radio Science Meeting (AT-AP-RASC), pp. 1–4, 2022

2022

[3] [3]

Advanced rfsoc readout for space- based superconducting sensor arrays

S. W. Hendersonet al., “Advanced rfsoc readout for space- based superconducting sensor arrays”, inMillimeter, Submil- limeter, and Far-Infrared Detectors and Instrumentation for Astronomy XI, vol. 12190, pp. 339–355, 2022

2022

[4] [4]

A born again 32-meter radio telescope for mexico

S. E. Kurtzet al., “A born again 32-meter radio telescope for mexico”, inGround-based and Airborne Telescopes IX, vol. 12182, pp. 929–937, 2022

2022

[5] [5]

Evaluatingdirectrfsam- plingperformanceforrfsoc-basedradio-frequencyastronomy receivers

C.Liu,L.Ruckman,andR.Herbst,“Evaluatingdirectrfsam- plingperformanceforrfsoc-basedradio-frequencyastronomy receivers”,arXiv preprint arXiv:2309.08067, 2023

arXiv 2023

[6] [6]

Developmentofrfsoc-baseddirectsam- pling highly multiplexed microwave squid readout for cmb andsubmillimetersurveys

C. Liu, Z. Ahmed, S. W. Henderson, R. Herbst, L. Ruckman, andT.Satterthwaite,“Developmentofrfsoc-baseddirectsam- pling highly multiplexed microwave squid readout for cmb andsubmillimetersurveys”,inMillimeter,Submillimeter,and Far-Infrared Detectors and Instrumentation for Astronomy XII, vol. 13102, pp. 422–429, 2024

2024

[7] [7]

Llrf system considerations for a compact,commercialc-bandacceleratorusingtheamdxilinx rf-soc

J. Einstein-Curtiset al., “Llrf system considerations for a compact,commercialc-bandacceleratorusingtheamdxilinx rf-soc”,arXiv preprint arXiv:2310.04209, 2023

arXiv 2023

[8] [8]

Direct rf sampling based llrf control system for c-band linear accelerator

C. L. et al., “Direct rf sampling based llrf control system for c-band linear accelerator”, inProc. IPAC’24, Nashville, TN, pp. 25–28, May 2024. doi:10.18429/JACoW-IPAC2024-MOCN2

doi:10.18429/jacow-ipac2024-mocn2 2024