pith. sign in

arxiv: 2605.25471 · v1 · pith:6TUISAC4new · submitted 2026-05-25 · ⚛️ physics.atom-ph · physics.app-ph· physics.ins-det

FPGA-based disturbance-observer servo for broadband noise suppression in laser frequency stabilization

Meung Ho Seo , Jae Hoon Lee , Young-Ho Park , Hyun-Gue Hong , Seji Kang , Myoung-Sun Heo , Sang-Bum Lee , Taeg Yong Kwon
show 2 more authors
Sangwon Seo Sang Eon Park
This is my paper

Pith reviewed 2026-06-29 19:39 UTC · model grok-4.3

classification ⚛️ physics.atom-ph physics.app-phphysics.ins-det
keywords disturbance observerFPGAlaser frequency stabilizationPID controllernoise suppressionmodulation transfer spectroscopy87RbAllan deviation
0
0 comments X

The pith

A disturbance observer on FPGA augments PID to cut laser frequency noise by 16.9 dB below 40 kHz.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper shows that adding a digital disturbance observer implemented on an FPGA to a standard PID controller can suppress broadband frequency noise in laser stabilization systems. This is achieved using a simple first-order exponential moving average filter as the Q-filter, which avoids the need for detailed frequency-domain plant identification. Measurements on two stabilized lasers using modulation transfer spectroscopy on the 87Rb D2 line demonstrate a 16.9 dB reduction in integrated rms frequency noise below 40 kHz, from about 140 kHz to 20 kHz. Short-term fractional frequency instability improves from 7.9 × 10^{-12} to 4.6 × 10^{-12} at 1 ms, while long-term stability stays the same. This approach provides a straightforward way to improve noise rejection in atomic physics experiments.

Core claim

The central claim is that a digital disturbance observer (DOB) with a first-order exponential moving average Q-filter, implemented on FPGA, augments conventional PID control to achieve broadband frequency-noise suppression in laser stabilization. Using modulation transfer spectroscopy on the 87Rb D2 line, the integrated rms frequency noise below 40 kHz is reduced by 16.9 dB compared to PID alone, from approximately 140 kHz to 20 kHz, with short-term fractional frequency instability improving from 7.9e-12 to 4.6e-12 at 1 ms and long-term stability unchanged.

What carries the argument

The disturbance observer (DOB) using a first-order exponential moving average filter as Q-filter, which approximates the plant dynamics through a single one-dimensional gain sweep instead of multi-parameter identification.

If this is right

  • The DOB-augmented architecture provides enhanced noise rejection in FPGA-based servo systems.
  • Short-term stability improves without impacting long-term performance.
  • The method simplifies implementation by replacing complex plant identification with a gain sweep.
  • It is suitable for atomic physics experiments requiring better frequency stabilization.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • This method could extend to other control applications where quick approximation of plant dynamics is needed without full modeling.
  • Potential for use in integrated photonics or portable atomic clocks to enhance performance.
  • Further tests with varying loop parameters might reveal the limits of the EMA filter approximation.

Load-bearing premise

The first-order exponential moving average filter sufficiently approximates the plant dynamics to enable broadband suppression without introducing instabilities.

What would settle it

A measurement showing that adding the DOB does not reduce the integrated rms frequency noise by approximately 16.9 dB below 40 kHz, or that it causes loop instability, would falsify the claim.

Figures

Figures reproduced from arXiv: 2605.25471 by Hyun-Gue Hong, Jae Hoon Lee, Meung Ho Seo, Myoung-Sun Heo, Sang-Bum Lee, Sang Eon Park, Sangwon Seo, Seji Kang, Taeg Yong Kwon, Young-Ho Park.

Figure 1
Figure 1. Figure 1: FIG. 1. Block diagram of the PID+DOB control architecture. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Optical layout of the modulation transfer spec [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Electronic signal flow diagram of the FPGA-based PID+DOB servo system. (a) External signal path connecting the [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. Dependence of the frequency-noise PSD on DOB [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6. Allan deviation of the beat-note frequency. Short [PITH_FULL_IMAGE:figures/full_fig_p006_6.png] view at source ↗
read the original abstract

We have demonstrated broadband frequency-noise suppression in a laser stabilization system by augmenting a conventional proportional-integral-derivative (PID) controller with a digital disturbance observer (DOB) implemented on a field-programmable gate array (FPGA). The DOB employs a first-order exponential moving average filter as its Q-filter, replacing multi-parameter frequency-domain plant identification with a single one-dimensional gain sweep. Using modulation transfer spectroscopy on the 87Rb D2 line at 780.24 nm, we have measured the frequency-noise power spectral density and the Allan deviation of the beat note between two independently stabilized lasers. The integrated rms frequency noise below 40 kHz decreased by 16.9 dB compared with PID alone, corresponding to a reduction from approximately 140 kHz to 20 kHz. The short-term fractional frequency instability improved from sigma_y(1 ms) = $7.9 \times 10^{-12}$ to $4.6 \times 10^{-12}$, while the long-term stability at tau > 1 s remained within statistical uncertainty. This DOB-augmented architecture offers a simple and effective route to enhanced noise rejection in FPGA-based servo systems for atomic physics experiments.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 1 minor

Summary. The paper demonstrates broadband frequency-noise suppression in a laser stabilization system by augmenting a conventional PID controller with a digital disturbance observer (DOB) implemented on an FPGA. The DOB uses a first-order exponential moving average (EMA) filter as its Q-filter, tuned via a single gain parameter rather than multi-parameter plant identification. Using modulation transfer spectroscopy on the 87Rb D2 line, beat-note measurements between two stabilized lasers show a 16.9 dB reduction in integrated rms frequency noise below 40 kHz (from ~140 kHz to 20 kHz) and short-term fractional frequency instability improvement from 7.9e-12 to 4.6e-12 at 1 ms, with long-term stability unchanged within uncertainty.

Significance. If the central experimental claims hold, the result offers a low-complexity route to enhanced broadband noise rejection in FPGA-based laser servos for atomic physics, avoiding detailed frequency-domain plant modeling. The direct comparison of two independently stabilized lasers via beat-note PSD and Allan deviation provides a falsifiable experimental test of the architecture.

major comments (1)
  1. [Abstract] Abstract: the claim that the first-order EMA Q-filter 'replaces multi-parameter frequency-domain plant identification' with a 'single one-dimensional gain sweep' is load-bearing for attributing the 16.9 dB rms-noise reduction to the DOB; however, no comparison is shown between the EMA transfer function and the measured plant (laser + AOM + MTS) response over the 40 kHz integration band, leaving open the possibility that unmodeled dynamics cause the observed suppression or stability changes.
minor comments (1)
  1. The abstract reports quantitative improvements (rms noise, Allan deviation) without stating the number of independent measurements, statistical methods, error bars, or data-exclusion criteria used to obtain the values 140 kHz to 20 kHz and 7.9e-12 to 4.6e-12.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their careful reading and constructive feedback on our manuscript. We address the single major comment below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the claim that the first-order EMA Q-filter 'replaces multi-parameter frequency-domain plant identification' with a 'single one-dimensional gain sweep' is load-bearing for attributing the 16.9 dB rms-noise reduction to the DOB; however, no comparison is shown between the EMA transfer function and the measured plant (laser + AOM + MTS) response over the 40 kHz integration band, leaving open the possibility that unmodeled dynamics cause the observed suppression or stability changes.

    Authors: The manuscript reports a direct head-to-head comparison of the identical laser stabilization system operated first with PID control alone and then with the PID augmented by the DOB; therefore the measured 16.9 dB reduction in integrated rms frequency noise and the improvement in short-term Allan deviation are attributable to the addition of the DOB. The EMA Q-filter is deliberately chosen so that its single tunable gain sets the effective bandwidth without any multi-parameter frequency-domain plant identification step, which is the central methodological claim. Because the underlying plant (laser + AOM + MTS) is unchanged between the two data sets, any unmodeled plant dynamics are common to both measurements and cannot account for the observed difference. While an overlay of the EMA transfer function against an independently measured plant response would be informative, it is not required to support either the performance result or the claim that detailed plant modeling has been avoided. revision: no

Circularity Check

0 steps flagged

No significant circularity; results are direct experimental measurements

full rationale

The paper presents an experimental demonstration comparing PID-only versus PID+DOB servo performance via measured frequency-noise PSD and Allan deviation on a beat note between two stabilized lasers. No derivation chain exists that reduces a claimed prediction or uniqueness result to fitted inputs or self-citations by construction. The Q-filter choice is presented as a practical simplification whose adequacy is tested by the reported suppression numbers rather than assumed; the 16.9 dB reduction and instability improvement are outcomes of hardware measurements, not tautological outputs of any model equation. This is the normal case of a self-contained experimental report.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The abstract invokes standard control-theory assumptions about the disturbance observer structure and the validity of the EMA approximation for the plant; no new entities are introduced and the only tunable element mentioned is the one-dimensional gain.

free parameters (1)
  • one-dimensional gain for Q-filter
    The abstract states that the method replaces multi-parameter identification with a single gain sweep, implying this gain is chosen or optimized for the servo.
axioms (1)
  • domain assumption The plant dynamics of the laser stabilization loop can be adequately captured by a first-order EMA filter inside the DOB without full frequency-domain modeling.
    This premise is required for the claimed simplification to hold and is stated in the abstract description of the Q-filter choice.

pith-pipeline@v0.9.1-grok · 5786 in / 1381 out tokens · 48210 ms · 2026-06-29T19:39:52.418629+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

37 extracted references

  1. [1]

    PID only (gray) is shown for reference

    at fixed gainK n = 400. PID only (gray) is shown for reference. Dotted vertical lines markf c for eachn. The green shaded regions indicate the integration window (up to 40 kHz). imal. Abovef c, traces with and without DOB converge, confirming the low-pass rejection predicted by Eq. (6). A relocated servo bump appears near 200 kHz, above the 40 kHz integra...

    2022

  2. [2]

    A. D. Ludlow, M. M. Boyd, J. Ye, E. Peik, and P. O. Schmidt, Optical atomic clocks, Rev. Mod. Phys.87, 637 (2015)

    2015

  3. [3]

    Wynands and S

    R. Wynands and S. Weyers, Atomic fountain clocks, Metrologia42, S64 (2005)

    2005

  4. [4]

    S. Bize, P. Laurent, M. Abgrall, H. Marion, I. Maksi- movic, L. Cacciapuoti, J. Gr¨ unert, C. Vian, F. Pereira dos Santos, P. Rosenbusch, P. Lemonde, G. Santarelli, P. Wolf, A. Clairon, A. Luiten, M. Tobar, and C. Sa- lomon, Cold atom clocks and applications, J. Phys. B 38, S449 (2005)

    2005

  5. [5]

    Kasevich and S

    M. Kasevich and S. Chu, Atomic interferometry using stimulated Raman transitions, Phys. Rev. Lett.67, 181 (1991)

    1991

  6. [6]

    Bongs, M

    K. Bongs, M. Holynski, J. Vovrosh, P. Bouyer, G. Con- don, E. Rasel, C. Schubert, W. P. Schleich, and A. Roura, Taking atom interferometric quantum sensors from the laboratory to real-world applications, Nat. Rev. Phys.1, 731 (2019)

    2019

  7. [7]

    C. D. Bruzewicz, J. Chiaverini, R. McConnell, and J. M. 7 Sage, Trapped-ion quantum computing: Progress and challenges, Appl. Phys. Rev.6, 021314 (2019)

    2019

  8. [8]

    Grotti, S

    J. Grotti, S. Koller, S. Vogt, S. Hafner, U. Sterr, C. Lis- dat, H. Denker, C. Voigt, L. Timmen, A. Rolland, F. N. Baynes, H. S. Margolis, M. Zampaolo, P. Thoumany, M. Pizzocaro, B. Rauf, F. Bregolin, A. Tampellini, P. Barbieri, M. Zucco, G. A. Costanzo, C. Clivati, F. Levi, and D. Calonico, Geodesy and metrology with a transportable optical clock, Nat. P...

    2018

  9. [9]

    M´ enoret, P

    V. M´ enoret, P. Vermeulen, N. Le Moigne, S. Bonvalot, P. Bouyer, A. Landragin, and B. Desruelle, Gravity mea- surements below 10 −9 gwith a transportable absolute quantum gravimeter, Sci. Rep.8, 12300 (2018)

    2018

  10. [10]

    Becker, M

    D. Becker, M. D. Lachmann, S. T. Seidel, H. Ahlers, A. N. Dinkelaker, J. Grosse, O. Hellmig, H. M¨ untinga, V. Schkolnik, T. Wendrich, A. Wenzlawski, B. Weps, R. Corgier, T. Franz, N. Gaaloul, W. Herr, D. L¨ udtke, M. Popp, S. Amri, H. Duncker, M. Erbe, A. Kohfeldt, A. Kubelka-Lange, C. Braxmaier, E. Charron, W. Ert- mer, M. Krutzik, C. L¨ ammerzahl, A. P...

    2018

  11. [11]

    Strangfeld, S

    A. Strangfeld, S. Kanthak, M. Schiemangk, B. Wiegand, A. Wicht, A. Ling, and M. Krutzik, A prototype of a compact rubidium-based optical frequency reference for operation on nanosatellites, J. Opt. Soc. Am. B38, 1885 (2021)

    2021

  12. [12]

    Schmidt, M

    M. Schmidt, M. Prevedelli, A. Giorgini, G. M. Tino, and A. Peters, A portable laser system for high-precision atom interferometry experiments, Appl. Phys. B102, 11 (2011)

    2011

  13. [13]

    Bechhoefer, Feedback for physicists: A tutorial essay on control, Rev

    J. Bechhoefer, Feedback for physicists: A tutorial essay on control, Rev. Mod. Phys.77, 783 (2005)

    2005

  14. [14]

    Preuschoff, M

    T. Preuschoff, M. Schlosser, and G. Birkl, Digital laser frequency and intensity stabilization based on the STEM- lab platform (originally Red Pitaya), Rev. Sci. Instrum. 91, 083001 (2020)

    2020

  15. [15]

    Wiegand, B

    B. Wiegand, B. Leykauf, R. J¨ ordens, and M. Krutzik, Linien: A versatile, user-friendly, open-source FPGA- based tool for frequency stabilization and spectroscopy parameter optimization, Rev. Sci. Instrum.93, 063001 (2022)

    2022

  16. [16]

    Neuhaus, M

    L. Neuhaus, M. Croquette, R. Metzdorff, S. Chua, P.- E. Jacquet, A. Journeaux, A. Heidmann, T. Briant, T. Jacqmin, P.-F. Cohadon, and S. Del´ eglise, Python Red Pitaya Lockbox (PyRPL): An open source software package for digital feedback control in quantum optics experiments, Rev. Sci. Instrum.95, 033003 (2024)

    2024

  17. [17]

    Avalos, X

    V. Avalos, X. Nie, A. Yang, C. He, S. Kumar, and K. Dieckmann, Field-programmable-gate-array- based digital frequency stabilization of low-phase-noise diode lasers, Rev. Sci. Instrum.94, 063001 (2023)

    2023

  18. [18]

    D. R. Leibrandt and J. Heidecker, An open source digital servo for atomic, molecular, and optical physics experi- ments, Rev. Sci. Instrum.86, 123115 (2015)

    2015

  19. [19]

    Appel, A

    J. Appel, A. MacRae, and A. I. Lvovsky, A versatile digi- tal GHz phase lock for external cavity diode lasers, Meas. Sci. Technol.20, 055302 (2009)

    2009

  20. [20]

    Di Domenico, S

    G. Di Domenico, S. Schilt, and P. Thomann, Simple ap- proach to the relation between laser frequency noise and laser line shape, Appl. Opt.49, 4801 (2010)

    2010

  21. [21]

    Numata, A

    K. Numata, A. Kemery, and J. Camp, Thermal-noise limit in the frequency stabilization of lasers with rigid cavities, Phys. Rev. Lett.93, 250602 (2004)

    2004

  22. [22]

    Ohnishi, M

    K. Ohnishi, M. Shibata, and T. Murakami, Motion control for advanced mechatronics, IEEE/ASME Trans. Mechatron.1, 56 (1996)

    1996

  23. [23]

    Umeno and Y

    T. Umeno and Y. Hori, Robust speed control of DC ser- vomotors using modern two degrees-of-freedom controller design, IEEE Trans. Ind. Electron.38, 363 (1991)

    1991

  24. [24]

    C. J. Kempf and S. Kobayashi, Disturbance observer and feedforward design for a high-speed direct-drive position- ing table, IEEE Trans. Control Syst. Technol.7, 513 (1999)

    1999

  25. [25]

    Schrijver and J

    E. Schrijver and J. van Dijk, Disturbance observers for rigid mechanical systems: equivalence, stability, and de- sign, ASME J. Dyn. Syst. Meas. Control124, 539 (2002)

    2002

  26. [26]

    Sariyildiz and K

    E. Sariyildiz and K. Ohnishi, A guide to design distur- bance observer, ASME J. Dyn. Syst. Meas. Control136, 021011 (2014)

    2014

  27. [27]

    W.-H. Chen, J. Yang, L. Guo, and S. Li, Disturbance- observer-based control and related methods—an overview, IEEE Trans. Ind. Electron.63, 1083 (2016)

    2016

  28. [28]

    Sariyildiz, R

    E. Sariyildiz, R. Oboe, and K. Ohnishi, Disturbance observer-based robust control and its applications: 35th anniversary overview, IEEE Trans. Ind. Electron.67, 2042 (2020)

    2042

  29. [29]

    Shim and N

    H. Shim and N. H. Jo, An almost necessary and sufficient condition for robust stability of closed-loop systems with disturbance observer, Automatica45, 296 (2009)

    2009

  30. [30]

    Chao, Z.-X

    Y.-X. Chao, Z.-X. Hua, X.-H. Liang, Z.-P. Yue, L. You, and M. K. Tey, Pound–drever–hall feedforward: laser phase noise suppression beyond feedback, Optica11, 945 (2024)

    2024

  31. [31]

    Li, W.-W

    L. Li, W.-W. Huang, X. Wang, Y.-L. Chen, and L. Zhu, Periodic-disturbance observer using spectrum-selection filtering scheme for cross-coupling suppression in atomic force microscopy, IEEE Trans. Autom. Sci. Eng.20, 2037 (2023)

    2037

  32. [32]

    J. H. Shirley, Modulation transfer processes in optical heterodyne saturation spectroscopy, Opt. Lett.7, 537 (1982)

    1982

  33. [33]

    D. J. McCarron, S. A. King, and S. L. Cornish, Modu- lation transfer spectroscopy in atomic rubidium, Meas. Sci. Technol.19, 105601 (2008)

    2008

  34. [34]

    H.-R. Noh, S. E. Park, L. Z. Li, J.-D. Park, and C.-H. Cho, Modulation transfer spectroscopy for 87Rb atoms: theory and experiment, Opt. Express19, 23444 (2011)

    2011

  35. [35]

    D. W. Allan, Statistics of atomic frequency standards, Proc. IEEE54, 221 (1966)

    1966

  36. [36]

    D. W. Allan, Time and frequency (time-domain) charac- terization, estimation, and prediction of precision clocks and oscillators, IEEE Trans. Ultrason. Ferroelectr. Freq. Control34, 647 (1987)

    1987

  37. [37]

    S. Lee, G. Moon, S. E. Park, H.-G. Hong, J. H. Lee, S. Seo, T. Y. Kwon, and S.-B. Lee, Laser frequency sta- bilization in the 10 −14 range via optimized modulation transfer spectroscopy on the 87Rb D2 line, Opt. Lett.48, 1020 (2023)

    2023