{"id":"13460a1e-7817-4209-b415-1cff0860e94c","arxiv_id":"2501.11820","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":2.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A MATLAB simulation study finds that state-feedback control with integral action outperforms P, PI, PID, and plain state-feedback controllers for DC servo motor position regulation.","lead":"This paper simulates five standard feedback controllers (P, PI, PID, state feedback, and state feedback with integral action) for DC servo motor position control in MATLAB. It reports that state feedback with integral action gives the best settling time and zero steady-state error, but the comparison relies on unreported tuning choices.","discovery_kind":"replication","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The SFCIA-best claim depends on undisclosed SFC/SFCIA pole placements; without a stated design criterion or the gain matrices, the comparison cannot be verified as fair.","rationale":"I agree with the reader's weakest assumption. The paper's modeling equations are coherent and the plant transfer function follows from the parameters, so the main risk is not the math but the uncontrolled degree of freedom in the state-feedback designs. With no pole locations or gains reported, Table III is not independently checkable, and the comparison lacks the fixed procedure that Ziegler-Nichols provides for classical controllers. The reader's conditional verdict is therefore appropriate: the central claim is plausible but contingent on disclosure. My stress-test does not move the verdict; it sharpens the condition. I also note the rise/peak time overreach, but the decisive check is to reveal and justify the pole placements.","tokens_in":7054,"tokens_out":3911,"duration_ms":40823,"concrete_test":"Obtain the exact Kc and Ka matrices (or the closed-loop pole vectors) used for SFC and SFCIA, then rerun the step-response comparison with SFC/SFCIA poles chosen by an explicit, stated criterion matched to the ZN-tuned PID's dominant dynamics (e.g., same real-part or same damping ratio for the dominant pair). If SFCIA no longer has the lowest overshoot and shortest settling time under matched design effort, the headline comparison is an artifact of undisclosed pole placement.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section II-F4 states that for SFC, closed-loop poles are 'carefully chosen' and MATLAB's place command is used, but neither the poles nor the resulting Kc are reported; Section II-F5 likewise omits Ka for SFCIA. The central claim that SFCIA beats P/PI/PID therefore rests on an unverifiable tuning step. Ziegler-Nichols gives P, PI, and PID a fixed, reproducible tuning, whereas the state-feedback designs have no stated performance criterion (e.g., dominant time constant, damping ratio, or LQR weights), so SFCIA's zero overshoot, zero steady-state error, and 0.26 s settling time could be manufactured by placing poles aggressively while the classical controllers are left with the overshoot-heavy ZN settings. Absent the matrices, an independent reader cannot rerun Table III or tell whether SFCIA's advantage is a property of the control structure or of the chosen pole locations. A secondary inconsistency: Table III and the Section III text concede SFCIA's rise time (0.15 s) and peak time (0.34 s) are slower than PID's (0.03 s and 0.09 s), so the abstract's claim of 'optimized rise and peak times' is not supported as stated, though this does not by itself overturn the ordering.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a third-order linear model of a DC servo motor from electrical and mechanical dynamics, then designs and simulates five controllers for position regulation: P, PI, and PID tuned via the Ziegler-Nichols method, plus state-feedback control (SFC) and state-feedback control with integral action (SFCIA) designed by pole placement. Based on MATLAB step responses, it reports performance metrics in Table III and concludes that SFCIA achieves zero overshoot, zero steady-state error, and the shortest settling time, while the abstract further claims 'optimized rise and peak times.' The central comparison rests on the simulation results; however, the SFC/SFCIA design details (pole locations, gain matrices, and design criteria) are not reported, and the rise/peak time claim is contradicted by the paper's own Table III.","tokens_in":7318,"tokens_out":4187,"duration_ms":41736,"significance":"If the missing design information were supplied, the paper could serve as a reproducible comparative case study for teaching position control of DC servo motors. The motor model in Equations (5)-(8) is algebraically consistent with the stated parameters, and the general conclusion that adding integral action to state feedback eliminates steady-state error is standard and plausible. The paper's strength is its explicit system parameters and a full transfer function, which allow independent verification given the omitted controller gains. The contribution is modest in novelty—the comparison is simulation-only and prior work already compares integral state feedback with PID—but a fully documented comparison would still be useful as an educational benchmark. The current significance is limited by the failure to report the SFC and SFCIA pole placements, which makes the headline comparison unverifiable.","major_comments":[{"comment":"The closed-loop pole locations chosen for SFC and SFCIA are not reported, and the resulting gain matrices Kc and Ka are not given. The text only says the poles were 'carefully chosen' and that MATLAB's 'place' command was used. Without a stated performance criterion (e.g., desired damping ratio, dominant pole time constant, or LQR weights) and the actual pole locations and gains, an independent reader cannot reproduce Table III or judge whether the comparison is fair relative to the fixed Ziegler-Nichols tuning of P, PI, and PID. This is load-bearing for the central claim that SFCIA outperforms the classical controllers.","section":"II-F4, II-F5, Table III"},{"comment":"The abstract states that SFCIA achieves 'optimized rise and peak times,' but Table III reports a rise time of 0.15 s and a peak time of 0.34 s for SFCIA, while PID achieves 0.03 s and 0.09 s, respectively. The text itself acknowledges in Section III that SFCIA is 'marginally slower' than PID on these metrics. The abstract should be corrected to claim zero overshoot, zero steady-state error, and the shortest settling time, but not optimized rise and peak times.","section":"Abstract, Section III, Table III"},{"comment":"The augmented state-space model for SFCIA is incompletely specified. The integral state w is defined as w = ∫(y−r)dt, so wdot = y − r = θ − r, but the matrix Aa as printed has row4 = [1 0 0 0], which gives wdot = θ, not θ − r. The reference-input matrix Br that would supply the −r term is never given. The author should provide Br explicitly or define wdot directly in terms of the error in the augmented state equation.","section":"II-F5, Eq. (16)-(17)"}],"minor_comments":[{"comment":"The text refers to 'Sliding Mode Control (SFC)' in the concluding paragraph of Section III and in Section IV; the abbreviation SFC has been used throughout as 'State Feedback Controller.' This typo should be corrected.","section":"Section IV and Section III, concluding paragraph"},{"comment":"There is a typographical error in 'euation' that should read 'equation.'","section":"Section II-C"},{"comment":"The computed Ziegler-Nichols parameters are not reported. The values of Kcr and Pcr, and hence the final Kp, Ti, and Td used for P, PI, and PID, should be listed to make the classical-controller comparison reproducible.","section":"Table II and Section II-F"},{"comment":"The denominator term 'K 2' in Equation (5) should be typeset as K^2 to avoid confusion with a variable K2.","section":"Eq. (5)"}],"recommendation":"major_revision","confidential_remarks":"This looks like a well-meaning student project. The contribution is modest and primarily pedagogical; the main technical result is plausible but cannot be accepted as presented because the state-feedback design specifications are missing. A full revision reporting pole locations, gain matrices, and the reference-input term for SFCIA would make the comparison reproducible. If the author also tones down the abstract's rise/peak time claim, the paper could be publishable as a short application/education note, though the novelty is limited relative to prior work, especially Ref. [16]."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague, here's my read of Akash's DC servo motor controller comparison. The headline: it's a clean, ordinary simulation study that confirms a well-known result—integral state feedback beats ZN-tuned PID on overshoot and settling time—but the paper cannot be fully checked because the SFC/SFCIA pole locations and gains are never reported.\n\nWhat it does well: the DC motor model in Eq. (5)-(6) is algebraically consistent with the parameters in Table I. The state-space matrices check out. The performance table is laid out clearly, and the Ziegler-Nichols tuning for P/PI/PID is reproducible. The paper is readable and the comparison methodology (same plant, same step input, five controllers, five standard metrics) is sound in spirit.\n\nThe soft spots, in order of importance. First, Section II-F4 and F5 say the SFC and SFCIA poles were 'carefully chosen' and MATLAB's place was used, but no poles or gain matrices are given. That means an independent reader cannot rerun Table III or verify that the SFCIA advantage is a property of the control structure rather than of chosen pole locations. This is the load-bearing gap. Second, the abstract claims SFCIA achieves 'optimized rise and peak times,' but Table III shows PID has faster rise (0.03 vs 0.15 s) and peak (0.09 vs 0.34 s). The Section III text actually concedes this; the abstract overstates. Third, the conclusion confusingly calls SFC 'Sliding Mode Control' when the paper is about State Feedback Control. That's a terminology slip that should be fixed. Fourth, as the paper's own references [16] and [17] show, the central comparison is not new; this is a parameter-specific replication. Nothing wrong with that if positioned as a tutorial or verification, but the paper should say so.\n\nNet: the central claim—SFCIA gives zero overshoot, zero steady-state error, and fastest settling for this plant—is plausible and likely true. But the missing design details mean the result is not yet verifiable. I'd send it to review with a request for the full controller gains and a rewritten abstract. There is enough there for a useful referee exchange, and the paper is honest in its limitations except for those two overstatements.","headline":"A clean but unreproducible textbook comparison: SFCIA beats ZN-tuned PID, yet the missing pole placements and an overstated abstract keep it from being accepted as-is.","tokens_in":7819,"tokens_out":2377,"would_cite":false,"duration_ms":24476,"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":"The paper claims that a state-feedback controller with integral action (SFCIA) outperforms P, PI, PID, and plain state feedback in simulated DC servo position regulation, achieving zero overshoot, zero steady-state error, and the shortest…","keywords":["DC servo motor","position control","state feedback controller","integral action","Ziegler-Nichols tuning","pole placement","overshoot","settling time"],"falsifier":"Re-run the five-controller comparison with a common design constraint, such as forcing SFCIA's closed-loop poles to give a rise time equal to PID's 0.03 s; if SFCIA then shows nonzero overshoot or a settling time no faster than PID's 0.32 s, the paper's ranking is an artifact of pole placement rather than a property of integral action.","tokens_in":6860,"feed_emoji":"⚙️","tokens_out":8611,"duration_ms":73871,"temperature":0.7,"pith_summary":"This paper asks which of five standard controllers—P, PI, PID, state feedback (SFC), and state feedback with integral action (SFCIA)—best regulates the angular position of a DC servo motor in simulation. It builds a third-order state-space model from the motor's electrical and mechanical dynamics, tunes the first three with the Ziegler-Nichols rule, places poles for the state-feedback designs, and compares step responses. The central claim is that SFCIA wins outright: zero overshoot, zero steady-state error, and the fastest settling time, with only a slightly slower rise and peak time than PID. A sympathetic reader would care because the paper offers a concrete, reproducible framework for choosing a controller when precision is the priority, and it isolates the value of integral action in a state-feedback design.","feed_headline":"Integral state feedback wins DC servo position control test","feed_subtitle":"Adding integral action to state feedback yields zero overshoot, zero error, and the fastest settling time in simulations.","key_machinery":"The central object is the augmented state-space system built from the three physical states $\\theta$, $\\dot{\\theta}$, and $i$ plus a fourth state $w = \\int (y - r)\\,dt$ representing the integral of the tracking error. The control input $u = -K_a x_a$ couples state feedback with integral action, and the combined gain $K_a$ is computed by a pole-placement routine. This mechanism removes steady-state error without introducing the overshoot typical of high-gain classical integral action. For P, PI, and PID, the mechanism is the Ziegler-Nichols rule producing $K_p$, $T_i$, and $T_d$ from the critical gain $K_{cr}$ and oscillation period $P_{cr}$.","core_discovery":"The author's discovery is that augmenting the state-space model with an integral-of-error state—so the control law becomes $u = -K_c x - K_i w$, with $w = \\int (y - r)\\,dt$—lets a state-feedback controller eliminate steady-state error while keeping the closed-loop response essentially free of overshoot. In the reported simulations, SFCIA reaches the reference with 0% overshoot, zero steady-state error, and a settling time of 0.26 s, compared with 56.13% overshoot and 0.32 s for PID, 46.31% and 0.60 s for P, 85.41% and 1.11 s for PI, and 2.39% overshoot with 0.90 steady-state error for SFC. The slightly longer rise (0.15 s) and peak (0.34 s) times than PID are the trade-off the paper accepts in exchange for precision.","pith_inferences":["The paper's ranking is contingent on the unstated pole-placement choices for SFC and SFCIA; a fairer comparison would fix a common design target, such as equal rise time or equal phase margin, across all five controllers.","If the physical system's parameters vary or sensor noise is present, the zero-overshoot and zero-steady-state-error claims likely degrade; the paper itself lists experimental validation as future work.","A direct extension would apply the same five controllers to a two-mass or flexible-load servo system, where the extra integral state may interact with resonance modes differently than in the rigid third-order model.","The transfer function used, $\\theta/V = K/[s((Ls+R)(Js+b)+K^2)]$, assumes $K_t = K_e = K$; if that equality is dropped, the SFCIA pole-placement calculation changes and the quantitative rankings could shift."],"forward_implications":["For high-precision position regulation of a DC servo motor in this simulated setting, SFCIA is the recommended choice: it is the only controller achieving both zero overshoot and zero steady-state error.","PID remains the best compromise when speed of response matters most, since it has the fastest rise (0.03 s) and peak (0.09 s) times while still reaching zero steady-state error.","Plain state feedback without integral action cannot reach the reference (steady-state error 0.90), so any precision application needs either integral action or a different compensation scheme.","P and PI controllers, as tuned here, are poorly suited to precision tasks because their overshoot (46.31% and 85.41%) and long settling times (0.60 s and 1.11 s) violate stability-critical requirements.","Because all states ($\\theta$, $\\dot{\\theta}$, $i$, and $w$) are assumed directly measurable, the SFCIA design needs no observer, simplifying implementation."],"supporting_citations":[{"why":"Supplies the Ziegler-Nichols tuning rules that set $K_p$, $T_i$, and $T_d$ for the P, PI, and PID controllers compared in the study.","marker":"[17]"},{"why":"Provides the integral-state-feedback controller design and its comparison with PID, which the paper's SFCIA controller directly builds on.","marker":"[16]"},{"why":"Supplies the pole-placement state-feedback method for DC motor position control that the SFC design uses.","marker":"[11]"}],"fun_headline_variants":["Integral state feedback achieves zero-error servo control","Zero overshoot, zero error: SFCIA tops DC servo test","SFCIA outpaces PID and PI in DC motor position control","Best DC servo control: state feedback with integral action"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the closed-loop pole locations chosen for SFC and SFCIA, which the paper does not report, were selected with tuning effort comparable to the Ziegler-Nichols settings used for P, PI, and PID.","fun_headline_variants_meta":{"raw":{"variants":["Integral state feedback achieves zero-error servo control","Zero overshoot, zero error: SFCIA tops DC servo test","SFCIA outpaces PID and PI in DC motor position control","Best DC servo control: state feedback with integral action"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000258,"raw_usage":{"total_tokens":1578,"prompt_tokens":935,"completion_tokens":643,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":551,"completion_tokens_details":{"reasoning_tokens":574}},"tokens_in":551,"tokens_out":643,"duration_ms":7129,"temperature":1.0,"reasoning_tokens":574,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T17:49:04.127218+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-run the five-controller comparison with a common design constraint, such as forcing SFCIA's closed-loop poles to give a rise time equal to PID's 0.03 s; if SFCIA then shows nonzero overshoot or a settling time no faster than PID's 0.32 s, the paper's ranking is an artifact of pole placement rather than a property of integral action.","supporting_citations":[{"cited_title":"Debnath, P","cited_arxiv_id":null,"evidence_quote":"Supplies the Ziegler-Nichols tuning rules that set $K_p$, $T_i$, and $T_d$ for the P, PI, and PID controllers compared in the study."},{"cited_title":"Ma’ari and N","cited_arxiv_id":null,"evidence_quote":"Provides the integral-state-feedback controller design and its comparison with PID, which the paper's SFCIA controller directly builds on."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the pole-placement state-feedback method for DC motor position control that the SFC design uses."}],"review_version":1}