Charge-Unified Semiconductor Switching Theory

Ameer Janabi; Biao Zhao; C.Q. Jiang; Feng An; Frede Blaabjerg; Hui Zhao; Jinwei Qi; Stephan Goetz; Teng Long; Wucheng Ying

REVIEW 3 major objections 4 minor 4 references

This paper claims that transistor turn-on switching can be derived from fundamental conservation laws alone, yielding an Eon prediction model that cuts average error roughly 17-fold versus the conventional waveform-based model.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-04 14:48 UTC pith:TO2DPEYT

load-bearing objection The model is a plausible extension of Kasper's iZVS loss model, but the '17-fold improvement' is not established as an independent prediction because the final equations still contain switching-waveform integrals. the 3 major comments →

arxiv 2510.17815 v2 pith:TO2DPEYT submitted 2025-09-26 eess.SY cs.SY

Charge-Unified Semiconductor Switching Theory

Wucheng Ying , Jinwei Qi , Hui Zhao , Feng An , Ziheng Yu , Yanfeng Shen , C.Q. Jiang , Yanghong Xia

show 5 more authors

Ameer Janabi Biao Zhao Stephan Goetz Frede Blaabjerg Teng Long

This is my paper

classification eess.SY cs.SY

keywords transistor switchingswitching lossturn-on energy predictioncharge conservationenergy conservationequivalent circuit modelMiller platformSiC MOSFET

verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper tries to establish that transistor switching, long treated as an empirical black box, can be derived from fundamental physical laws. It proposes a charge-unified view in which every circuit element is a medium for charge redistribution, with the switch under study represented by a single non-linear resistance RS1 and the complementary switch's dynamics included explicitly. From this view it derives a closed-form turn-on energy (Eon) prediction model twice—once from charge conservation and once from energy conservation—and shows the two derivations yield the same expression. Measured against SiC MOSFETs, the model cuts average prediction error roughly 17-fold relative to the conventional model. If right, switching analysis shifts from waveform-fitting to first-principles reasoning.

Core claim

The central claim is that the turn-on switching event is governed by the gate-voltage-driven collapse of a single non-linear equivalent resistance RS1, and that all observed waveform features—current commutation, voltage fall, the Miller platform, and energy loss—are consequences of charge and energy conservation applied to a half-bridge that includes the complementary switch's non-linear capacitance and reverse-recovery dynamics. The paper derives an Eon prediction model in two independent ways, from charge conservation and from energy conservation, and reports that the two expressions are identical, which it presents as the first intrinsic unification of the two conservation laws in switch

What carries the argument

The load-bearing object is the non-linear equivalent resistance RS1 = vds,S1/iRS1, defined by Ohm's law at each operating point and applied to all conduction modes. Alongside it, the model uses lumped non-linear capacitances for output charge and reverse recovery, and treats the complementary switch's junction-capacitance transition as an essential input rather than an ignorable detail. The derivations then apply KCL and KVL plus charge and energy conservation to the half-bridge, producing Eq (13) and Eq (30), which are claimed to be identical.

Load-bearing premise

The derivation assumes the waveform-level integrals in the final Eon expression—the complementary switch's current, the load-inductor term, and the shoot-through dissipation term—are known independently of the Eon being predicted, but the paper does not explain how a user obtains them a priori; if they come from measured or simulated switching waveforms, the '17-fold improvement' is a post-hoc reconstruction rather than a first-principles prediction.

What would settle it

Take a switching-ON event for which no measured waveforms are available; compute the proposed Eon using only datasheet curves and the circuit schematic, and compare against a double-pulse measurement. If the error returns to the conventional model's ~50% level, the claim of first-principles prediction fails. A quicker check: set the complementary switch's reverse-recovery capacitance Crr,S2 to zero and see whether the model's error degrades; if it does not, the CC term is not the source of the improvement.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Share X Bluesky LinkedIn Reddit HN

If this is right

If the model is correct, turn-on loss can be predicted from physical laws and datasheet-level parameters without relying on observed switching waveforms, reducing the need for repeated double-pulse measurements.
The physical origin of the Miller platform is identified as negative feedback between the complementary switch's dv/dt and gate-drain capacitance charging, with distinct sub-phase mechanisms in hard, zero-voltage, and incomplete-zero-voltage switching.
Charge conservation and energy conservation yield the same closed-form Eon expression, so either law alone is sufficient to derive switching loss—a unification that could extend to other switching metrics and other device types.
The same paradigm accounts for zero-voltage, hard, and incomplete-zero-voltage switching within one framework, replacing scenario-specific empirical models with a single causal description.
The framework opens a path to designing converters for minimum switching loss before hardware is built, if all required inputs can be sourced from datasheet curves and circuit conditions.

Where Pith is reading between the lines

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

The claimed 17-fold improvement is only meaningful if the integrals involving the complementary switch's waveforms in Eq (30) are obtained independently of the measured Eon; if they are read from the same measured waveforms, the comparison partly measures in-sample fit rather than predictive power.
The equivalence of the charge- and energy-conservation derivations may be a mathematical identity given the same approximations (neglecting gate voltages, lumping capacitances) rather than two fully independent physical validations, since both paths share those approximations.
A natural testable extension is to apply the model to IGBTs or GaN devices, where reverse-recovery and output-capacitance characteristics differ; the framework predicts the same structural form should hold.
If the model's inputs can be sourced purely from datasheet curves and a circuit schematic, it would enable predictive loss optimization in converter design before any hardware is built.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Desk Editor's Note

The model is a plausible extension of Kasper's iZVS loss model, but the '17-fold improvement' is not established as an independent prediction because the final equations still contain switching-waveform integrals.

read the letter

Dear X,

The real contribution is concrete: for incomplete ZVS turn-on in a half-bridge, the authors add charge-commutation energy and load-inductor work to the energy-conservation model from Kasper et al., and they show the same Eon expression can be derived from charge conservation and from energy conservation. The experimental sweep is wide—two SiC MOSFETs, two DC voltages, many ΔV points—and the model tracks the measured Eon much better than the conventional model. If the model can be evaluated from datasheet parameters and known operating conditions, that is a genuine improvement to loss estimation.

But I don't see how it can be, and the paper doesn't say. Equations (13) and (30) still contain integrals of i_RS2, v_ds,S2 times i_L, and v_ds,S2 times i_RS2 over the switching interval. Those are exactly the instantaneous waveform quantities a predictive model is supposed to avoid. The paper claims to derive everything 'solely from fundamental physical laws' and not to rely on observed waveforms, yet it never gives a procedure for computing those integrals from VDC, ΔV, gate-drive conditions, load current, and device capacitance curves. Without that, the '17-fold error reduction' is not an independent prediction; it is an energy-balance reconstruction that uses the measured or simulated switching waveforms as inputs. The provenance of the 'calculated values' in Table 1 is never disclosed, and there is no code. That omission is load-bearing, not cosmetic.

Also, the packaging is out of proportion with the content. The arXiv metadata promises a 'Charge-Unified Semiconductor Switching Theory' of sweeping scope, while the manuscript is a focused iZVS loss-modeling study. The 'textbook-level foundation' and 'first-principles paradigm' language is doing a lot of work that the equations don't support.

What's worth keeping: the physical narrative about the complementary switch's nonlinear dynamics and the Miller platform is thoughtful, and the algebraic cross-check between charge and energy conservation is a useful sanity check, even if it doesn't establish predictivity. The paper cites the right prior work (Kasper, Perera, Zhang) and builds on it honestly.

Who is this for? Power electronics researchers who care about SiC switching-loss modeling. They might get a useful correction to the Kasper model—if the authors can close the a priori gap. As it stands, I would send it to peer review rather than desk-reject, because the experimental data and the model idea are worth referee scrutiny, and a good referee can force the authors to state exactly how the integrals are evaluated. But I wouldn't accept it in current form.

Best

Referee Report

3 major / 4 minor

Summary. The manuscript proposes 'Charge-Unified Semiconductor Switching Theory' (CUSST), a framework that models a transistor as a nonlinear equivalent resistance (RS1) plus equivalent capacitances, including a reverse-recovery capacitance Crr,S2 for the complementary switch. It offers qualitative explanations of ZVS, hard-switching, and two iZVS turn-on scenarios, and then derives an Eon prediction model for case-2 iZVS using both charge conservation (Eq. 13) and energy conservation (Eq. 30). The two derivations are claimed to be equivalent for the first time. Experimental validation with two CREE SiC MOSFETs in Table 1 reports an average 17-fold reduction in Eon prediction error relative to a conventional model. The paper also claims that the paradigm explains transistor switching 'solely from fundamental physical laws' without reliance on observed waveforms.

Significance. If the central prediction claim were established, a first-principles Eon model with a 17-fold error reduction over the conventional model would be a significant contribution to power electronics, with implications for loss modeling, device characterization, and circuit design. The qualitative emphasis on the complementary switch's nonlinear dynamics and on RS1 variation as a causal factor is a useful conceptual contribution, and the algebraic cross-check between charge- and energy-conservation derivations is a reasonable bookkeeping exercise. However, the quantitative core of the paper is not currently supported: the final model retains integrals of instantaneous switching waveforms whose a priori evaluation is never demonstrated. As presented, the experimental validation cannot be distinguished from a post-hoc energy-balance reconstruction, so the claimed predictive power and the 'solely from fundamental physical laws' statement are not substantiated.

major comments (3)

[Methods, Eq. (30) and Eq. (13)] The central Eon model is not a closed-form prediction. Both Eq. (13) and Eq. (30) contain the switching-interval integrals ∫ i_RS2 dt, ∫ v_ds,S2 i_L dt, and ∫ v_ds,S2 i_RS2 dt, along with capacitance-energy terms. These integrals depend on the instantaneous S2 voltage/current waveforms over the same interval whose dissipated energy the model is supposed to predict. The paper never expresses these integrals as functions of quantities known before the switching event (VDC, ΔV, load current, gate-drive conditions, datasheet capacitances). Thus Eq. (30) is an energy-balance identity, not a predictive formula. This directly undermines the Abstract's claim that the model is derived 'solely from fundamental physical laws' and the Experimental validation's claim of a 17-fold improvement.
[Experimental validation / Table 1] The provenance of the 'calculated values' in Table 1 is not disclosed. The paper states 'Code availability: No custom code' and 'Data availability: data available in this manuscript,' but no evaluation procedure is given for the waveform integrals in Eq. (30). If those integrals were evaluated from measured oscilloscope traces or LTspice simulations, then the reported errors (0.88%–11.60%) and the 17-fold improvement are a measure of how well the model reconstructs the measured Eon from measured waveforms, not a measure of independent predictive accuracy. Without a stated a priori evaluation protocol—e.g., formulas for the integrals in terms of datasheet parameters and operating conditions—the validation in Table 1 cannot support the paper's central claim.
[Methods, Eq. (12) and Eq. (25)] The charge-conservation derivation introduces lumped equivalent capacitances Cpar,S2 and Crr,S2 with no identification procedure. In particular, Crr,S2 is invoked in the model description as an 'equivalent reverse-recovery capacitance,' but the manuscript does not state how it is obtained or whether it is fitted from the same experimental waveforms. If Crr,S2 or Cpar,S2 is calibrated to the measured data used in Table 1, the model contains free parameters that are not accounted for in the error comparison. This is a separate, load-bearing gap in the claim of a parameter-free first-principles derivation.

minor comments (4)

[Abstract and Introduction] The phrase 'near a century since the transistor's invention' is an overstatement (the transistor was invented in 1947). The introduction also repeatedly claims that no prior work recognizes complementary-switch dynamics; such strong novelty claims should be softened or supported with a more detailed historical comparison.
[Throughout Methods] The equations are hard to verify because of inconsistent notation and apparent OCR artifacts. For example, subscripts such as 'par', 'Cpar', and 'Crr' are used without clear definition in every instance, and some integrals in Eqs. (4), (11), (13), and (30) are difficult to parse. The authors should rewrite the derivation with consistent notation and numbered definitions for all capacitance terms.
[Table 1] The 'average 17-fold reduction' is reported without a standard deviation or definition of averaging. The individual ratios vary from 4.8 to 58.8, so the average is not representative unless accompanied by dispersion measures. The table also omits test conditions such as gate resistance, load current, and temperature, which are needed to reproduce the measurements.
[Data and code availability] The statements 'Data availability: data in this manuscript' and 'Code availability: No custom code' are inconsistent with a model whose evaluation requires the waveform integrals of Eq. (30). The authors should specify exactly how the calculated values in Table 1 were produced, including any simulation or post-processing steps.

Circularity Check

1 steps flagged

The proposed Eon 'prediction model' leaves the switching-interval waveform integrals unevaluated, so the claimed 17x improvement in Table 1 is a post-hoc reconstruction rather than a first-principles prediction.

specific steps

fitted input called prediction [Methods, 'Case-2 iZVS Eon prediction model derivation', Eq. (30) and identical Eq. (13); Experimental validation, Table 1]
"It is noteworthy that (30) is identical to (13), indicating that the proposed prediction model derived from the law of energy conservation is fundamentally equivalent to that obtained from the law of charge conservation, both grounded in the physical insights of the proposed paradigm."

The predicted quantity is defined in Eq. (1) as Eon,S1 = ∫ vds,S1 iRS1 dt over the switching interval. The final 'prediction model' in Eq. (13)/(30) still contains switching-interval integrals of iRS2, vds,S2 iL, and vds,S2 iRS2 (the DC-source work term VDC∫iRS2 dt, the load-inductor work term ∫vds,S2 iL dt, and the S2 shoot-through dissipation ∫vds,S2 iRS2 dt). These are exactly the waveform-level quantities that must be observed or simulated during the turn-on event being predicted. The paper supplies no procedure for evaluating these integrals from pre-switching inputs (VDC, ΔV, gate-drive conditions, datasheet capacitance/charge curves), and 'Code availability: No custom code' makes the calculation unrepeatable. The Table 1 'calculated values using proposed prediction' therefore reduce

full rationale

Most of the manuscript is a qualitative causal narrative about RS1 variation, the complementary switch's non-linear dynamics, and the Miller platform. That narrative is self-contained and does not rest on any load-bearing self-citation. The central quantitative claim, however—the average 17-fold error reduction over the conventional model—is not supported as a prediction. The derivation starts from Eq. (1), which defines Eon as a switching-interval waveform integral, and after KCL/KVL substitutions and capacitance-energy approximations arrives at Eq. (13)/(30), which still contains switching-interval waveform integrals of the same event. No a priori evaluation of those integrals is given, and the only data made available are the final numbers in Table 1. The claimed equality between the charge-conservation and energy-conservation derivations is an algebraic identity of the same bookkeeping, not an independent cross-validation. Thus the 17x improvement is, as stated, a reconstruction of the measured/simulated switching waveforms rather than a derivation 'solely from fundamental physical laws.' Score 7: the central claim reduces by construction to its own waveform inputs, but the qualitative switching analysis retains independent content and there is no self-citation circularity. (Data/Code availability statements: 'The data presented in this study are available in this manuscript' and 'No custom code was used in this study'—neither provides a reproducible evaluation procedure.)

Axiom & Free-Parameter Ledger

1 free parameters · 5 axioms · 2 invented entities

The central claim rests on standard circuit laws plus several domain assumptions about where and when losses occur. The free parameter Crr,S2 is introduced to represent reverse recovery. The model also leans on waveform-level integrals that are not derived from first principles, which is the main circularity concern.

free parameters (1)

Crr,S2 equivalent reverse-recovery capacitance = not specified
Introduced in Eq (13)/(30) to represent reverse-recovery charge of the complementary switch S2; its value and voltage dependence are not derived from first principles and would need to be extracted from device characterization or simulation.

axioms (5)

standard math Kirchhoff's current and voltage laws apply to the half-bridge equivalent circuit
Used throughout Methods, e.g., Eq (2), Eq (7), Eq (8), Eq (22).
domain assumption The entire Eon,S1 is incurred in the nonlinear resistance RS1
Methods, before Eq (1): 'it is valid to assume that the entire Eon,S1 is incurred in RS1'. This is a modeling assumption about where energy dissipates.
domain assumption Vast majority of energy dissipation during S1's switching-ON occurs during (t0-t3)
Methods, before Eq (1) and Eq (14): used to truncate the integration interval to (t0-t3). No quantification or sensitivity analysis is provided.
domain assumption vgs,S1 and vgs,S2 are negligibly small compared to ΔV and VDC-ΔV during (t0-t3)
Methods, near Eq (17): used to approximate gate voltages to 0 V in the energy-balance expressions. The validity depends on gate-drive levels and switching times.
domain assumption Energy dissipated in ESR of CS1 is negligible compared to losses in RS1
Methods, before Eq (1): stated without numeric support, requiring that capacitor series resistance losses are negligible.

invented entities (2)

Crr,S2 equivalent reverse-recovery capacitance no independent evidence
purpose: Lumped representation of reverse-recovery charge of the complementary switch in the equivalent circuit
The paper does not provide a measured value or a first-principles derivation for this capacitance; it is introduced to extend the equivalent circuit model.
RS1 nonlinear equivalent resistance independent evidence
purpose: Lumped nonlinear resistance representing all non-displacement current paths of the switch under study
Defined as RS1=vds,iRS1 and intended to be read from device output and third-quadrant characteristics; this has a datasheet-observable basis, so it is not a purely invented entity.

pith-pipeline@v1.3.0-alltime-deepseek · 19596 in / 11719 out tokens · 148298 ms · 2026-08-04T14:48:12.118604+00:00 · methodology

0 comments

Cite this review

Pith. "Pith review of Charge-Unified Semiconductor Switching Theory." pith.science (2026). https://pith.science/paper/TO2DPEYT

@misc{pith2026251017815,
  author       = {Pith},
  title        = {Pith review of: Charge-Unified Semiconductor Switching Theory},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TO2DPEYT}},
  note         = {Machine review of arXiv:2510.17815}
}

Share X Bluesky LinkedIn Reddit HN

read the original abstract

Semiconductors and their downstream applications sustain the electronic, information, energy and industrial systems underpinning modern society. Improving their sustainability is therefore an urgent global priority, particularly as global electricity generation is projected to increase more than 2.5 fold by 2050. Yet, since the invention of the transistor in 1947, a unified, global view of circuit elements as media for charge redistribution and transfer one that reveals switching inertia and the dynamical nature of switching while connecting microscopic and macroscopic domains across the semiconductor value chain through a common theoretical language has remained absent. Switching consequently lacks a unified mechanistic account of its physical origins and spatiotemporal evolution, with fundamental disconnects between charge- and energy-conservation frameworks, among carrier dynamic mechanisms and across equivalent-circuit formalisms. These limitations fragment research domains and impede sustainability gains, particularly those requiring cross-domain causal information. Here, we present Charge-Unified Semiconductor Switching Theory (CUSST), a general theory that unifies circuit elements through a charge-mediated view, reveals switching inertia and the dynamical nature of switching, bridges these long-standing disconnects and establishes a unified conceptual, mechanistic, formal and analytical framework. Through these unifications, CUSST provides an unusually simple representation of otherwise fragmented switching phenomena. It establishes a unified micro-macro spatiotemporal view of switching, generalizes circuit theory, extends the application of conservation laws and provides a foundation for developing new theoretical systems.

discussion (0)

Reference graph

Works this paper leans on

4 extracted references

[1]

14V 0.2 10 10m 0.4 14V vgs,S1 RS1 -6V 4V RS1 (e) -6V 4V0 vgs,S1 5 5

Unified equivalent -circuit model ing across device types, switching scenarios and switching phases (detailed in Methods) — including a non - linear equivalent resistance (e.g., RS1 model), consistent with both semiconductor physics and Ohm’s law, and equivalent capacitance models (e.g., Crr,S2 model) to represent reverse-recovery charge. 14V 0.2 10 10m 0...
[2]

RS1 variation as the primary cause — defining the switching -ON criterion (the initial rapid drop of RS1 when vgs,S1 exceeds the threshold; detailed in Methods), and governing the entire turn-on process as the main driver
[3]

Coupling effects of the complementary switch ’s non -linear dynamics — incorporating non-linear junction-capacitance transitions of the complementary switch as essential inputs beyond the recognition of the conventional analysis
[4]

Figure 2

Revealing the role of load current — exemplified by explicit recognition of the existence of CC in iZVS and its associated dissipated energy, as well as the participation and work done by the load inductor throughout the process, both beyond recognition of conventional analysis. Figure 2. (a) Simulated waveforms during a typical ZVS process in LTspice. (b...

arXiv 2025

[1] [1]

14V 0.2 10 10m 0.4 14V vgs,S1 RS1 -6V 4V RS1 (e) -6V 4V0 vgs,S1 5 5

Unified equivalent -circuit model ing across device types, switching scenarios and switching phases (detailed in Methods) — including a non - linear equivalent resistance (e.g., RS1 model), consistent with both semiconductor physics and Ohm’s law, and equivalent capacitance models (e.g., Crr,S2 model) to represent reverse-recovery charge. 14V 0.2 10 10m 0...

[2] [2]

RS1 variation as the primary cause — defining the switching -ON criterion (the initial rapid drop of RS1 when vgs,S1 exceeds the threshold; detailed in Methods), and governing the entire turn-on process as the main driver

[3] [3]

Coupling effects of the complementary switch ’s non -linear dynamics — incorporating non-linear junction-capacitance transitions of the complementary switch as essential inputs beyond the recognition of the conventional analysis

[4] [4]

Figure 2

Revealing the role of load current — exemplified by explicit recognition of the existence of CC in iZVS and its associated dissipated energy, as well as the participation and work done by the load inductor throughout the process, both beyond recognition of conventional analysis. Figure 2. (a) Simulated waveforms during a typical ZVS process in LTspice. (b...

arXiv 2025

Charge-Unified Semiconductor Switching Theory

Core claim §

What carries the argument

Load-bearing premise §

What would settle it §

If this is right

Where Pith is reading between the lines

discussion (0)

Core claim

Load-bearing premise

What would settle it