The nonlinear device: near-open at working voltage, overwhelming in a surge — current follows a power of the voltage.
I = C·Vα under the material state you select — the state is part of every result.
A varistor is a bilateral nonlinear resistor — in its modern form a metal-oxide varistor (MOV) of sintered zinc-oxide (ZnO) grains — whose current follows a power of the voltage, I = C·Vα (Book One B, Ch 23; the System Design book writes the same law I = k·Vα). The exponent α is a material state, and the books state it in bands, each carried here with its source: ZnO α 20–80 (Book One B, Ch 23); ZnO α 30–80 and legacy silicon carbide (SiC) α 3–7 (Book Three, Ch 10); ZnO α 30–50 (Book Seven, Ch 33) and α ≥ 40 (Book Seven, GEM passage). The two ZnO lower bounds — 20 in One B, 30 in Three — stand as a stated difference between the books; they are not merged, and the state you select is the state the result is computed in. At any point the device presents two resistances (Book One B, Ch 17): the static resistance R = V ÷ I, and the dynamic (differential) resistance dV ÷ dI = R ÷ α — the steepness of the law at the point, and the reason a small rise in voltage across a clamping device draws an overwhelming rise in current. Book Nine carries the device in working form, without an α value: the middle stage of a staged protection scheme (gas discharge tube → MOV → TVS) and the absorber in an HVDC breaker. One documented instance stands in two states and is carried in both: Book One A’s worked example (α = 35 as printed, at 230 V and 265 V) prints ≈ 268.8 A and ≈ 0.986 Ω, which do not recompute from the passage’s own law; the review’s recomputation gives I(230 V) ≈ 6.86 mA with R ≈ 33.5 kΩ and I(265 V) ≈ 0.977 A with R ≈ 271 Ω — a collapse of about two orders of magnitude, not the printed “more than eight.” Neither side is adopted here: this tool’s grounded values are the material bands above and the conversion bound below, and every figure it returns is computed from the values you enter, in the state you select.
Two points on the device’s curve determine α and C — the bench measurement of the same law.
From two points the law gives its own constants: α = ln(I2 ÷ I1) ÷ ln(V2 ÷ V1), and C follows from either point. The recovered α is then read against every band the books state — the reading is reported band by band, because the bands are the books’ statements, not this tool’s: an α can stand inside Book One B’s ZnO band and below Book Three’s at once, the two lower bounds standing as the stated difference they are. Points at the same voltage, or carrying the same current, do not determine α and are reported as such.
The same nonlinearity read as a frequency converter: the efficiency of the Nth harmonic is bounded.
Book Six (Ch 18) states the varistor in a second precept: not the surge device but a mode of nonlinear harmonic generation, bounded by Page’s and Pantell’s theorems — the conversion efficiency to the Nth harmonic cannot exceed ηmax = 1 ÷ N². The bound at the orders the book works: N = 2 gives 25% (−6.02 dB) and N = 3 gives 11.1% (−9.54 dB). The two precepts — device and mode — share the one nonlinear mechanism and are kept distinct here, as the books keep them: this bound governs a varistor worked as a converter, not a varistor worked as a clamp.