How a sweep is set up 01 Load a netlist or preset. The RC low-pass preset already contains .ac dec 40 10 1meg, so the tab opens in AC mode. 02 Pick the sweep law. dec places points at a fixed number per decade, oct per octave, lin evenly in hertz. dec 40 over 10 Hz to 1 MHz gives five decades, so 201 points.03 Name the probes. The output probe defaults to V(out); an optional reference probe turns the absolute response into a ratio such as a transfer function. 04 Read the plot and the corner. The status line names the −3 dB frequency; for a first-order low-pass it should land at 1/(2πRC) — here 1 kHz — with −45° of phase there.
Worked readouts The RC preset (1 kΩ, 159.1549 nF) swept with dec 40 from 10 Hz to 1 MHz:
Frequency
Magnitude
Phase
100 Hz
−0.0432 dB
−5.711°
1 kHz
−3.0103 dB
−45.000°
10 kHz
−20.0432 dB
−84.289°
100 kHz
−40.0004 dB
−89.427°
The interpolated corner reads 1000.097 Hz with −3.0107 dB and −45.003° ; from 10 kHz to 100 kHz the magnitude falls 19.9572 dB , the textbook one-pole slope. Point counts behave as the law says: dec 10 from 10 Hz to 1 kHz gives 21 points; oct 8 from 100 Hz to 800 Hz gives 25 ; lin 5 from 100 Hz to 500 Hz gives 5 .
A series RLC stage (R = 20 Ω, L = 1 mH, C = 1 µF, output across C) shows the other shape: f₀ = 5032.921 Hz , ζ = 0.316228 , Q = 1.581139 , a peak of +4.4351 dB at the best dec 40 grid point, 4466.836 Hz (the analytic peak for this ζ is +4.437 dB at 4501.6 Hz ), and half-power points at 2249.18 Hz and 5954.78 Hz .
Why the points are not evenly spaced A Bode plot is drawn against log frequency, so a sweep that is even in hertz spends most of its points at the top of the range and almost none where the breakpoints live. dec N fixes N points per decade of ratio instead, so a corner is captured with the same resolution whether the range is 10 Hz–1 MHz or 1 kHz–100 kHz. Magnitude is 20·log₁₀|H|, so ten times the response adds exactly 20 dB — the RC rows step by about 20 dB per decade. Phase is unwrapped before plotting so the curve does not jump as it passes ±180°.
The −3 dB corner is not a grid point. Both probes are computed at every frequency, and the corner is interpolated against log frequency between the two samples that bracket −3.0103 dB . That is why it reads 1000.097 Hz rather than a round 1000, and why the same routine returns nothing when a target is never crossed instead of inventing a frequency.
Limits and edges One sweep, linearised. The analysis uses a single small-signal operating point; it cannot show large-signal distortion, slew limiting or noise. Sweep a source with the DC sweep tab to see how the operating point moves.
Spacing is a modelling choice. Coarse sweeps miss narrow peaks entirely — a high-Q resonance needs enough points per decade to be visible, and the peak height read from a coarse grid is always slightly below the true value because the exact peak frequency rarely falls on a point. What it does not do. It does not integrate in time or transform measured samples; use the transient tab for waveforms and the Science Lab for spectra of real data.
Reading a corner without over-reading it Two habits cause most misreadings. The first is quoting the corner to four decimals: 1000.097 Hz is an interpolation between sampled points, so the digits past 1000 belong to the sweep density, not to the circuit. The second is treating the resonant peak as Q. The peak sits below f₀, but its reported position and height are partly the sweep's: the dec 40 grid's best point is 4466.8 Hz at +4.4351 dB , while the analytic peak for ζ = 0.316228 is +4.437 dB at 4501.6 Hz . The spacing of the half-power points is not exactly f₀/Q for finite Q. Use the closed-form card for ζ, Q and f₀, and the sweep for the shape.
Where it earns its place Checking a filter before lab The RC preset answers the standard questions in one pass: where the corner is, what the slope above it is, and how much phase lag the stage adds at the frequency that matters. If the measured corner differs from 1 kHz , the components or the loading are the first suspects, not the solver.
Finding the shape of a resonance The RLC sweep gives a peak of +4.4351 dB at 4466.8 Hz and a bandwidth of 3705.6 Hz between its half-power points. That is the data a damping resistor is chosen against: doubling R halves the peak height and flattens the phase transition.
Privacy Netlists, sweeps and plots are computed inside your browser; nothing is uploaded.
References
Wikipedia, Bode plot , en.wikipedia.org (访问日期:2026-10-07)— dB magnitude and log frequency.
Wikipedia, RC circuit , en.wikipedia.org (访问日期:2026-10-07)— the one-pole corner and −45° phase.
Wikipedia, RLC circuit , en.wikipedia.org (访问日期:2026-10-07)— natural frequency, damping and Q.
Wikipedia, SPICE , en.wikipedia.org (访问日期:2026-10-07)— the .ac directive and its sweep suffixes.
MIT OpenCourseWare 6.002, Circuits and Electronics , ocw.mit.edu (访问日期:2026-10-07)— frequency-response fundamentals.
Calculators in this hub
Hand-picked tools, one click away. The mini versions compute live and carry your values into the full calculator.
Sources & review
Reviewed by CalcX Editorial Team
Updated 2026-10-07