Skip to content
RFrftools.io

FDTD Transmission Line Simulator

Solve microstrip stubs, coupled-line filters, via transitions, and step discontinuities across a ladder of named solve modes — from a 2D cross-section in seconds to a fine 3D openEMS field solve. S-parameters, Smith chart, and impedance analysis.

Cross-section
3.0 mm1.6 mm
Plan view · not to scale
30 mm15 mm
Z₀
50.04 Ω
near 50 Ω
ε_eff
3.3312
λg at centre
68.44 mm
Stub electrical length
78.9°
Solve mode

Checking which modes your account can run…

How It Works

FDTD (Finite-Difference Time-Domain) solves Maxwell's curl equations directly on a 3D Yee mesh by time-stepping electric and magnetic field components. A Gaussian pulse excitation covers the entire frequency band in a single simulation run, and S-parameters are extracted from port voltages and currents via Fourier transforms.

Solving is a choice among four named modes, not a mesh-density knob. Instant is a 2D cross-section field solve cascaded with published discontinuity models — engineering accuracy in seconds, and explicitly not a full-wave simulation. Express, Normal and Fine are all real 3D FDTD runs through openEMS with CSXCAD geometry, differing in how finely the trace and substrate are resolved and how long the time record runs. Instant and Express are available on every tier; Normal and Fine are Pro, and Fine carries a monthly allowance.

A mode your tier cannot run is rejected, not quietly downgraded — a fine solve is never returned wearing a coarser mesh. Every result states the mode requested and the method that actually produced it, and if a full-wave solve could not run, the result is marked degraded and says why rather than being presented as full-wave.

Four structure types are supported:

  • Microstrip Open Stub — A parallel open-circuit stub produces a frequency-selective reflection notch at the quarter-wave resonance of the stub.
  • Coupled-Line Filter Section — Edge-coupled microstrip lines with even/odd mode impedances form a bandpass section.
  • Through-Via Transition — A cylindrical via modelled as series inductance plus shunt pad capacitance; critical for signal integrity above a few GHz.
  • Step Discontinuity — A change in trace width creates fringing capacitance and a partial reflection; quantified by the mismatch S11.

Related Calculators

Read the Guide

FAQ

How long does a simulation take?+

It depends on the solve mode and on the hardware your tier runs on, so each mode card quotes the runtime measured for your own account rather than a general figure. Instant is seconds; the 3D modes run from a couple of minutes to several. No runtime is advertised for a mode that has not been measured on the hardware it would run on.

Which solve mode should I pick?+

Start with Instant while you are still moving dimensions around — it is a 2D cross-section solve with published discontinuity models, fast enough to iterate against, and it reports itself as quasi-static rather than full-wave. Move to Express once the geometry is roughly right and you want a real 3D field solve, then Normal or Fine when you need the trace and substrate resolved more finely.

What happens if I pick a mode my tier does not include?+

The job is rejected with the reason, not silently downgraded. An earlier version clamped an out-of-tier request to a coarser mesh and returned it under the requested mode’s label, which meant you could not tell what you had actually been given.

Why does a via transition offer fewer modes?+

The full-wave modes need a three-dimensional model of the structure, and the through-via transition does not have one — it is solved from a lumped series-inductance and shunt-capacitance model instead. Rather than accept the job and hand back a quasi-static answer under a full-wave label, the tool marks those modes unavailable for that structure and says why.

Which substrate should I choose?+

FR-4 is the standard low-cost PCB substrate (εr ≈ 4.4, tan δ ≈ 0.02). Rogers 4350B and Rogers 3003 are high-frequency laminates with tighter tolerances and lower loss — preferred for designs above 5 GHz.

Why does the stub notch not appear at the expected frequency?+

The notch frequency depends on the effective dielectric constant, not just the physical length. The Hammerstad-Jensen model automatically accounts for the microstrip fringing field, but you may need to reduce stub length slightly to hit a precise target due to the T-junction effect.