2D Layout
Stackup
Simulate
3D / Pattern
Matching
Tuner (VNA)
Signal Integrity
Report
x — , y —
Layer stackup
Top of table = top of board. Imported .PcbDoc stackups land here; edit εr / thickness / material freely. Copper rows are bound to layout layers.
Presets:
Simulation
–
GHz
| on | centre (GHz) | bandwidth (MHz) | |
|---|---|---|---|
| T1 | |||
| T2 | |||
| T3 | |||
| T4 |
Advanced settings
How finely the mesh resolves the narrowest copper trace (cells across it). Too coarse makes thin traces look wider and the resonance read too high. Set by Accuracy.
Largest grid the run may use. If the model needs more, the mesh is coarsened automatically. Each million cells needs about 50 MB of GPU memory. If the GPU drops out the run is retried once at half the cells. Set by Accuracy.
mm
Free space between the model and the absorbing boundary of the simulation box. 0 = auto (one eighth of a wavelength). More air gives a cleaner pattern and efficiency but costs cells.
How long the time-domain run may last. auto = the excitation pulse plus about 10 ns of ring-down, stopping early once the field has died out. Enter a number only if the log says the field had not decayed.
auto runs on the graphics card (WebGPU) and falls back to the processor. The CPU is many times slower — use it only to cross-check a result.
Uniform uses the finest step everywhere (default). Graded keeps the fine step at copper edges and detail and grows cells up to 1 mm elsewhere — about 2.5× faster; in validation the resonance landed within 1 % of the uniform mesh. Use graded to iterate, uniform for final numbers.
Models copper with its real conductivity (skin effect), so efficiency includes conductor loss. Off = perfect conductor.
Records the live wave-propagation animation. Off makes runs faster.
mm
mm
mm
Ready.
S11 (return loss) · hover for a frequency readout
Smith chart (Z₀ ref)
Wave propagation (Ez just above top copper)
runs live during simulation
no pattern yet — run a simulation
drag = orbit · wheel = zoom
Antenna matching network
Ladder from the port towards the antenna (element 1 is closest to the feed). Values are ideal lumped components applied to the simulated port impedance. Assign each element to an imported landprint/pad for documentation.
No matching elements — S11 shown as simulated.
S11 — raw vs matched · hover to read, click to pin a cursor
Smith — raw vs matched
Target bands
up to 4 fixed guides, always shown while tuning — the same bands as in the Simulate tab
| on | CF (GHz) | BW (MHz) | from band | verdict | |
|---|---|---|---|---|---|
| T1 | |||||
| T2 | |||||
| T3 | |||||
| T4 |
cursors ("taps"): click the S11 plot — click a cursor again to remove it · each cursor also gets gain (dBi) on the next ▶ Run
Impedance tuner — measured circuit (VNA)
Standalone bench tool, independent of the simulation. Enter the impedance you measured
(or click the Smith chart / import a Touchstone .s1p), describe the matching network that is already
fitted, then tweak or add components until the source sees the impedance you want.
Element #1 is closest to the source, the last element is closest to the load.
Source ZS
+j
Ω
Load ZL
+j
Ω
add:
Smith chart — click to set ZL
Measured points · |Γ| vs frequency
Signal integrity — impedance along a conductor
Pick a source pad and a destination pad (dropdowns, or click pads in the board view). The copper route between them is traced
through tracks, vias and — optionally — series 2-pin components; each piece gets its characteristic impedance from the stackup and the reference plane
actually under it, and the cascade is solved for TDR, step response and return loss. Independent of the antenna simulation.
Ω
%
Ω
Ω
pF
ps
GHz
Board — click a pad: · wheel = zoom, drag = pan
■ within tolerance ■ up to 2× ■ beyond · ■ via / component
Impedance along the conductor · sections (model) and TDR as seen with the chosen rise time
Step response · voltage at source pad and at destination pad, 1 V driver step
Frequency response · return loss at the source, insertion loss to the load
Run a simulation to populate the report.