Antenna
Segmentation rules
- solve to check
Frequency sweep
Compare with a measurement
The measured angle is read as phi from +x on the x-y plane and as theta from +z on a vertical plane, so pick the plane the antenna was mounted to sweep; Rotate lines up the chamber's zero.
Both traces are normalized to their own peak before they are differenced — a simulated pattern is absolute dBi and a chamber cut is raw S21 through cables and fixture, so nothing else puts them on one axis. The difference is clamped at the plot floor, because a null one degree off its predicted angle differences to tens of dB and means almost nothing.
Export
The pattern file drops straight onto a run in the chamber
app: its importer matches the param, angle_deg,
freq_hz and gain_dbi columns by name.
Results
3D pattern
The radius in each direction is the gain in that
direction and the color is the same number, always in dBi. The
radius scale changes the shape, not the pattern: linear power is
the donut from the lesson, and a dB radius inflates everything far from
the peak until a dipole looks nearly round — useful for reading
sidelobes, misleading about beam shape. The wire is drawn to scale at the
center, so the nulls along its axis are the ones you can see. This is one
more RP card than a cut — NEC has always computed the
sphere.
Convergence
Convergence means the answer stopped moving under refinement. It is not the same claim as the answer being right.
NEC input file from the form
These cards are the actual interface to NEC, and they are
identical in 4nec2, xnec2c and this tool. Edit them here and press
Run these cards, or copy them into 4nec2. nec_lab reads
CM CE GW GE GN EX FR RP XQ EN — any number of wires,
any number of sources, and a perfect or real ground plane. Anything else
— a load, a transmission line, an arc — is refused by name,
because the rules, the energy audit and the studies on this page all
assume the structure is what those cards describe.