Heuristic RCS · radar range equation · notional radar seeds

Fighter RCS Estimator

Order-of-magnitude radar cross section from airframe class, aspect, radar band and external stores, with the detection-range effect via the radar range equation. Treat outputs as bands, not numbers.

σ total

–

dBsm

–

Detect range

–

Airframe clean, frontal, X-band

Type

Geometry and band

Aspect off the target's nose0°

0 nose4590 beam135180 tail

Radar band LO surface condition

Loadout external stores

Store (incl. pylon/launcher)m² eachQty

Weapon bay doors open (internal-carriage types; transient spike)

Internal weapons with bay doors closed add nothing. Store values are editable; the low/high band scales with your edit.

Radar model

Notional seeds

Transmitter and waveform

Peak power (kW) Pulse width (µs) PRF (kHz)

Antenna

Frequency (GHz) Aperture dia (m) Efficiency (%)

Gain override (dBi) ESA scan angle (°)

Receiver and losses

Noise figure (dB) System losses (dB)

Timing

Track dwell (ms) Search frame time (s) Search azimuth (°) Search elevation (°)

Detection criterion

Pd Pfa Target model

R⁴ = P_avg · T_dwell · G² · λ² · σ ÷ ((4π)³ · k · T₀ · F · L · SNR_req)

P_avg = P_peak × pulse width × PRF. Search dwell = frame time ÷ beam positions in the scan volume.

Result

–

Airframe Stores Bay

Detection range (nominal): –
Band (low–high σ): –
Search range (nominal): –
Same jet, clean, same aspect: –
Range penalty from loadout: –

Range vs RCS noise-limited

Track Search This target, low–high σ x: σ in m² (log) · y: nm

Aspect profile dBsm, nose up

Configured Low–high band Clean airframe

How the model works, and where it is weak

Airframe baselines are commonly cited open-source estimates and manufacturer claims, not measured data. Real figures are classified. Treat the low–high range as the honest answer.
Aspect, band and stores factors are my engineering heuristics, not sourced values. They encode the right physics direction (beam flashes, resonance at low frequency, stores dominating LO jets) but the magnitudes are assumptions.
Summation is incoherent: σ adds as power. Stores carry an installation factor of 1.0 / 1.2 / 1.5 (low / nominal / high) for pylon junctions and multi-bounce. Real RCS is a coherent, wildly scintillating function of angle; a 1° change can move it 10+ dB.
Band matters most for shaped jets. Shaping works when features are many wavelengths long; at VHF/UHF, tails and edges approach resonance and the advantage collapses. Legacy jets barely change across bands in this model.
Elevation is ignored. Look-up/look-down and the planform from above or below change RCS a lot, especially with stores and open bays.
Reference mode scales range as R ∝ σ^1/4 from one known point. Halving range needs 16× less σ.
Physics mode solves the radar range equation in energy form with T₀ = 290 K. Only average power × dwell sets noise-limited range; PRF and pulse width enter through duty cycle. Integration over the dwell is treated as ideal coherent. Real radars mix coherent and non-coherent integration and switch PRFs; put that shortfall into system losses.
Required SNR uses Albersheim's approximation for a steady target, or the exact single-look Swerling 1 result (SNR = ln Pfa ÷ ln Pd − 1) for a fluctuating one. Swerling 1 is the default because aircraft scintillate.
Antenna gain = η(πD/λ)². Beamwidth from θ² ≈ 26,000 ÷ G (deg²). ESA scan loss is cos θ one way (cos² two way), applied to track only. Search beam count ignores overlap and overhead; a mechanically scanned radar's search and TWS dwell are the same thing.
Radar seeds are notional. They give plausible orders of magnitude and are not taken from any real system.
Not modelled: clutter (look-down is usually clutter-limited), jamming, radar horizon and terrain masking, atmospheric attenuation, multipath and lobing, ambiguity-resolution losses, and cumulative detection over several scans. Treat physics ranges as a noise-limited upper bound.
Sphere equivalent is a perfectly conducting sphere in the optical region (σ = πr²). Below ~10 cm diameter at X-band the sphere is resonant, so tiny equivalents are illustrative only.