On this page
Scaling Laws & Dimensionless Numbers
1. Electrical size: $ka$
\[\boxed{ka=\frac{2\pi a}{\lambda}}\]- $ka\ll1$: quasistatic/small-scatterer regime.
- $ka\sim1$: full-wave effects are central.
- $ka\gg1$: electrically large structure.
This single ratio is useful across antennas, scattering, resonators, EMC, nanoparticles and spacecraft.
2. Electrical length of an interconnect
\[\theta=\beta l=\frac{2\pi l}{\lambda_g}.\]When phase varies appreciably along an interconnect, treat it as a transmission line rather than an ideal wire.
3. Wavelength with frequency
\[\lambda\propto f^{-1}.\]Doubling frequency halves free-space wavelength. That affects antenna size, array spacing, diffraction, waveguide cutoff, electrical length and scattering.
4. Far-field distance
\[R_{FF}\sim\frac{2D^2}{\lambda}.\]For fixed physical aperture $D$, far-field distance grows approximately linearly with frequency because $1/\lambda\propto f$.
5. Skin depth
For a good conductor,
\[\delta\approx\sqrt{\frac{2}{\omega\mu\sigma}} \propto(f\mu\sigma)^{-1/2}.\]A useful dimensionless parameter is conductor thickness divided by skin depth:
\[\frac{t}{\delta}.\]If $t\gg\delta$, current is concentrated near the surface.
6. Waveguide cutoff
For rectangular TE$_{10}$ in an air-filled guide,
\[f_c=\frac{c}{2a}.\]So cutoff scales as
\[f_c\propto a^{-1}.\]Smaller waveguide dimensions move the operating band upward.
7. Antenna aperture and gain
For effective aperture $A_e$,
\[G=\frac{4\pi A_e}{\lambda^2}.\]For fixed physical/effective aperture,
\[G\propto\lambda^{-2}\propto f^2.\]This is why statements such as “free-space loss always gets worse with frequency” need context. In Friis, fixed dimensionless gains give one scaling; fixed physical apertures give another.
8. Diffraction-limited beamwidth
A characteristic aperture beamwidth scales roughly as
\[\theta\sim\frac{\lambda}{D}.\]Larger electrical aperture gives a narrower beam.
9. Phased-array spacing
To avoid visible grating lobes over broad scan regions, spacing is commonly kept near or below
\[d\lesssim\frac{\lambda}{2}\]with the exact requirement depending on scan angle and geometry.
For a fixed physical array area, the number of half-wavelength-spaced elements increases approximately as $1/\lambda^2$.
10. Friis transmission
\[P_r=P_tG_tG_r\left(\frac{\lambda}{4\pi R}\right)^2.\]With fixed gains,
\[P_r\propto\lambda^2R^{-2}.\]With fixed physical apertures, each antenna gain itself scales approximately as $1/\lambda^2$, which changes the frequency dependence.
11. Radar range scaling
For a monostatic radar,
\[P_r\propto\frac{P_tG^2\lambda^2\sigma}{R^4}.\]Solving for detection range gives a fourth-root dependence:
\[R_{max}\propto P_t^{1/4}G^{1/2}\sigma^{1/4}.\]A 16× increase in transmitter power gives only about 2× range if all else remains fixed.
12. Doppler
For monostatic radial motion,
\[f_D=\frac{2v}{\lambda}\propto vf.\]Higher carrier frequency gives larger Doppler shift for the same velocity.
13. Range resolution
For bandwidth $B$,
\[\Delta R\approx\frac{c}{2B}.\]Resolution improves as bandwidth increases. Carrier frequency by itself does not determine this ideal range resolution.
14. Thermal noise
\[P_n=k_BT B.\]Therefore
\[P_n\propto TB.\]Doubling bandwidth doubles noise power (+3 dB). Ten times bandwidth adds 10 dB.
15. Resonance quality factor
\[Q\approx\frac{f_0}{\Delta f}.\]For fixed $f_0$, larger $Q$ means narrower bandwidth and longer energy-storage time. High $Q$ improves selectivity and field enhancement but can reduce response speed and tolerance to detuning.
16. Fresnel number
A useful diffraction parameter is
\[N_F=\frac{a^2}{\lambda L}.\]It distinguishes diffraction regimes for an aperture of characteristic radius $a$ observed over distance $L$.
17. Loss tangent
For dielectric loss,
\[\tan\delta=\frac{\epsilon''}{\epsilon'}.\]The dimensionless loss tangent helps compare material dissipation independent of geometry.
18. Reflection coefficient
\[\Gamma=\frac{Z_L-Z_0}{Z_L+Z_0}.\]| $ | \Gamma | $ directly gives reflected-to-incident voltage-wave amplitude ratio. Reflected power fraction is $ | \Gamma | ^2$ for the usual normalized case. |
19. Cyclotron motion versus collisions
For charged particles in a collisional gas,
\[\boxed{\frac{\omega_c}{\nu_{coll}}}\]is more informative than cyclotron frequency alone.
- $\omega_c/\nu_{coll}\gg1$: many radians of gyromotion between collisions.
- $\omega_c/\nu_{coll}\ll1$: collisions strongly interrupt magnetic deflection.
This is especially important when interpreting electron/ion motion in vapor cells.
20. Rabi coupling versus decoherence
\[\frac{\Omega}{\Gamma}\]compares coherent drive strength with linewidth/decoherence.
- $\Omega\ll\Gamma$: weak-drive regime.
- $\Omega\gtrsim\Gamma$: coherent splitting/dynamics become resolvable depending on system details.
21. Far-detuned drive
\[\frac{\Omega}{\Delta}\]| controls perturbative validity. When $ | \Omega/\Delta | \ll1$, a far-detuned AC-Stark approximation is often useful. As the ratio grows, full dressed-state treatment becomes increasingly important. |
22. Optical depth
A common resonant optical-depth scale is
\[OD\sim n\sigma L.\]Larger optical depth increases absorption/interaction but can also make propagation, reabsorption and nonlinear effects more important.
23. Atomic magnetometry
For a magnetic resonance,
\[\omega_L=\gamma B.\]A useful resolution scale is
\[\frac{\gamma B}{\Gamma}.\]This compares field-induced precession/splitting to the resonance linewidth.
24. Rydberg scaling with principal quantum number
For hydrogenic/high-$n$ intuition away from strong perturbations and resonances:
| Quantity | Approximate scaling |
|---|---|
| orbital radius | $n^2$ |
| adjacent-level spacing | $n^{-3}$ |
| radiative lifetime | roughly $n^3$ |
| neighboring-state electric dipole matrix element | roughly $n^2$ |
| DC polarizability | often roughly $n^7$ |
25. A model-selection checklist
Before solving:
| Question | Useful ratio |
|---|---|
| Is the object electrically small? | $ka$ |
| Is an interconnect distributed? | $\beta l$ |
| Is conductor loss surface dominated? | $t/\delta$ |
| Am I in the far field? | $R\lambda/D^2$ |
| Is a resonance narrow/slow? | $Q$ |
| Is magnetic particle motion collision dominated? | $\omega_c/\nu_{coll}$ |
| Is atomic driving coherent enough to resolve? | $\Omega/\Gamma$ |
| Is far-detuned perturbation valid? | $\Omega/\Delta$ |
| Is optical propagation weak or strong? | $OD$ |
That table is often a better starting point than choosing an equation by memory.
Advanced scaling — sensitivity is not determined by one exponent
Scaling laws are valid only while the underlying model assumptions remain unchanged. A nominally high-order antenna gain or atomic polarizability scaling does not translate directly into a whole-system sensitivity improvement if bandwidth, matching, readout noise or dynamic range changes simultaneously.
For a receiver with equivalent noise temperature (T_{\mathrm{sys}}) and effective bandwidth (B), available noise power is
\[P_n=k_BT_{\mathrm{sys}}B.\]For coherent narrowband measurements, integrating longer reduces estimation variance only when the underlying signal remains coherent and the noise process supports that averaging law. Technical drifts, frequency instability and flicker noise can prevent indefinite improvement.
Advanced exercise: compare measured SNR versus averaging time against the expected white-noise scaling, then compute Allan deviation to identify the timescale where drift begins to dominate. When reporting a claimed sensitivity scaling with Rydberg principal quantum number, hold the relevant optical linewidth, atomic density, RF polarization, measurement bandwidth and calibration assumptions explicit.
Further reading: NIST measurement resources; Allan deviation reference.