This calculator was generated by Claude (Anthropic) as part of the Wireless 101 course materials. Nearly every number on this page is computed live from a stated formula — there is no lookup table anywhere in it. Three kinds of number appear, and each is labelled where it is used:
Computed — all path loss, the noise floor, sensitivity, received power, the margin, the maximum range and the coverage probability. Published figure — the physical constants (k = 1.380649×10⁻²³ J/K, c, T₀ = 290 K), the model coefficients, the model validity ranges, and the standard channel bandwidths in the bandwidth menu. Illustrative — the equipment parameters in four of the seven presets, which are marked as such in the menu and in the line under the knobs.
It is a model evaluator, not a measurement. Three of the presets reproduce this course's own worked examples digit for digit — the 2.4 GHz office link and the 900 MHz uplink/downlink pair of M9-L4 — and a checker asserts that they still do. A propagation model asked a question outside its validity range does not fail loudly; it returns a confident number no measurement supports, so the range warnings here are part of the interface, not a footnote.
Models: H. T. Friis, A Note on a Simple Transmission Formula, Proc. IRE 34(5):254–256, 1946 · Y. Okumura et al., Rev. Elec. Commun. Lab. 16(9–10), 1968 · M. Hata, Empirical Formula for Propagation Loss in Land Mobile Radio Services, IEEE Trans. Veh. Technol. 29(3):317–325, 1980 · COST Action 231 final report, EUR 18957, 1999 · J. B. Johnson and H. Nyquist, Phys. Rev. 32, 1928 (thermal noise) · M. Abramowitz and I. A. Stegun, Handbook of Mathematical Functions, 26.2.17 (the normal CDF approximation) · P. J. Acklam's rational inverse-normal approximation. Course sources: M7-L2 antenna gain, M8-L1 Friis and EIRP, M8-L4 log-distance, Hata, COST-231 and shadowing, M9-L1 kTB, noise figure and sensitivity, M9-L4 the budget and its margins.
Where did the power go, how far does the link reach, and how much does the choice of propagation model change the answer? Move any knob and the waterfall, the four models on one pair of axes, the coverage ring and the noise floor all recompute. Every decibel is evaluated from a formula printed on this page.
The waterfall below is the detailed view; this is the map. Every thumbnail is drawn from the same computation as the big panels, so nothing here is decoration — the path-loss block really plots loss against distance with your distance marked, and the comparison block really plots received power against sensitivity. Hover or focus a block and it will highlight its row of the waterfall and the knobs that move it.
The signal path — what arrives
The receiver’s own floor — what it must beat
Each bar is one line of the budget, starting at the transmit power and ending at what actually arrives. Gains climb, losses fall, and the two subtotals — EIRP and received power — are drawn as full columns from the baseline because they are absolute levels rather than steps. The dashed line is the receiver's sensitivity, computed in the noise panel below; the gap between it and the received power is the link margin. The table repeats every number, so the chart is not the only way to read it.
A waterfall chart of the link budget. The same values are listed in the table immediately below this chart.
This is the panel worth the most. All four models are plotted over the same distance range at the same frequency, so the vertical gap between them is the cost of choosing one. Distance is logarithmic, which is why every model is a straight line: they are all of the form “a constant plus a slope times log₁₀ d”, and the slope is 10n dB per decade. A curve is drawn dashed and pale outside its published validity range, and the warnings underneath say which range and by how much. Extrapolating quietly is the classic beginner error; this panel refuses to do it quietly.
The range at which the margin has been eaten exactly to zero, using the selected model and the required margin from the panel below. The ring is a plan view with the frame labelled, and the radius is solved in closed form by inverting the model — not searched for.
Nothing here is typed in. The floor is kTB evaluated with the SI value of Boltzmann's constant at the 290 K reference, converted to dBm once; the receiver adds its noise figure, and the demodulator demands its SNR.
A single row labelled “margin” hides which risk it covers, so it cannot be argued with or reduced when a measurement earns the right to reduce it (M9-L4). The shadowing row is the only one this page computes for you: shadow fading is log-normal, so the margin needed for a target cell-edge reliability is z σ with z = Φ⁻¹(p). Read the two together: the margin scales with σ, not with the path loss, so a better-characterised environment is cheaper to cover.