RF Physics for DeWi Investors (Pt 1)

Why this matters

Networks are at least as valuable as the sum of their parts. If we understand the value of individual deployments, we can extrapolate a lower bound for the entire network. This Part One studies the relationship between costs and coverage; Part Two will study coverage and value.

What is a wireless deployment?

• Radios generate electromagnetic signals on integrated circuits.
• Antennas couple those signals into 3‑D space to extend range.
• Towers (or any elevated structure) lift radios/antennas to widen useful coverage.

Costs of deployments

Telcos spend ~$50B/yr in US capex; each deployment mixes up‑front and recurring costs.

Up‑front

Radios. Differentiated by EIRP, power draw, bands, channel width, throughput, max connections, and size/weight. “Carrier‑grade” SDR small cells (e.g., Airspan AirStrand 2200) can cost $50K+, but DeWi focuses on mass‑market radios (~$750–$7.5K) to broaden participation. See AirStrand 2200 and Pollen shop.

Antennas. Think of gain as a vector field: beamwidth trades off coverage vs capacity. Antenna gain (dBi) is logarithmic: +5dBi ≈ 3× amplification; +17dBi ≈ 50× along the main lobe. Highly directional antennas paired with high‑frequency links enable PtP mesh at 10 Gbps over ~10 km on 60–80 GHz (example vendor).

Other up‑front. Mounts, cabling, casings, licensing, and installation (often 2–4 technician hours) vary and are secondary drivers.

Recurring

Rent. Consumer installs can be $0; hosted models share 20–70% of rewards (e.g., Fairspot, HeliumDeploy); traditional tower space may be $100/mo (rooftop) to $1,000+/mo (towers), with change fees and annual escalators.

Backhaul. A 10 Gbps circuit may cost $1–2K/mo and be resold to thousands of households via an IXP path; poor utilization drives unit costs higher (see WISP failures like Starry).

Power. Outdoor DeWi radios at ~50–60 W cost roughly $5–6/mo at $0.15/kWh; globally, telco power use is material (~470 TWh).

Other. Maintenance, insurance, administration as applicable.

From costs to useful coverage

Useful coverage is driven by three things:

1. How powerful is the equipment? (TX power + antenna gain + RX gain, subject to regs)
2. What spectrum is used? (frequency ↔ wavelength ↔ capacity/propagation)
3. How high is the antenna vs surroundings? (line‑of‑sight geometry)

1) Equipment power & FCC limits

Engineers use dBm: 0 dBm = 1 mW; +10 dBm = 10×; +30 dBm = 1 W. In CBRS, Class A devices are limited to 30 dBm (1 W), Class B to 47 dBm (50 W). Licensed adjacent bands allow far higher EIRP (72–75 dBm). With a +17 dBi antenna on a 30 dBm radio, EIRP hits the 47 dBm ceiling; receivers (phones) can add ~+3 dBi.

Selected readings: AT&T 62 dBm ask, DISH 72 dBm ask, CTIA paper, industry letters.

2) Spectrum 101

Frequency determines wavelength (wavelength(m) ≈ 300 / frequency(MHz)). Lower bands (e.g., 700–900 MHz) propagate far and resist interference but have limited throughput; higher bands (e.g., 3.5 GHz CBRS, 24–39 GHz mmWave) offer high throughput but short range and higher susceptibility to interference/attenuation. See calculator and mmWave deployment.

3) Path loss (distance × frequency)

Log‑domain model: PathLoss(dB) = 20·log10(d[km]·f[MHz]) + 32.44. Example: at 0.1 km, 900 MHz loses ~72 dB; at 3.6 GHz, ~84 dB. Above 11 GHz, loss worsens by another order of magnitude — densification needs explode.

Interference & attenuation

Higher frequencies magnify both. Rain fade: ~1 dB/km at 10 GHz vs ~0.01 dB/km at 3 GHz. Concrete walls: −10 dB at 2.4 GHz vs −20 dB at 5 GHz.

Tower height & line‑of‑sight

Elevation clears local obstructions. A 100‑ft tower ~1 mi away from a 20‑ft house (5× taller) yields a “shadow” extending ~⅕ mi behind the house; a 400‑ft tower reduces the shadow to ~1⁄20 mi. In dense high‑rise cores, building‑side radios beat ever‑taller towers; in flat terrain, 400‑ft structures can maintain LOS for miles (subject to path loss).

Implications for CBRS

CBRS (3.55–3.70 GHz) vs mid‑band 2.3–2.5 GHz: ~16 dB lower power limits plus ~3–5 dB extra path loss → ~20 dB deficit. Empirically, OpenSignal observed ~15 dB deltas — yet faster DL speeds on CBRS due to available spectrum. Density is the lever.

Takeaways

• High‑frequency spectrum (incl. CBRS) is the path to 100+ Mbps for the masses — but it requires at least an order‑of‑magnitude densification under legacy models.
• CBRS is valuable already, but FCC EIRP caps (47 dBm) constrain potential vs adjacent licensed bands (72–75 dBm).
• Retail‑priced DeWi radios will be eclipsed within 3–5 years on capacity, connections, and power — plan for refresh cycles.

Looking ahead (Part Two)

Next: how much is coverage worth? Consider density, demographics, capture rates, elasticity, pricing currency (inflation), and more.