Wireless Link Planning & RF Engineering

Overview

The three parts were originally documented separately. They are presented together here because a real link decision needs all of them: a good link budget cannot rescue an obstructed path, and a clear path does not prove there is enough signal at the receiver. Each part below states its inputs, its method and the evidence a reviewer should expect to inspect.

Planning workflow

Each question is answered with its own method, then recorded in a single link decision record. Prediction and measurement are kept apart throughout: a model result is never presented as a field measurement.

1 · Coverage and point-to-point site planning

Wireless coverage and point-to-point connectivity were modelled together, considering antenna alignment, capacity requirements and site coverage. Coverage describes where access is available; point-to-point planning describes the transport between sites. Antenna placement and alignment influence both the proposed link and its installation.

Coverage asks where a client can maintain the required radio service. Backhaul asks whether the path between fixed endpoints has adequate propagation and clearance. Capacity asks whether the radio and network can deliver the required traffic under the defined load. Treating these as separate review paths keeps a nominally covered area from being presented as guaranteed throughput.

A planning record needs the radio settings, antenna pattern, installation geometry and environmental assumptions associated with its prediction. Alignment should be reviewed against the intended endpoint and polarisation before optimising observed signal. No frequencies or sector counts are reconstructed from the low-resolution preview.

2 · Line of sight and Fresnel-zone clearance

A point-to-point link needs more than a visible sight line. The planning tool addresses line-of-sight and Fresnel-zone clearance: the link budget evaluates power, while the path profile evaluates geometry and obstruction.

The first Fresnel radius varies along the path. With wavelength λ and distances d1 and d2 from the inspected point to the endpoints, r1 = sqrt(λ d1 d2 / (d1 + d2)), using metres consistently. The maximum radius occurs at the midpoint for a level path, but the critical clearance point can lie elsewhere where terrain or an obstacle rises.

All lengths use metres and frequency uses Hz. A 60% first-zone clearance is an initial planning target, subject to the path, refraction and availability assumptions; it is not a guarantee of link performance. ITU-R P.530 considers climate, effective Earth radius, diffraction and availability requirements when specifying clearance. Avoid applying the curvature correction twice.

A complete review combines endpoint elevations and antenna heights with a terrain profile, obstacle heights and the selected refraction assumptions. Straight-line visibility alone can conceal Fresnel obstruction. Any recommended mast height must identify its survey inputs; the available material does not support a site-specific height recommendation.

3 · RF link budget and propagation analysis

The VB.NET engineering calculator addresses path loss, receiver sensitivity, thermal noise and link-budget calculations. A walkthrough must expose input units and assumptions as well as the output, so a reviewer can recompute a result without access to the original source code.

The interface takes frequency, distance, antenna heights, transmit power, antenna gains, cable and polarisation losses, system margin, terrain and the effective-Earth factor, and groups results into path loss, Fresnel clearance, link budget, receiver sensitivity and a received-power chart.

Gains are dBi; losses and margins are dB; absolute powers and sensitivity are dBm. Free-space path loss is only one part of the propagation model: obstacles, diffraction, atmospheric effects and interference can change a field link, so a positive calculated margin is not a recorded availability or throughput result. The receiver threshold depends on bandwidth, modulation and required quality, so datasheet sensitivity is kept distinct from a sensitivity calculated from noise and required SNR.

Recalculation of the displayed example

For 5800 MHz and 5.02 km, with 20 dBm transmit power, 15 dBi gain at each end and 1 dB cable loss at each end, assuming free-space propagation and no other losses:

The displayed received power agrees with the displayed loss, but that loss does not agree with the visible frequency and distance. This is a discrepancy in the poster; without application source it does not identify the cause or prove a software defect. The original image is preserved as source evidence. At 20 MHz bandwidth and 5 dB noise figure, equivalent receiver noise is approximately −95.99 dBm; the poster’s roughly −101 dBm corresponds to thermal noise before adding noise figure, so the two quantities are labelled separately.

Putting it together: one link, three checks

The same illustrative path is carried through all three stages to show how the checks combine into a decision. The values are independently computed example inputs, not recovered deployment parameters.

A passing link budget does not remove a clearance deficit, and a clear path does not prove margin. The decision record therefore lists all three results with their assumptions, then hands over to alignment and field validation, where measured values are recorded separately from these predictions.

Historical implementation; the RF calculations shown are a subsequent engineering review, not commissioning measurements. The original implementation and the later review are separate evidence categories.

Atma to Al-Sharq engineering case

Download the engineering report (PDF) · Open the calculation explorer · Read the engineering audit

Source requirements and engineering inputs

The original requirements specify 8.5 km and 16.5 km. The retained terrain profiles yield 8.277 km and 15.556 km, while UISP displays 8.31 km and 15.64 km. Keep all three distance bases labelled; confirm actual mounting points before adopting a final path length.

The historical budget lists one link on each hop: PowerBeam M5 400 ($180) and a 34 dBi RocketDish link ($1,200). A link requires both endpoints; the second hop also requires compatible Rocket radios. Exact procurement quantities must be reconciled.

View the original requirements document

Actual UISP simulation results

The two result images below were captured from Ubiquiti UISP Design Center on 12 October 2026. Guest mode, manual product selection, Noise Free, 20 MHz and 6 m antenna heights were used. These are actual tool outputs, not recreated interface illustrations.

UISP reports that LIDAR data is unavailable in this area. Its frequency and loss assumptions differ from the independent 5.8 GHz calculation. The displayed 84.5 Mbps is vendor-predicted capacity, not an iperf measurement or a guaranteed end-to-end service rate.

Terrain and Fresnel clearance

The nominal sampled paths clear 60% of the first Fresnel zone. The midpoint first-zone radii are about 10.34 m and 14.18 m at 5.8 GHz. Curvature is applied once, with k = 4/3. Ground elevations are in metres above mean sea level; antenna heights are above local ground.

These profiles do not resolve every tree, rooftop or narrow ridge. A direct survey of the endpoint approaches remains necessary. A zero-height result from a sparse model is not an installation height recommendation.

Explore the terrain views

RF budget and MIMO capacity

Each device uses its own manufacturer TX table. MCS0-7 denotes one stream; MCS8-15 denotes two streams. The highest two-stream example is MCS15, 130 Mbps PHY at 20 MHz/800 ns GI. At an assumed 50% payload efficiency, 65 Mbps is an illustrative per-hop estimate, shared across subscribers and directions.

The RF equations use dBm for absolute power, dBi for antenna gain and dB for losses. H1 has an integrated antenna and no separately added external RF-feed loss. H2 assumes 0.5 dB feed loss per end. Both include 2 dB other losses.

EIRP and uncertainty change the decision

Hardware maximum TX power is a capability input, not a permitted operating setting. The illustrative EIRP sweep shows how lower radiated power changes top-rate margin. It does not assert any legal EIRP limit for Yemen. Confirm the applicable band, channel, country profile and radiated-power limit.

For a 36 dBm illustrative EIRP cap, the MCS15 margins fall to 7.93 dB on H1 and 10.95 dB on H2. H1 then falls below the 10 dB screen at MCS15 and requires a lower modulation rate. Device TX and receiver sensitivity tolerances alone may consume another 4 dB.

Mast-height sensitivity and design boundaries

The Monte Carlo screen varies assumed DEM error, clutter up to 15 m and effective Earth factor in 5,000 scenarios per height. A scenario pass fraction is not uptime, a measured obstruction probability or a confidence level. Endpoint elevation error is fixed and terrain between samples is unresolved.

Increasing H2 mast height improves this sampled-point screen, but neither 10 m nor 15 m is a certified height. Surveyed clearance, safe access, foundation/mast structure, wind loading and permissions determine the final installation.

The legacy annual rain/multipath outage figures are withdrawn from current conclusions: local climate, propagation-model applicability and annual versus worst-month time basis were not established.

Seven-zone distribution and client coverage

The original map identifies seven retained zones. The equipment totals are 16 NanoStation M2 units, six local PowerBeam units and one sector line. A NanoStation has a directional antenna; a nominal 500 m reach is not omnidirectional coverage or a guaranteed high-rate boundary.

The 5 GHz sector and 2.4 GHz NanoStation operate in different bands and cannot directly associate over RF. Confirm the sector radio, its exact SKU and any Ethernet interconnection. Coverage below assumes on-axis 11 dBi antennas at both ends, an n = 3.2 path-loss exponent and 8 dB shadowing; none is fitted to local measured coverage.

Equipment reference imagery

These are authentic manufacturer product-family illustrations reproduced from official Ubiquiti datasheets. They identify the hardware families and are not photographs of the installed sites. Exact SKU identification still requires an inventory or device label.

PowerBeam PBE-M5-400: 25 dBi integrated antenna. RocketDish RD-5G34: 34 dBi antenna with approximately 3 degree beamwidth, requiring a compatible radio. NanoStation M2: directional 2.4 GHz MIMO device.

The existing VB.NET RF and Fresnel project posters remain available on the website. They are source document previews; the new calculation plots and UISP captures are separately labelled. No new VB.NET execution or field configuration screenshot is claimed.

Network operation and relay-site power

Separate management, backhaul and subscriber traffic. Record the actual router configuration and Internet-line health checks. For co-located MIMO radios, keep both polarisation chains matched; determine channel and physical separation through spectrum and desensitisation measurements rather than a fixed universal spacing.

The remote relay load is assumed as two 8 W radio allowances plus 10% conversion/cable overhead: 17.6 W, or 422.4 Wh/day. Actual injector/DC conversion and auxiliary consumption must be measured.

A system near 200 W is the mathematical scenario, not a verified winter design. Check minimum-month insolation, recharge after prolonged low sun, temperature derating, battery condition and all loads. Grounding, surge protection, weatherproofing and structural work require site inspection.

Historical source subtotal: backhaul $1,380 + zone equipment $2,090 + two core routers $420 = $3,890. Cable, modems, accessories, uncertain currency lines, missing radio quantities and solar items prevent a complete present-day project total.

Engineering audit and remaining evidence

The review corrects the following issues while preserving the project implemented status and the original documentation. The full audit in the repository records completion evidence for each remaining item.

The main remaining evidence is site photographs, surveyed antenna positions/heights, device inventory, actual settings, per-chain signal/noise, channel scans, traffic tests and power/uptime logs. Product-family images and planning screenshots cannot substitute for those records.

Field acceptance record and references

Use this evidence register during verification of the installed system. Agree service throughput, latency, packet loss and availability targets with the operator before applying pass/fail decisions. A seven-day log is an initial stability screen, not annual availability proof.

Reproducible calculation, input tables, results and engineering audit

Engineering approach

• Separate coverage, path clearance, link budget and capacity, then combine them in one decision record.
• Expose input units and assumptions with every calculated result so it can be recomputed independently.
• Treat standards such as ITU-R P.530 as design context that does not replace a radio data sheet or site survey.
• Keep predictions, calculator outputs and field measurements clearly distinguished.

The two VB.NET calculators support analysis and are distinct from the Ubiquiti wireless network work, which is documented through the source requirements and planning deliverables.

Verification context

The functional diagrams explain the engineering responsibilities of each part. The original source previews appear above. As-built addressing, configuration exports and raw acceptance records are not included in the public release. The current report adds actual UISP planning captures and computed plots. Surveyed heights, installation photographs and measured traffic records remain to be supplied.

Detailed verification boundaries and technical source context

Source documentation

• PowerBeam M5 400
• Rocket M5
• NanoStation M2
• RocketDish
• UISP Design Center
• Elevation dataset
• Free-space propagation
• Terrestrial LOS design
• Map attribution

• Wireless coverage and network planning — rf-mpls-network-design
• Fresnel and link-budget calculators — dotnet-monitoring-apps
• ITU-R P.530 — terrestrial line-of-sight link design
• ITU-R P.525 — free-space attenuation
• LinkedIn project record