Microwave Backhaul Link Commissioning Project

This project commissions a point-to-point microwave backhaul link between two telecommunications sites. The goal is to produce a defensible commissioning package: design assumptions, link-budget calculation, antenna alignment method, spectrum evidence, service tests, acceptance thresholds, failure-mode review, and final handover record.

The project is not a generic wireless overview. It asks whether a real installed link can carry a backhaul service with enough margin, timing performance, resilience, and evidence for operations teams to support it after commissioning.

Project Objective

Commission a microwave backhaul link that connects a remote access site to an aggregation site. The final engineering deliverable should answer:

1. Does the installed RF path meet the calculated received-power and SNR margin?
2. Are antenna alignment, polarization, feeder loss, and site losses consistent with the design basis?
3. Does local spectrum evidence support the selected channel?
4. Does the service meet throughput, latency, jitter, packet loss, and availability requirements?
5. Which alarms, monitoring thresholds, and records should operations inherit?
6. What degraded modes or failure modes remain after commissioning?

The deliverable is a commissioning report and acceptance checklist. It should be detailed enough for a field engineer, network engineer, or design reviewer to reproduce the decision.

Baseline Scenario

Use the following scenario or replace it with site data.

Parameter	Value
Link purpose	remote site backhaul
Frequency	18\ \text{GHz}
Path length	8.0\ \text{km}
Channel bandwidth	56\ \text{MHz}
Required net throughput	150\ \text{Mbit/s}
One-way latency target	less than 3.0\ \text{ms}
Peak-to-peak jitter target	less than 0.8\ \text{ms}
Packet loss target	less than 0.1\% under acceptance load
Required fade and implementation margin	at least 20\ \text{dB}
Transmitter power	20\ \text{dBm}
Antenna gain at each end	38\ \text{dBi}
Feeder and connector loss at each end	2.0\ \text{dB}
Miscellaneous path allowance	3.0\ \text{dB}
Receiver noise figure	5.0\ \text{dB}
Required detector SNR	18\ \text{dB}

These values are simplified. A real commissioning package should use licensed frequency data, final radio configuration, measured feeder loss, antenna model, polarization, tower geometry, path profile, rain zone, equipment firmware, traffic class policy, and calibrated measurement records.

Step 1: Define the Service Boundary

The service boundary should state what is being accepted. For this project, accept the path from the Ethernet handoff at Site A to the Ethernet handoff at Site B, including radios, antennas, feeders, grounding, power, management access, and monitoring.

Do not accept only the RF carrier. A backhaul link can have good received power and still fail because queueing, duplex mismatch, traffic shaping, packet loss, timing variation, poor alarms, or weak power autonomy degrade the service.

The commissioning record should include:

• site names and coordinates;
• radio model, serial numbers, firmware version, and licensed channel;
• antenna model, diameter or aperture class, gain, polarization, azimuth, and elevation;
• feeder type, connector count, measured loss, grounding, and weatherproofing;
• expected received signal level and acceptance tolerance;
• throughput, latency, jitter, packet loss, and monitoring thresholds;
• photographs or screenshots that document final field condition.

Step 2: Build the Link Budget

Free-space path loss using distance in kilometers and frequency in MHz is:

For the baseline:

Therefore:

Received power is:

Substitute the baseline values:

This is the predicted received signal level at the receiver input after the stated losses.

Engineering Comment

The result is credible for a short high-gain 18 GHz link, but it is only useful if the site losses and antenna gains are real. A commissioning report should compare this predicted value with the measured received level after final antenna alignment. A difference of several dB should trigger investigation before acceptance.

Step 3: Estimate Receiver Noise and SNR Margin

Receiver noise power for a bandwidth B is:

For a 56\ \text{MHz} channel:

With NF=5.0\ \text{dB}:

Available detector SNR estimate:

Margin above the required detector SNR:

The baseline SNR margin exceeds the required 20\ \text{dB} allowance by:

Engineering Comment

The calculation suggests enough margin before local interference, rain attenuation, antenna misalignment, equipment aging, and implementation uncertainty consume it. Because the link operates at 18\ \text{GHz}, rain fade and alignment should be treated seriously. A strong commissioning package separates fade margin, interference margin, and installation tolerance instead of hiding them inside one unexplained number.

Step 4: Define Antenna Alignment Procedure

Antenna alignment should be a controlled test, not a visual aim. Use safe access procedures and follow the radio vendor’s alignment mode if available.

Recommended alignment sequence:

1. Confirm both radios are configured for the intended frequency, channel width, polarization, and transmit power.
2. Verify antenna mounting, mechanical tilt, polarization orientation, connector weatherproofing, and grounding before transmitting.
3. Set both antennas to calculated azimuth and elevation.
4. Sweep slowly in azimuth while monitoring received signal level.
5. Peak the main lobe and reject side-lobe false peaks by checking expected beamwidth and signal level.
6. Sweep elevation and recheck azimuth.
7. Tighten hardware to specified torque and confirm the signal level after tightening.
8. Record final received level at both ends, alignment screenshots, weather, and technician names.

For this project, predicted received power is -46.6\ \text{dBm}. A practical acceptance tolerance might be:

Measured level	Decision
-49.6\ \text{dBm} or stronger	accept RF level if other tests pass
between -49.6 and -52.6\ \text{dBm}	investigate alignment and losses
weaker than -52.6\ \text{dBm}	do not accept without root cause

Engineering Comment

The tolerance should reflect measurement uncertainty, weather, radio calibration, and expected loss uncertainty. A link that is 6\ \text{dB} weaker than predicted may still pass a quick throughput test, but it has already consumed margin needed for rain, aging, and interference.

Step 5: Collect Spectrum and Interference Evidence

Commissioning should include evidence that the selected channel is usable in the field. A receiver-sensitivity calculation does not reveal local emitters, reflections, intermodulation, or intermittent interference.

Minimum evidence:

• spectrum scan near each antenna direction where safe and practical;
• radio built-in interference or noise-floor report;
• occupied channel, adjacent-channel occupancy, and guard-band observation;
• scan settings: resolution bandwidth, detector, averaging, span, date, time, and antenna setup;
• note on weather, nearby transmitters, temporary equipment, and visible obstructions.

Acceptance example:

Evidence item	Pass condition
in-channel interference	no persistent interferer within operating channel
adjacent channel	no adjacent energy high enough to trigger receiver desensitization
noise floor	within 3\ \text{dB} of expected receiver or site baseline
intermittent events	none observed during minimum observation window
screenshot metadata	settings and timestamp recorded

Engineering Comment

Spectrum evidence is not permanent truth. It is commissioning evidence. If a new transmitter, crane, building, temporary generator, or antenna is later added, the spectrum basis may change and should be rechecked.

Step 6: Validate Service Performance

Run service tests after RF alignment and spectrum review. The service test should exercise both directions and the intended traffic class.

For this project, use the following acceptance plan:

Test	Method	Acceptance
Throughput	bidirectional traffic test for 15\ \text{min}	at least 150\ \text{Mbit/s} net service throughput
Latency	timestamped packet probe under acceptance load	95th percentile one-way latency less than 3.0\ \text{ms}
Jitter	packet-delay variation from the same run	peak-to-peak jitter less than 0.8\ \text{ms}
Packet loss	counted sent and received packets	less than 0.1\%
Modulation stability	radio diagnostics during load	no repeated fallback below accepted operating mode
Alarm visibility	disconnect, low-RSL threshold, power alarm if safe	alarms visible in monitoring system

Example measurement:

Quantity	Measured value
Net throughput	168\ \text{Mbit/s}
95th percentile one-way latency	2.2\ \text{ms}
Peak-to-peak jitter	0.55\ \text{ms}
Packet loss	0.03\%
Minimum received signal during test	-47.8\ \text{dBm}

All service metrics pass the stated baseline acceptance thresholds.

Engineering Comment

Passing throughput alone is not enough. A backhaul link may deliver peak rate while producing poor jitter, packet loss under small packets, or unstable modulation. Record the traffic profile, packet size, duration, direction, and load condition so the result can be interpreted later.

Step 7: Review Failure Modes

A commissioning package should identify what can still fail after acceptance.

Failure mode	Likely symptom	Commissioning control
antenna misalignment after wind loading	reduced received signal level	final torque record and RSL baseline
water ingress in connector	slow margin degradation	weatherproofing photos and inspection interval
channel interference	packet loss or modulation fallback	spectrum evidence and monitoring threshold
power interruption	link outage or radio reboot	power autonomy test and alarm verification
queue saturation	high latency and jitter	load test and QoS configuration record
firmware or configuration drift	unexpected mode, bandwidth, or alarms	configuration export and version record
shared tower or route dependency	correlated outage	dependency map and site-risk note

The project should not claim that these risks are eliminated. It should show which risks are controlled, which are monitored, and which require operational procedures.

Step 8: Define Monitoring and Handover

Operations should inherit useful thresholds, not only a statement that the link passed. Recommended monitoring items:

• received signal level at both ends;
• modulation and coding mode;
• error counters and retransmission indicators;
• link availability and flap count;
• throughput, queue drops, latency, and jitter probes;
• radio temperature and power supply state;
• configuration backup and firmware version;
• alarm thresholds tied to acceptance margin.

If predicted received level is -46.6\ \text{dBm} and accepted final level is -47.8\ \text{dBm}, a reasonable early-warning threshold might be set near -52\ \text{dBm}, with a higher-severity alarm near the minimum level required for the accepted modulation. The exact values should come from vendor thresholds, fade plan, and operations policy.

Engineering Comment

Monitoring thresholds should be related to measured commissioning baselines. A generic default alarm may be too loose to catch margin degradation early or too tight for a link with expected weather fade.

Final Deliverable

The project deliverable is a microwave backhaul commissioning package containing:

1. service boundary and acceptance requirements;
2. final equipment, firmware, frequency, bandwidth, antenna, and site data;
3. link-budget calculation with assumptions and margin;
4. measured received levels before and after final alignment;
5. spectrum evidence and interference notes;
6. throughput, latency, jitter, packet-loss, and modulation-stability results;
7. alarm and monitoring verification;
8. failure-mode review and residual-risk notes;
9. configuration exports, screenshots, photographs, and signoff record.

Acceptance should be explicit:

Area	Decision
RF level and SNR margin	pass if measured level is within tolerance and margin remains above requirement
spectrum evidence	pass if no credible in-channel or adjacent-channel interference is observed
service performance	pass if throughput, latency, jitter, and loss meet thresholds
operations handover	pass if alarms, baselines, records, and ownership are complete

If any area fails, the report should state the corrective action: realign, repair feeder loss, change channel, reduce service target, adjust QoS, improve power support, retest under load, or redesign the path.

Common Project Mistakes

Common mistakes include accepting the link from throughput alone, using transmitter power instead of EIRP, ignoring feeder and connector losses, peaking on an antenna side lobe, failing to record spectrum settings, accepting a link during unusually good weather without margin review, and omitting monitoring thresholds from handover.

Another frequent mistake is leaving commissioning evidence in field screenshots that operations cannot find later. A backhaul link is not commissioned until the evidence is attached to the service record that future engineers will actually use.