Wireless access point deployment runs from site survey through design, physical installation, and validation, and its success depends on capacity, roaming, and interference rather than coverage alone. Deployments typically fail when coverage testing passes but users still see dropped calls, slow transfers in rooms with full signal bars, and applications that stall when the space fills.
Signal strength is the easiest thing to measure and the least predictive of experience. A deployment designed around coverage alone will meet its stated requirement and still be considered a failure by the people using it.
Why Design Wi-Fi for Capacity Instead of Coverage?
Because a network every client can hear may still be unable to serve them all at once. Coverage answers whether a client can hear an access point. Capacity answers whether the access point can serve every client that will associate with it.
These produce different designs. A warehouse with a handful of scanners needs few access points placed for reach. A conference center with the same square footage and several hundred simultaneous devices needs many more access points at lower transmit power, placed so that each serves a smaller cell.
Start from device count and application profile rather than square footage. How many devices per area at peak, what they are doing, and what each of those activities requires in throughput and latency. Voice and video are latency-sensitive and modest in bandwidth. Bulk transfer is the reverse. A design that satisfies one may not satisfy the other.
Density changes the tuning. Higher access point counts mean smaller cells, which means reducing transmit power so that cells do not overlap destructively — a counterintuitive step that installers frequently skip, leaving high-density deployments with every radio at full power and clients unable to make good roaming decisions.
What Are the Three Types of Wireless Site Survey?
Predictive, active, and post-installation validation, and each answers a different question.
Predictive survey models the environment in software from floor plans, with wall materials and attenuation values applied. It produces a design to build against and a bill of materials. It is only as good as the accuracy of the plans and the material assumptions, and it is routinely optimistic about interior walls.
Active survey measures the real environment with equipment on site. Pre-deployment, it validates the predictive model using a test access point in candidate positions. This is where the model's wrong assumptions surface — the wall that turned out to be foil-backed insulation, the elevator shaft that blocks more than expected.
Post-installation validation measures the finished deployment as it will be used. Signal, signal-to-noise ratio, throughput, roaming behavior, and interference, walked through the actual space at client height.
Skipping the middle step is the common shortcut. It converts a design error into an installation error, which is discovered after the mounting hardware is in the ceiling.
How Do You Plan Channels Around Interference?
Survey for every source of interference first, because the spectrum is shared and everything in 802.11 bands competes.
Non-Wi-Fi sources matter more than most designs account for: microwave ovens, some cordless phones, wireless cameras, medical telemetry, and in industrial settings a range of equipment nobody thinks of as a radio. A spectrum analyzer during survey finds these. A Wi-Fi-only survey tool does not.
Neighboring networks are the other half. In multi-tenant buildings, the adjacent floors are part of your radio environment whether or not you have any control over them.
Channel planning follows from this. The 2.4 GHz band has three non-overlapping channels and should be treated as a legacy-support band rather than a capacity band. The 5 GHz band has enough channels for real planning, with attention to the ones carrying radar-avoidance requirements — clients handle those inconsistently, and an access point that vacates a channel on a radar event takes its clients with it. The 6 GHz band offers substantial clean spectrum where client support exists.
Mounting and the Physical Install
Access points belong in the ceiling, exposed, oriented as the manufacturer specifies. The antenna pattern is designed around that orientation, and installations that deviate produce coverage that does not match any model.
The recurring problem is aesthetics. Access points get moved above ceiling tiles, into corners, behind obstructions, or relocated during construction to positions that look tidier — and the deployment quietly stops matching its design. Where concealment is required, plan for it in the design phase with hardware intended for it, rather than accepting a field change.
Cabling runs to each position, and each position needs power over ethernet at the correct IEEE 802.3 standard. Current-generation access points draw enough that older switches cannot power a full complement at once. Verify the switch power budget against the full deployment, not against a single port.
Validation and Handover
Test what users will do. Walk the space with a client device, measure signal and noise, run throughput tests, and roam between cells while a call is active. Roaming is where deployments most often fall short, and it is invisible to a stationary test.
Test under load where the environment will be loaded. An empty conference hall and a full one are different radio environments — human bodies absorb 2.4 and 5 GHz signal measurably.
Document the final state: access point positions, names, channel and power settings, the validation results, and the cable runs feeding each. When a user reports a problem in a specific room six months from now, that document is the difference between an answer and an investigation.