Wireless Site Survey and AP Placement: An RF Planning Reference

The short answer

A wireless site survey plans access point (AP) placement, channels, and transmit power so that every area where devices are used receives adequate signal and a clean signal-to-noise ratio. The two numbers that anchor the design are RSSI (received signal strength) and SNR (signal-to-noise ratio). Widely used enterprise targets are a cell-edge RSSI of -67 dBm for voice and real-time traffic, roughly -70 dBm for reliable general data, and an SNR of at least 25 dB. SNR is RSSI minus the noise floor: with a typical indoor noise floor near -92 dBm, a -67 dBm signal yields about 25 dB of SNR. RSSI alone does not guarantee performance; a strong signal sitting in a noisy channel still delivers poor throughput, which is why both figures are validated together.

Channel planning is the other half. The 2.4 GHz band offers only three non-overlapping 20 MHz channels in North America (1, 6, and 11), so it is reused carefully and reserved for range and legacy devices. The 5 GHz band provides up to 25 twenty-MHz channels in the US across the UNII-1, UNII-2, UNII-2C, and UNII-3 sub-bands, including DFS channels that share spectrum with radar and must vacate on detection. Wider 40, 80, and 160 MHz bonded channels raise peak throughput but shrink the count of non-overlapping channels, increasing co-channel interference. Adjacent AP cells are designed to overlap roughly 15-20 percent so clients roam before they drop. AP count follows client density and airtime demand, not floor area alone. Surveys are predictive (modeled), passive (measured coverage), or active (measured while associated).

How to Approach an RF Design Decision

Start by defining the application, because it sets the RSSI and SNR targets that everything else follows. Coverage-only networks (asset tags, guest browsing, warehouse scanners tolerant of retries) can be designed to a cell-edge RSSI near -70 dBm. Data networks that carry business applications are commonly designed to -67 dBm. Voice over Wi-Fi and real-time video hold the tightest budget: -67 dBm at the cell edge with SNR of at least 25 dB, and cell overlap large enough that a handset re-associates to the next AP before the current signal degrades. High-density spaces such as lecture halls, arenas, and conference centers are driven by client count and airtime, not coverage, and often need more, lower-powered APs rather than fewer strong ones.

Next, decide the band strategy. Treat 5 GHz as the primary capacity band and 2.4 GHz as a fallback for reach and older devices, because 2.4 GHz has only three non-overlapping channels and far more non-Wi-Fi interference. Choose channel width deliberately: 20 or 40 MHz in dense multi-AP deployments preserves the number of reusable channels, while 80 or 160 MHz suits low-AP-count sites like a single-tenant office or a point-to-point link. Finally, match the survey method to the project phase. Predictive modeling scopes a design from floor plans before cabling; a passive or active survey after installation validates that the deployed network actually meets the RSSI, SNR, overlap, and data-rate targets on the floor.

• Voice / real-time: cell-edge RSSI -67 dBm, SNR >= 25 dB, generous overlap for fast roaming.
• Reliable data: cell-edge RSSI around -70 dBm; -67 dBm where business apps are latency-sensitive.
• Coverage-only / scanners: -70 dBm is often acceptable where retransmission is tolerable.
• High-density venues: size by concurrent clients and airtime, then lower transmit power and add APs.
• Band strategy: 5 GHz for capacity, 2.4 GHz as reach/legacy fallback with channels 1, 6, 11 only.
• Survey phase: predictive to design, passive/active to validate the built network against targets.

RSSI, SNR, and the Noise Floor

RSSI is the received power of the Wi-Fi signal, expressed in dBm on a negative logarithmic scale where numbers closer to zero are stronger. A reading of -50 dBm is excellent, -67 dBm is the common voice cell-edge threshold, -70 dBm supports reliable data, and signals weaker than about -80 dBm are typically unusable for modern applications. Because the scale is logarithmic, every 3 dB change represents a doubling or halving of power, and roughly every 6 dB of additional path loss halves the usable distance in free space. Free-space path loss rises about 20 dB per decade of distance, so signal falls off quickly and interior walls, glass, and metal add further attenuation that a survey must measure rather than assume.

SNR is the difference between the signal and the noise floor, in dB. If the signal is -67 dBm and the noise floor is -92 dBm, SNR is 25 dB. SNR, not raw RSSI, governs which modulation and coding a client can sustain: higher-order modulation such as 256-QAM (used by 802.11ac/Wi-Fi 5) or 1024-QAM (802.11ax/Wi-Fi 6) requires a high SNR to decode, so a link with strong RSSI but a raised noise floor will down-shift to slower, more robust rates. Enterprise voice designs target 25 dB SNR; data can function at lower SNR but at reduced data rates. Lowering AP transmit power to control cell size does not help if it drops SNR below what the target data rate needs.

• RSSI scale: ~-50 dBm excellent, -67 dBm voice edge, -70 dBm reliable data, below ~-80 dBm unusable.
• 3 dB = double/half power; ~6 dB extra path loss roughly halves usable free-space distance.
• SNR (dB) = RSSI (dBm) minus noise floor (dBm); design target >= 25 dB for voice.
• Typical indoor noise floor near -92 dBm; -67 dBm signal over it yields ~25 dB SNR.
• High-order modulation (256-QAM, 1024-QAM) needs high SNR; noisy channels force slower rates.
• Validate both RSSI and SNR on the floor; strong signal in a noisy channel still underperforms.

Channel Planning: 2.4 GHz, 5 GHz, DFS, and Bonding

The 2.4 GHz band spans roughly 2.400-2.483 GHz. Although channels 1 through 11 are usable in North America (up to 13 in much of the rest of the world under ETSI), each 20 MHz channel is wide enough that only channels 1, 6, and 11 do not overlap. Every viable 2.4 GHz plan reuses just those three, so the band is congested and better reserved for coverage and legacy clients. 40 MHz bonding in 2.4 GHz is discouraged because it leaves no room for a clean reuse pattern.

The 5 GHz band offers far more room: up to 25 non-overlapping 20 MHz channels in the US across UNII-1 (36, 40, 44, 48), UNII-2A (52, 56, 60, 64, DFS), UNII-2C (100-144, DFS), and UNII-3 (149, 153, 157, 161, 165). DFS (Dynamic Frequency Selection) channels share spectrum with radar; an AP must monitor for radar and, on detection, vacate the channel for a period, which can briefly interrupt clients, so DFS channels add capacity but carry that trade-off. Channel bonding combines adjacent 20 MHz channels into 40, 80, or 160 MHz to raise peak throughput, but each doubling halves the number of non-overlapping channels available for reuse. An 80 MHz plan leaves only a couple of fully non-overlapping non-DFS options, which is why dense multi-AP sites usually stay at 20 or 40 MHz.

• 2.4 GHz (~2.400-2.483 GHz): only channels 1, 6, 11 are non-overlapping in North America.
• 5 GHz US: up to 25 non-overlapping 20 MHz channels across UNII-1, UNII-2A, UNII-2C, UNII-3.
• Non-DFS 5 GHz 20 MHz channels: 36, 40, 44, 48 (UNII-1) and 149, 153, 157, 161, 165 (UNII-3).
• DFS channels (52-64 and 100-144) require radar detection and channel evacuation on a hit.
• Bonding to 40/80/160 MHz raises peak rate but halves usable channels at each step.
• Dense deployments favor 20 or 40 MHz; 80/160 MHz suits low-AP-count offices and PtP links.

Co-Channel and Adjacent-Channel Interference, and Cell Overlap

Two failure modes shape AP layout. Co-channel interference (CCI) occurs when nearby APs and their clients share the same channel: Wi-Fi is a polite, contention-based medium, so co-channel radios take turns and split airtime rather than colliding outright, which caps aggregate throughput as more devices contend. Adjacent-channel interference (ACI) is worse: it occurs when APs use partially overlapping channels (for example channels 1 and 3 in 2.4 GHz), so their transmissions bleed into each other as raw noise the radios cannot decode or defer to, raising the noise floor and cutting SNR. A correct plan eliminates ACI entirely by using only non-overlapping channels, then manages the unavoidable CCI through channel reuse spacing and transmit-power control.

Cell overlap is designed, not accidental. Adjacent AP coverage cells are planned to overlap roughly 15-20 percent for general data so a moving client finds a viable next AP before it loses the current one. Voice and real-time roaming use larger overlap, and the design ensures the client reaches the roaming handoff threshold while both cells still meet the RSSI and SNR targets. Too little overlap creates coverage holes and sticky-client drops at the edge; too much overlap raises co-channel contention and wastes airtime. Transmit power is tuned so cell sizes match the reuse plan, and 2.4 GHz radios are frequently run at lower power (or some disabled) because their cells naturally reach farther than 5 GHz.

• Co-channel interference: same-channel radios share airtime politely; throughput drops as contention rises.
• Adjacent-channel interference: partially overlapping channels inject undecodable noise, lowering SNR.
• A correct plan uses only non-overlapping channels, eliminating ACI by design.
• Design overlap ~15-20% between adjacent cells for data roaming; larger for voice.
• Too little overlap = coverage holes and sticky clients; too much = wasted airtime and CCI.
• Tune transmit power to fit cell size to the reuse plan; often reduce or disable extra 2.4 GHz radios.

AP Density and Survey Methods

AP count is driven by capacity and airtime, not floor area alone. A coverage-first design may place APs by signal reach, but a capacity design starts from the number of concurrent clients and their traffic. Because a channel is a shared medium, every associated device competes for airtime; a single AP serving many active clients divides that airtime, so dense spaces need more APs on more channels to spread the load, each running lower power to keep cells small. Client capabilities matter too: a legacy device that transmits at a low data rate consumes disproportionate airtime, dragging down the cell for everyone (the airtime-fairness problem). Planners also account for band steering, so dual-band clients prefer the roomier 5 GHz band and leave 2.4 GHz for devices that need it.

Three survey methods correspond to project phases. A predictive survey models coverage from floor plans, wall materials, and AP specifications in software before any hardware is installed; it scopes AP quantity and cabling and is only as accurate as its inputs. A passive survey walks the built space with a receiver that measures RSSI, SNR, channel usage, and interference from all nearby APs without associating, mapping real coverage. An active survey associates a client to the network and measures actual throughput, retransmission, packet loss, and roaming behavior, revealing performance a passive scan cannot. Rigorous projects combine them: predictive to design, then passive and active to validate against the RSSI, SNR, overlap, and data-rate targets.

• Size AP count by concurrent clients and airtime demand, not square footage alone.
• Shared-medium airtime: many active clients on one AP split capacity; add APs on more channels.
• Legacy low-rate clients consume outsized airtime; airtime fairness and band steering help mitigate.
• Predictive survey: software model from floor plans; scopes design pre-installation, input-dependent.
• Passive survey: measures RSSI/SNR/interference from all APs without associating; maps real coverage.
• Active survey: associates and measures throughput, loss, and roaming; validates real performance.

Frequently asked questions