- Purple
- WiFi RF engineering and troubleshooting: a complete guide
- Understanding RSSI and Signal Strength for Optimal Channel Planning
Understanding RSSI and Signal Strength for Optimal Channel Planning
You will be able to read RSSI correctly, set -67 dBm and 25 dB SNR design targets, and decide AP placement, power, channel width and data rates from survey evidence. You will also know where RSSI misleads, and how to separate a coverage problem from co-channel interference before buying more access points.
📖 13 min read2,967 words3 worked examples3 practice questions12 key definitions
Video overview
Listen to this guide
View podcast transcript
Understanding RSSI and Signal Strength for Optimal Channel Planning
A Purple WiFi Intelligence Briefing
[INTRODUCTION & CONTEXT - approximately 1 minute]
Welcome to the Purple WiFi Intelligence Briefing. I'm your host, and today we're getting into the fundamentals that underpin every high-performing wireless network: RSSI, signal strength, and how they drive optimal channel planning.
If you're an IT manager, network architect, or venue operations director, you've almost certainly encountered the frustration of a WiFi network that looks fine on paper but performs poorly in practice. Guests complaining about dropped connections. Handheld scanners losing signal mid-transaction. Video calls breaking up in the boardroom. The root cause, more often than not, comes back to a misunderstanding of what RSSI actually tells you - and more importantly, what it doesn't.
In the next ten minutes, I want to give you a clear, practical framework for understanding these metrics and translating them into better channel planning decisions. This isn't academic theory. This is the kind of briefing I'd give a client before a major deployment.
Let's get into it.
[TECHNICAL DEEP-DIVE - approximately 5 minutes]
So, what is RSSI? RSSI stands for Received Signal Strength Indicator. It's a relative measurement of the power level of a radio frequency signal as received by a client device. It's expressed in negative decibels relative to a milliwatt - so negative dBm. The closer to zero, the stronger the signal. Minus 30 dBm is excellent. Minus 90 dBm is effectively unusable.
But here's the critical point that many deployments get wrong: RSSI alone does not tell you whether a connection is good. It tells you how loud the signal is. It does not tell you how clear it is.
That's where Signal-to-Noise Ratio - SNR - comes in. SNR is the difference in decibels between your received signal and the ambient noise floor. If your RSSI is minus 65 dBm and your noise floor is minus 90 dBm, your SNR is 25 dB. That's the minimum you need for the high-order modulation schemes - things like 256-QAM - that deliver real throughput in 802.11ac and 802.11ax networks.
Think of it this way. Imagine you're in a quiet library. Someone whispers to you from across the room. You can hear them clearly - that's a good SNR. Now imagine you're at a stadium during a match. Someone shouts at you from the same distance. The signal is louder, but the noise is also much higher. You might struggle to understand them. That's exactly what happens in a noisy RF environment.
Now, why does this matter for channel planning?
WiFi is a shared medium. Every device on the same channel has to take turns transmitting, governed by a protocol called CSMA/CA - Carrier-Sense Multiple Access with Collision Avoidance. Before transmitting, every device listens to check if the channel is clear. If it hears another device, it backs off and waits.Co-Channel Interference - CCI - occurs when multiple access points on the same channel can hear each other. They all back off. They all wait. Channel utilisation goes through the roof, and latency spikes, even when actual client traffic is low. This is one of the most common performance killers in enterprise deployments, and it's entirely avoidable with proper channel planning.
Adjacent Channel Interference - ACI - is a different problem. In the 2.4 GHz band, channels are only 5 MHz apart but each channel is 22 MHz wide. So they overlap. If you put an AP on channel 3 next to an AP on channel 1, the RF energy from channel 3 bleeds into channel 1, raising the noise floor and degrading SNR. The solution in 2.4 GHz is to use only channels 1, 6, and 11 - the three non-overlapping channels.
In the 5 GHz band, you have far more spectrum to work with. You can use DFS channels - Dynamic Frequency Selection - to expand your available channel set, though you need to be aware that radar detection can force a channel change, which causes a brief disruption.
Now, a word on channel widths. There's a temptation to use wider channels - 40, 80, or even 160 MHz - because they offer higher theoretical throughput. And in a low-density environment, that's fine. But in a high-density venue - a hotel, a stadium, a conference centre - wider channels mean fewer non-overlapping options, which means more CCI. In those environments, 20 MHz channels in 2.4 GHz and 20 or 40 MHz in 5 GHz is almost always the right call.
Let me talk about AP placement and power tuning, because this is where I see the most mistakes in the field.
There's a common misconception that more transmit power equals better coverage equals better performance. It's wrong. Setting AP transmit power too high creates what we call an asymmetric link. The AP can shout loudly, and the client can hear it clearly from a long distance. But the client - a smartphone, a laptop, a handheld scanner - has a much weaker transmitter. It can't shout back with the same power. So the AP can't hear the client reliably.
This also creates the "sticky client" problem. A device in a far corner of the building can still hear the AP at minus 70 or minus 75 dBm. It decides the connection is acceptable and stays put, even when it moves physically closer to a different AP. The client doesn't roam. Performance degrades. The fix is to tune AP transmit power down - typically to 10 to 14 dBm - to match the client's capabilities, and to ensure sufficient AP density so that clients are always close to an AP.
To facilitate seamless roaming, you should implement the 802.11k, 802.11v, and 802.11r protocols. 802.11k provides clients with a neighbour report - a list of nearby APs they could roam to. 802.11v allows the network to suggest that a client roam to a better AP. And 802.11r enables fast BSS transition, dramatically reducing the time it takes to re-authenticate when roaming. Together, these protocols ensure that roaming decisions are driven by RSSI thresholds rather than client inertia.
[IMPLEMENTATION RECOMMENDATIONS & PITFALLS - approximately 2 minutes]
Right. Let's talk about implementation. Here are the key steps I'd walk through with any client.
First, define your requirements before you touch any hardware. What is the minimum RSSI you need to support your most demanding application? For voice over WiFi, you need minus 65 dBm or better. For high-throughput data, minus 70 dBm. For basic connectivity, minus 75 dBm. And critically, identify your Least Capable, Most Important device - the device with the weakest radio that absolutely must work reliably. Design for that device.
Second, conduct a proper site survey. Not just a predictive survey using software, but an active survey with real hardware in the real environment. Measure RSSI and SNR. Use a spectrum analyser to identify non-WiFi interference sources - microwave ovens, Bluetooth devices, DECT phones, even some industrial equipment. These raise the noise floor and degrade SNR without showing up on a standard WiFi scan.
Third, plan your channels before you deploy. In 2.4 GHz, stick to 1, 6, and 11. In 5 GHz, create a channel reuse plan that maximises the physical separation between APs on the same channel. Use 20 MHz channels in dense environments.
Fourth, tune your transmit power down. Match it to your client devices. Ensure 15 to 20 per cent cell overlap to support seamless roaming.
Fifth, set minimum mandatory data rates. Disable the legacy rates - 1, 2, 5.5, and 11 Mbps in 2.4 GHz. This forces clients to roam sooner when RSSI degrades, rather than clinging to a distant AP at a low data rate.
Now, the pitfalls. The most common one I see is over-reliance on automatic channel assignment. Most enterprise AP vendors offer automatic radio resource management - it sounds great in theory. In practice, in complex environments, it can make poor decisions. Always validate the channel plan manually after deployment.
The second pitfall is ignoring the noise floor. A network can look fine on an RSSI heatmap but perform terribly because the noise floor is elevated. Always measure SNR, not just RSSI.
The third pitfall is deploying a guest WiFi solution without thinking about the RF implications. Captive portals, analytics platforms, and location services all depend on a well-architected RF environment. If the RF is broken, the analytics will be inaccurate and the guest experience will be poor.
[RAPID-FIRE Q&A - approximately 1 minute]
Let me run through a few quick questions I hear regularly.
What RSSI do I need for a reliable connection? Minus 65 dBm or better for primary coverage. Minus 70 dBm for roaming overlap zones.
Should I use 80 MHz channels in a stadium? Almost never. The reduction in available non-overlapping channels causes CCI that far outweighs the throughput benefit.
My site survey shows good RSSI but performance is still poor. What's wrong? Check your SNR. Check your channel utilisation. Check for sticky clients. One of these three is almost certainly the culprit.
Is 2.4 GHz still worth deploying? Yes, for legacy device compatibility and penetration through walls. But limit it to channels 1, 6, and 11, and consider disabling it on every other AP in dense environments to reduce CCI.
[SUMMARY & NEXT STEPS - approximately 1 minute]
Let me wrap up with the key takeaways.
RSSI tells you signal strength. SNR tells you signal quality. Always optimise for SNR, not just RSSI.
Design for capacity, not coverage. More APs at lower power beats fewer APs at high power in any dense environment.
Use non-overlapping channels. In 2.4 GHz, that's channels 1, 6, and 11. In 5 GHz, build a proper channel reuse plan.
Implement 802.11k, v, and r to ensure roaming is driven by RF conditions, not client stubbornness.
Validate with a real active site survey. Software predictions are a starting point, not a final answer.
And finally, remember that your RF architecture is the foundation for everything else - your guest WiFi experience, your analytics, your location services, your operational efficiency. Get the RF right, and everything else becomes much easier.
If you want to go deeper on channel width selection, check out the Purple guide on 20 MHz versus 40 MHz versus 80 MHz. And if you're looking at deploying guest WiFi with analytics at scale, the Purple platform is hardware-agnostic and integrates with your existing infrastructure.
Thanks for listening. Until next time.
Part of our core series: WiFi RF Engineering Guide →
- What is RSSI?
- RSSI, RCPI and dBm: which number are you reading?
- RSSI thresholds by application
- Why does RSSI matter for channel planning?
- How does RSSI work with SNR, attenuation and channel width?
- SNR: the figure that sets your data rate
- Attenuation: how building materials change RSSI
- Channel width and the 3 dB rule
- How do you turn RSSI targets into a channel plan?
- Step 1: design for your least capable, most important device
- Step 2: run an active survey as well as a predictive one
- Step 3: match AP power to client power
- Step 4: remove low data rates
- Step 5: enable 802.11k, 802.11v and 802.11r
- Step 6: audit automatic channel assignment
- What does this look like in a real venue?
- Worked scenario: a 200-room hotel
- Worked scenario: a grocery store with handheld scanners
- Worked scenario: a 3,000-seat conference hall
- Where does RSSI planning sit alongside the WiFi platform you already run?
- What are the limits of RSSI you should know about?
- What should you do next?
- Frequently asked questions
- Does Purple work with the access points we already own?
- Will moving our guest WiFi to Purple change our RF design?
- What RSSI should we design for in a hotel or retail store?
- Do we need a site survey before adding more access points?
- Is 6 GHz worth enabling at our venues?
- Does collecting RSSI and presence data create GDPR obligations?
- Does automatic radio resource management replace manual channel planning?
Interactive engineering tool
WiFi RSSI, SNR & channel planning advisor
Simulate enterprise received signal strength (RSSI), evaluate signal-to-noise ratio (SNR), prevent co-channel interference, and calculate roaming handoff boundaries for high-density venues.
-65 dBm
-90 dBm (weak)-65 dBm (target)-40 dBm (strong)
-90 dBm
-100 dBm (clean)-90 dBm (standard)-75 dBm (noisy)
14 dBm (~25 mW)
8 dBm (microcell)14 dBm (balanced)24 dBm (maximum)
Signal-to-noise ratio
25 dB
Good link condition
Target: >= 25 dB for Voice over WiFi & video conferencing
Roaming handoff margin
+7 dB
Reliable roaming cell
Trigger threshold: -72 dBm (802.11k/v/r)
Co-channel interference
moderate
Power gap: 2 dB vs client
Venue ceiling: 12 dBm
RF design analysis & engineering recommendations
- Modulation and data rate capacity: MCS 5 to 8 (64-QAM up to 256-QAM at 3/4) - High throughput for business workflows; 256-QAM MCS 8-9 needs roughly 28 to 32 dB SNR.
- Transmit power balance: Transmit power is slightly elevated for this environment. Monitor co-channel contention and consider reducing power by 2 to 3 dBm.
- Roaming handoff behaviour: Solid primary coverage (+7 dB above roaming trigger). Fast roaming candidates are evaluated proactively without audio or video packet loss. Ensure secondary AP cell overlap reaches at least -67 dBm for uninterrupted voice and video handoff.
- Channel width policy: 20 MHz recommended (40 MHz maximum). Narrower channels double available channel count and reduce noise floor by 3 dB per step.
Related network engineering resources:WiFi channel selection guide802.11k/v/r roaming & handoffGuest WiFi master guide
Useful? Link to this tool
RSSI (Received Signal Strength Indicator) measures how strongly a device hears a WiFi signal. Tools show it in dBm as a negative number. Readings near -30 dBm are excellent, -67 dBm is the usual floor for voice, and below -80 dBm is unreliable. Channel planning pairs RSSI with signal-to-noise ratio (SNR) to set access point placement, power and channel reuse.
What is RSSI?
RSSI is the power a client device or access point (AP) measures on the radio frames it receives. Every survey tool, controller dashboard and laptop status icon reports it first. WiFi signals arrive at tiny fractions of a milliwatt, so RSSI uses a logarithmic scale in decibel-milliwatts (dBm). Zero dBm equals one milliwatt, so real WiFi readings are negative. The closer the number is to zero, the stronger the signal.
RSSI, RCPI and dBm: which number are you reading?
The IEEE 802.11 standard defines RSSI as an optional, relative index from 0 to 255. Each chipset vendor sets its own scale. That is why two laptops in the same spot can disagree by several decibels. Drivers and survey tools convert the index to dBm, but each vendor calibrates differently.
IEEE 802.11k added the Received Channel Power Indicator (RCPI), defined in dBm in 0.5 dB steps. APs that support 802.11k use RCPI in beacon and neighbour reports. The lesson for planners is consistency. Survey with one calibrated adapter, and never merge phone, laptop and controller readings into one dataset.
RSSI thresholds by application
| RSSI (dBm) | Signal quality | What it reliably supports |
|---|---|---|
| -30 to -50 | Excellent | Highest data rates, including 256-QAM where SNR allows |
| -51 to -67 | Good | Voice over WiFi, video calls, location services, high-throughput data |
| -68 to -70 | Fair | Web browsing, email, guest login pages |
| -71 to -80 | Poor | Basic connectivity only; voice and video unstable |
| Below -80 | Unusable | Close to the -82 dBm sensitivity floor; frequent disconnections |
The -67 dBm boundary and its 25 dB SNR companion come from Cisco's Voice over WLAN design guidance. This guide uses them as the design target for real-time traffic. The -82 dBm floor is the minimum receiver sensitivity the 802.11ac standard sets for its lowest data rate on a 20 MHz channel.
Why does RSSI matter for channel planning?
Channel planning decides which APs share airtime. RSSI decides how far each AP's signal travels before its neighbours stop hearing it. Get both wrong and the heatmap looks green while the network fails at peak.
WiFi shares airtime through carrier sense multiple access with collision avoidance (CSMA/CA). Every radio listens before it transmits. The 802.11 standard requires a radio to defer when it detects a WiFi preamble at -82 dBm or stronger on a 20 MHz channel. Two APs on channel 36 that hear each other at -78 dBm therefore split one channel's airtime. That is co-channel interference (CCI). It is a capacity problem, not a noise problem.
Adjacent channel interference (ACI) is different. It occurs when channels overlap in frequency, raising the noise floor and cutting SNR. In 2.4 GHz, only channels 1, 6 and 11 avoid overlap. Any other assignment pushes ACI into its neighbours.
The planning goal follows from both. Every visitor's device should hear one AP at -67 dBm or better. Ideally, it should hear other APs on the same channel below -82 dBm. RSSI is the measurement that proves you have achieved it.
How does RSSI work with SNR, attenuation and channel width?
SNR: the figure that sets your data rate
SNR is RSSI minus the noise floor. An RSSI of -65 dBm over a noise floor of -90 dBm gives 25 dB of SNR. Radios pick their modulation and coding scheme (MCS) from link quality, so SNR matters more than raw signal. A strong RSSI over a raised noise floor still produces slow, unstable links.
Most non-WiFi interference sits in 2.4 GHz: microwave ovens, Bluetooth, video senders and older 2.4 GHz cordless phones. European DECT phones use 1880-1900 MHz, so they do not interfere with WiFi in the UK. Use a spectrum analyser to find these sources, because a WiFi scanner cannot see them.
Higher-order modulation needs more signal. The 802.11ac standard sets a minimum receiver sensitivity of -82 dBm for MCS 0 on a 20 MHz channel. For 256-QAM (quadrature amplitude modulation) at MCS 9, it sets -57 dBm. That 25 dB gap separates a link that merely works from one that runs at full speed.
Attenuation: how building materials change RSSI
Decibels are logarithmic. A 3 dB loss halves signal power, a 10 dB loss cuts it tenfold, and 20 dB cuts it a hundredfold. A signal crossing two brick walls can arrive with around 1% of the power that reached the first wall.
| Material | Typical planning loss at 5 GHz |
|---|---|
| Plasterboard partition | 3-4 dB |
| Standard glass | 2-3 dB |
| Brick wall | 8-12 dB |
| Concrete wall | 12-15 dB |
| Reinforced concrete or steel | 15-25 dB |
| Metal shelving with stock | 10-20 dB |
Treat these as starting assumptions for predictive design; your active survey replaces them. Loss through most materials rises with frequency, so 6 GHz cells shrink faster than 5 GHz cells behind the same wall. Some environments add extreme losses. Lead-lined imaging rooms on Healthcare sites need dedicated coverage. So do metallised carriage windows on Trains; the next room or carriage cannot cover them.
Channel width and the 3 dB rule
Thermal noise across a 20 MHz channel is about -101 dBm, and every doubling of width adds 3 dB. An 80 MHz channel therefore needs 6 dB more signal than a 20 MHz channel for the same SNR. The 802.11ac sensitivity table shows exactly this: 256-QAM needs -57 dBm at 20 MHz and -51 dBm at 80 MHz.
Width also cuts channel count. In the UK, ETSI bands A and B give 19 non-overlapping 20 MHz channels at 5 GHz, but only four at 80 MHz. Fewer channels mean more APs reuse each one, which means more CCI. Dynamic Frequency Selection (DFS) channels must also check for radar before use.
| Band | 20 MHz channels (UK) | DFS radar checks | Wall penetration | Best fit |
|---|---|---|---|---|
| 2.4 GHz | 3 (1, 6, 11) | No | Strongest | Legacy and IoT devices |
| 5 GHz | 19 (ETSI bands A and B) | Yes, on 15 of 19 | Moderate | Guest and staff data, voice |
| 6 GHz (5925-6425 MHz) | 24 | No | Weakest | WiFi 6E and WiFi 7 capacity layer; WPA3 mandatory |
How do you turn RSSI targets into a channel plan?
Step 1: design for your least capable, most important device
Identify the least capable, most important (LCMI) device: the weakest radio that must work everywhere. In a warehouse, it is often an ageing handheld scanner. In a hospital, it may be a specific infusion pump or nurse call handset. In a hotel, it is an older smartphone. Design RSSI and SNR to that device's data sheet, and newer devices perform better. Plan around 20% cell overlap, as Cisco's voice guidance recommends, so clients always have a second AP to roam to.
Step 2: run an active survey as well as a predictive one
Predictive software models your floor plan using the material losses above. An active survey measures real RSSI, noise floor and SNR with a connected client. Record noise everywhere, because it caps the SNR you can reach. Survey at device height, where people actually hold phones. Bodies absorb WiFi energy, so repeat the survey in occupied conditions where you can.
Step 3: match AP power to client power
Louder is not better. Phones typically transmit at lower power than enterprise APs. At maximum AP power, the client hears the AP clearly but the AP struggles to hear the client. That asymmetry creates sticky clients, which cling to a distant AP at -80 dBm while a closer one sits idle. Check your LCMI device's transmit power and set AP power close to it.
Step 4: remove low data rates
Disable 1, 2, 5.5 and 11 Mbps in 2.4 GHz, and 6 and 9 Mbps in 5 GHz, where your LCMI device allows. Clients then drop a weak AP earlier and roam sooner. Beacons transmit at the lowest enabled rate, so removing slow rates returns airtime to every visitor.
Step 5: enable 802.11k, 802.11v and 802.11r
802.11k gives clients a list of neighbouring APs. 802.11v lets the network steer a client to a better AP. 802.1X, the IEEE standard for port-based access control, normally runs a full authentication on every new connection. 802.11r, Fast BSS Transition, lets clients skip that exchange when they roam. Together, the three move roaming decisions from client inertia to RF conditions.
Step 6: audit automatic channel assignment
Every enterprise vendor ships radio resource management (RRM). It reacts to what APs hear from each other, not what devices hear at desk height. Audit the plan after deployment and pin channels or power where survey data disagrees.
| Venue type | Density | 2.4 GHz width | 5 GHz width | 6 GHz width |
|---|---|---|---|---|
| Stadium, arena or conference hall | High | 20 MHz, alternate radios off | 20 MHz | 20 MHz |
| Hotel or retail store | Medium | 20 MHz | 40 MHz | 80 MHz |
| Hospital ward | Medium, voice-critical | 20 MHz | 20 MHz | 40 MHz |
| Office or meeting suite | Low | 20 MHz | 80 MHz | 160 MHz |
Got questions about your specific setup?
Our team works with venue operators, IT managers, and network engineers across 80,000 venues. Book a 20-minute call and we will show you how others like you solved it.
What does this look like in a real venue?
These worked scenarios are illustrative designs built on the figures above, not named deployments.
Worked scenario: a 200-room hotel
Situation. A 200-room hotel runs one AP in every second corridor bay at full power. Rooms sit behind brick partitions and en-suite bathroom pods. A survey shows -76 to -80 dBm at the bed and desk, over a noise floor of -92 dBm. SNR is 12-16 dB, so video calls drop. The captive portal, the login page guests see before getting online, loads slowly.
Action. The team moves to in-room APs in every third room. APs are staggered by floor so same-channel radios never stack vertically. 5 GHz runs at 40 MHz on a fixed plan, with AP power matched to a typical smartphone. Rates of 1-11 Mbps are disabled, and 802.11k, 802.11v and 802.11r are enabled.
Outcome. The validation survey shows -55 to -65 dBm in every room and SNR of 27 dB or more. That clears the -67 dBm and 25 dB voice target with margin. Same-channel APs on adjacent floors now read below -82 dBm, so each AP has its channel to itself. In-room design like this is common across Hotels.
Worked scenario: a grocery store with handheld scanners
Situation. A 3,000 m² grocery store uses handheld scanners for stock checks and click-and-collect picking. The scanners are the LCMI device, with lower transmit power than the APs. Tall metal shelving adds 10-20 dB of loss between aisles. Scanners disconnect at aisle ends, and shoppers' guest WiFi struggles at the tills.
Action. APs move to aisle ends, with directional antennas pointed down each aisle rather than across shelf tops. Power is set to the scanner's transmit level. The team surveys with the scanner itself, at waist height, with shelves fully stocked.
Outcome. Every aisle reaches -67 dBm or better at scanner height, with at least 25 dB SNR. The asymmetric links disappear, so scanners roam between aisles instead of clinging to the entrance AP. The same layout applies across Retail estates with similar shelving.
Worked scenario: a 3,000-seat conference hall
Situation. A conference hall hosts 3,000 attendees for keynotes. Every seat reads -50 dBm or stronger. Yet channel utilisation exceeds 50% before the first session, and attendees cannot load the event app.
Action. The problem is CCI, not coverage. The team cuts 5 GHz from 80 MHz to 20 MHz, raising the channel count from four to 19. Ceiling APs are replaced with lower-power directional mounts, each covering one block of seats. 2.4 GHz radios are disabled on alternate APs, and a 6 GHz layer is added for WiFi 6E devices.
Outcome. Same-channel APs now hear each other at or below -82 dBm, so each cell gets its own airtime. In the full-occupancy validation test, peak utilisation stays below 40%. The hall now runs 19 independent channels at 5 GHz, plus 24 at 6 GHz.
Where does RSSI planning sit alongside the WiFi platform you already run?
RSSI and channel planning live in your WLAN controller or cloud dashboard. Purple does not replace that layer. Purple is a hardware-agnostic cloud overlay for Cisco Meraki, HPE Aruba, Ruckus, Juniper Mist, Ubiquiti UniFi, Cambium, Extreme and Fortinet. Your APs, channels and power settings stay where they are.
Purple works above the RF layer. Guest WiFi runs the captive portal and authentication. Its conscious-choice opt-ins capture first-party data. WiFi Analytics turns connection data into visit reporting by venue. For staff, Identity-Based Networks and SecurePass replace shared passwords with individual credentials. Purple integrates with Microsoft Entra ID, Okta and Google Workspace to revoke access automatically when someone leaves.
All of it depends on the RF work in this guide. A captive portal served over a -80 dBm link at the lowest data rate feels broken, however well designed. Analytics built on patchy coverage under-count visits in weak areas. Purple runs across 80,000+ live venues in more than 90 countries and handles 500 million users and logins a year.
What are the limits of RSSI you should know about?
- It is vendor-relative. The 802.11 RSSI index has no common scale, so readings differ by chipset. Compare trends from one device, not absolute values across many.
- It is one-directional. RSSI reports what the receiver heard. The AP may hear the client far worse, which is why power matching matters.
- It ignores noise and contention. A -55 dBm reading says nothing about the noise floor or channel utilisation. Pair it with SNR and airtime data.
- It changes with people. Bodies absorb energy, so an empty-venue survey overstates RSSI for a sell-out crowd.
- It does not identify devices reliably. Since iOS 14 and Android 10, phones use randomised MAC addresses per network by default. Presence data from unconnected devices' probe RSSI is less accurate as a result. Authenticated, opt-in data through a captive portal avoids the problem.
- DFS can move your channel. Under ETSI EN 301 893, a radio must check a DFS channel for radar for 60 seconds before use. Weather radar channels need 10 minutes. A detection forces a channel change and a brief interruption, so keep critical areas on non-DFS channels.
What should you do next?
- Pull your LCMI device's data sheet and note its transmit power and receive sensitivity.
- Commission an active survey that records RSSI, noise floor and SNR at device height, ideally when occupied.
- Compare results against the -67 dBm and 25 dB targets, area by area.
- Fix power, channel width and data rates before buying more APs.
- Re-survey, then audit RRM decisions monthly against that baseline.
Once the RF layer is sound, Purple adds guest onboarding, analytics and staff access on the hardware you already own, with no rip and replace.
Frequently asked questions
Does Purple work with the access points we already own?
Yes. Purple is hardware-agnostic and runs as a cloud overlay on Cisco Meraki, HPE Aruba, Ruckus, Juniper Mist, Ubiquiti UniFi, Cambium, Extreme and Fortinet networks. Your APs, controller and channel plan stay in place. Purple adds the captive portal, authentication and analytics layer on top, so there is no rip and replace. The Purple support centre covers the integration steps for each vendor platform.
Will moving our guest WiFi to Purple change our RF design?
No. Purple's Guest WiFi sits on your existing SSIDs and leaves AP placement, channels, transmit power and data rates untouched. Those settings stay in your controller or cloud dashboard. Migration usually means pointing your guest SSID at Purple's captive portal and authentication service. Still check RSSI at onboarding points such as lobbies and entrances. A weak cell edge makes any login page feel broken.
What RSSI should we design for in a hotel or retail store?
Design for -67 dBm or stronger with at least 25 dB SNR wherever guests or staff use voice, video or handheld devices. That target follows Cisco's Voice over WLAN design guidance. It also leaves margin above the -82 dBm sensitivity floor in the 802.11ac standard. Areas used only for light browsing can tolerate around -70 dBm. Measure at device height, not at ceiling level.
Do we need a site survey before adding more access points?
Yes. A predictive survey shows where APs should go, but only an active survey measures real RSSI, noise floor and SNR in your building. Furniture, stock, people and neighbouring networks all change the result. Survey before and after deployment with the same calibrated adapter, and repeat after any major refit. Adding APs without a survey often increases co-channel interference rather than capacity.
Is 6 GHz worth enabling at our venues?
Yes, where enough of your visitors carry WiFi 6E or WiFi 7 devices. In the UK, Ofcom's lower 6 GHz band adds 24 clean 20 MHz channels with no DFS radar checks. Every 6 GHz network must use WPA3, which strengthens guest security. The trade-off is range: 6 GHz loses more signal through walls. Plan it as a capacity layer alongside 5 GHz, not a replacement.
Does collecting RSSI and presence data create GDPR obligations?
Yes, when the data links to an identifiable device. The ICO's guidance on WiFi location analytics treats MAC addresses as potentially personal data. You therefore need a lawful basis, clear notices and retention limits. Purple is ISO 27001 certified and GDPR compliant, and collects first-party data through conscious-choice opt-ins at the captive portal. Survey data held without device identifiers carries far lower risk.
Does automatic radio resource management replace manual channel planning?
No. Radio resource management (RRM) from your AP vendor is a sound baseline, and Cisco Meraki, HPE Aruba and Juniper Mist all ship it. RRM reacts to what APs hear from each other, not what devices hear in a guest room or aisle. Use it for day-to-day adjustments. Then audit the plan after deployment and pin channels or power where survey data disagrees.
Key Definitions
RSSI (Received Signal Strength Indicator)
An optional, relative index from 0 to 255 defined in the IEEE 802.11 standard, with each chipset vendor setting its own scale. Drivers and survey tools convert it to dBm, but calibration differs by vendor.
IT teams see RSSI first in every survey tool and controller dashboard. Because it is vendor-relative, compare trends from one device rather than absolute values across many.
dBm (decibel-milliwatts)
A logarithmic power scale where 0 dBm equals one milliwatt. A 3 dB loss halves power, 10 dB cuts it tenfold and 20 dB cuts it a hundredfold, so WiFi readings are negative.
Every RSSI target in a channel plan is expressed in dBm. Understanding the logarithmic scale explains why two brick walls can leave around 1% of the original signal power.
RCPI (Received Channel Power Indicator)
A measurement added by the IEEE 802.11k amendment, defined in dBm in 0.5 dB steps. APs that support 802.11k use RCPI in beacon and neighbour reports.
RCPI gives a more consistent reading than vendor RSSI. It matters when you rely on 802.11k neighbour reports to guide client roaming.
SNR (signal-to-noise ratio)
RSSI minus the noise floor, in dB. Radios choose their modulation and coding scheme from link quality, so SNR sets the achievable data rate more than raw signal does.
Active surveys must record the noise floor everywhere. A strong RSSI over a raised noise floor still produces slow, unstable links, as the hotel scenario's 12-16 dB SNR showed.
CSMA/CA (carrier sense multiple access with collision avoidance)
The IEEE 802.11 medium access method in which every radio listens before transmitting. The standard requires a radio to defer when it detects a WiFi preamble at -82 dBm or stronger on a 20 MHz channel.
CSMA/CA is why same-channel APs that hear each other split airtime. It defines the -82 dBm separation target in channel reuse planning.
Co-channel interference (CCI)
Airtime contention between radios on the same channel that hear each other above the 802.11 -82 dBm preamble detection threshold. It is a capacity problem, not a noise problem.
CCI explains why a dense venue can read -50 dBm at every seat yet fail at peak. Narrower channels and lower-power directional APs reduce it.
Adjacent channel interference (ACI)
Interference from channels that overlap in frequency, raising the noise floor and cutting SNR. In 2.4 GHz, only channels 1, 6 and 11 avoid overlap.
Any 2.4 GHz assignment outside 1, 6 and 11 pushes ACI into neighbouring cells. Fixed, non-overlapping plans keep SNR high.
MCS (modulation and coding scheme)
The data rate setting a radio selects from link quality. The IEEE 802.11ac standard sets minimum receiver sensitivity of -82 dBm for MCS 0 and -57 dBm for 256-QAM at MCS 9 on a 20 MHz channel, rising to -51 dBm at 80 MHz.
The 25 dB gap between MCS 0 and MCS 9 separates a link that merely works from one that runs at full speed. Wider channels need 3 dB more signal per doubling.
DFS (Dynamic Frequency Selection)
A radar-avoidance requirement under ETSI EN 301 893. A radio must check a DFS channel for radar for 60 seconds before use, or 10 minutes on weather radar channels, and change channel on detection.
DFS applies to 15 of the 19 UK 5 GHz channels. Keep critical areas such as voice zones on non-DFS channels to avoid interruptions.
LCMI (least capable, most important) device
The weakest radio that must work everywhere on the network, whose data sheet transmit power and receive sensitivity set the RSSI and SNR design targets.
Examples include ageing warehouse scanners, infusion pumps or older smartphones. Designing to this device means newer devices perform better.
802.11r (Fast BSS Transition)
An IEEE 802.11 amendment that lets clients skip the full IEEE 802.1X port-based authentication exchange when roaming between APs. It works alongside 802.11k neighbour lists and 802.11v network-directed steering.
Enabling 802.11k, 802.11v and 802.11r together moves roaming decisions from client inertia to RF conditions, which supports voice and handheld devices.
RRM (radio resource management)
Automatic channel and power assignment shipped by enterprise AP vendors, including Cisco Meraki, HPE Aruba and Juniper Mist. It reacts to what APs hear from each other, not what devices hear at desk height.
RRM is a sound baseline for day-to-day adjustment. Audit it against survey data monthly and pin channels or power where it disagrees.
Worked Examples
A 200-room hotel runs one AP in every second corridor bay at full power. Guest rooms read -76 to -80 dBm over a -92 dBm noise floor, SNR is 12-16 dB, video calls drop and the captive portal loads slowly. What should change?
The team moved to in-room APs in every third room, staggered by floor so same-channel radios never stack vertically. 5 GHz runs at 40 MHz on a fixed plan, with AP power matched to a typical smartphone. Rates of 1-11 Mbps were disabled, and 802.11k, 802.11v and 802.11r were enabled. Corridor APs had to cross brick partitions and bathroom pods, which wiped out SNR. The validation survey shows -55 to -65 dBm in every room and SNR of 27 dB or more, clearing the -67 dBm and 25 dB voice target. Same-channel APs on adjacent floors now read below -82 dBm, so each AP has its channel to itself.
A 3,000 m² grocery store uses handheld scanners with lower transmit power than its APs. Tall metal shelving adds 10-20 dB of loss, scanners disconnect at aisle ends and shopper guest WiFi struggles at the tills. How do you fix it?
The scanners are the LCMI device, so the design follows their data sheet. APs moved to aisle ends with directional antennas pointed down each aisle, not across shelf tops, avoiding the shelving loss. AP power was set to the scanner's transmit level to remove the asymmetric link that made scanners cling to the entrance AP. The team surveyed with the scanner itself, at waist height, with shelves fully stocked, so the measurements reflected real conditions. Every aisle now reaches -67 dBm or better at scanner height with at least 25 dB SNR, and scanners roam between aisles instead of sticking to a distant AP.
A 3,000-seat conference hall reads -50 dBm or stronger at every seat, yet channel utilisation exceeds 50% before the first session and attendees cannot load the event app. What is wrong?
The problem is co-channel interference, not coverage. At 80 MHz, 5 GHz offers only four channels, so many APs shared each one and split airtime under CSMA/CA. The team cut 5 GHz to 20 MHz, raising the channel count from four to 19. Ceiling APs were replaced with lower-power directional mounts, each covering one block of seats. 2.4 GHz radios were disabled on alternate APs, and a 6 GHz layer was added for WiFi 6E devices. Same-channel APs now hear each other at or below -82 dBm, so each cell gets its own airtime. In the full-occupancy validation test, peak utilisation stays below 40%.
Practice Questions
Q1. You are designing a WiFi network for a 40,000-seat stadium. The venue operator wants maximum throughput for concurrent video streaming and social media uploads during events. You are considering using 80 MHz channels in the 5 GHz band to maximise per-client throughput. Is this the recommended approach, and what channel plan would you implement instead?
Hint: Consider the number of non-overlapping 80 MHz channels available in the 5 GHz band versus 20 MHz channels, and the impact of Co-Channel Interference in an open, high-density environment.
View model answer
No. Using 80 MHz channels in a stadium is strongly contraindicated. In the standard 5 GHz UNII-1/2/2e bands, there are only a handful of non-overlapping 80 MHz channels, meaning that with the AP density required for 40,000 concurrent users, severe CCI is inevitable. The correct approach is to use 20 MHz channels throughout, which provides up to 24 non-overlapping channels in 5 GHz (including DFS), maximising channel reuse. Directional sector antennas should be used to tightly control RF cell coverage, pointing down into seating sections rather than radiating omnidirectionally. AP density should be calculated based on a target of no more than 30-50 clients per AP radio, with transmit power tuned to match the coverage area of each sector.
Q2. A warehouse deployment uses handheld barcode scanners that frequently drop connections when operators move between aisles. The APs are configured at maximum transmit power (23 dBm) to ensure full coverage. The scanners run a legacy WMS application that requires sub-100ms latency. What is the likely cause and what steps would you take to resolve it?
Hint: Consider the transmit power capabilities of a small handheld scanner versus an enterprise AP, and the implications for the link budget in both directions.
View model answer
The likely cause is the sticky client problem resulting from an asymmetric link budget. The APs are transmitting at 23 dBm, so the scanners hear them well across the entire warehouse and do not initiate a roam. However, the scanners' internal radios typically transmit at only 15-17 dBm, meaning the AP cannot reliably receive the scanner's transmissions when it is far away. The solution is to lower AP transmit power to 10-12 dBm to match the scanners' capabilities, ensuring that the coverage cells are appropriately sized and that scanners roam when they move out of range. Enable 802.11k/v/r to facilitate fast roaming. Set minimum mandatory data rates to 12 Mbps to force earlier roaming decisions. Validate with an active site survey using the actual scanner hardware to confirm -65 dBm RSSI and 25 dB SNR throughout all aisles.
Q3. During a site survey for a new hospital wing, you measure an RSSI of -58 dBm from the primary AP throughout the target area. However, the noise floor measured by a spectrum analyser is consistently -72 dBm due to legacy medical monitoring equipment operating in the 2.4 GHz band. The hospital requires reliable VoWiFi for clinical communications. Will this network support VoWiFi, and what actions would you recommend?
Hint: Calculate the SNR and evaluate it against the minimum requirement for VoWiFi. Consider which frequency band is affected and what mitigation options are available.
View model answer
No, this network will not reliably support VoWiFi in its current state. The SNR is calculated as -58 dBm - (-72 dBm) = 14 dB. This falls below the minimum 20 dB SNR required for VoWiFi and well below the 25 dB target for high-quality voice. Despite the strong RSSI of -58 dBm, the elevated noise floor from the medical equipment degrades the link quality to an unacceptable level. Recommended actions: First, migrate VoWiFi traffic to the 5 GHz band, which is largely unaffected by the legacy 2.4 GHz medical equipment. Second, increase AP density in the affected areas to improve RSSI to -50 dBm or better, which would yield an SNR of 22 dB even with the elevated noise floor - marginally acceptable for VoWiFi. Third, engage the biomedical engineering team to assess whether the legacy equipment can be replaced or shielded. Fourth, implement QoS (WMM) with voice traffic prioritisation to protect VoWiFi traffic from competing with data traffic during periods of congestion.
Frequently asked questions
Does Purple work with the access points we already own?
Yes. Purple is hardware-agnostic and runs as a cloud overlay on Cisco Meraki, HPE Aruba, Ruckus, Juniper Mist, Ubiquiti UniFi, Cambium, Extreme and Fortinet networks. Your APs, controller and channel plan stay in place. Purple adds the captive portal, authentication and analytics layer on top, so there is no rip and replace. The Purple support centre covers the integration steps for each vendor platform.
Will moving our guest WiFi to Purple change our RF design?
No. Purple's Guest WiFi sits on your existing SSIDs and leaves AP placement, channels, transmit power and data rates untouched. Those settings stay in your controller or cloud dashboard. Migration usually means pointing your guest SSID at Purple's captive portal and authentication service. Still check RSSI at onboarding points such as lobbies and entrances. A weak cell edge makes any login page feel broken.
What RSSI should we design for in a hotel or retail store?
Design for -67 dBm or stronger with at least 25 dB SNR wherever guests or staff use voice, video or handheld devices. That target follows Cisco's Voice over WLAN design guidance. It also leaves margin above the -82 dBm sensitivity floor in the 802.11ac standard. Areas used only for light browsing can tolerate around -70 dBm. Measure at device height, not at ceiling level.
Do we need a site survey before adding more access points?
Yes. A predictive survey shows where APs should go, but only an active survey measures real RSSI, noise floor and SNR in your building. Furniture, stock, people and neighbouring networks all change the result. Survey before and after deployment with the same calibrated adapter, and repeat after any major refit. Adding APs without a survey often increases co-channel interference rather than capacity.
Is 6 GHz worth enabling at our venues?
Yes, where enough of your visitors carry WiFi 6E or WiFi 7 devices. In the UK, Ofcom's lower 6 GHz band adds 24 clean 20 MHz channels with no DFS radar checks. Every 6 GHz network must use WPA3, which strengthens guest security. The trade-off is range: 6 GHz loses more signal through walls. Plan it as a capacity layer alongside 5 GHz, not a replacement.
Does collecting RSSI and presence data create GDPR obligations?
Yes, when the data links to an identifiable device. The ICO's guidance on WiFi location analytics treats MAC addresses as potentially personal data. You therefore need a lawful basis, clear notices and retention limits. Purple is ISO 27001 certified and GDPR compliant, and collects first-party data through conscious-choice opt-ins at the captive portal. Survey data held without device identifiers carries far lower risk.
Does automatic radio resource management replace manual channel planning?
No. Radio resource management (RRM) from your AP vendor is a sound baseline, and Cisco Meraki, HPE Aruba and Juniper Mist all ship it. RRM reacts to what APs hear from each other, not what devices hear in a guest room or aisle. Use it for day-to-day adjustments. Then audit the plan after deployment and pin channels or power where survey data disagrees.
Continue reading in this series
20MHz vs 40MHz vs 80MHz: Which Channel Width Should You Use?
This guide provides a definitive, vendor-neutral technical reference for IT managers, network architects, and venue operations directors on selecting the correct WiFi channel width - 20MHz, 40MHz, or 80MHz - across enterprise deployments in hospitality, retail, events, and public-sector environments. It covers the underlying IEEE 802.11 mechanics, real-world capacity trade-offs, and step-by-step deployment guidance to help teams make the right call this quarter. Understanding channel width selection is one of the highest-leverage decisions in any wireless LAN design, directly impacting throughput, interference, client density support, and the reliability of guest-facing services.
WiFi 6 vs WiFi 5: Does it Solve Channel Interference?
This guide provides a technical deep-dive into how WiFi 6 (802.11ax) addresses channel interference in high-density enterprise environments through OFDMA and BSS Coloring. It equips IT managers, network architects, and CTOs with actionable deployment strategies, real-world case studies from hospitality and healthcare, and a framework for evaluating the ROI of infrastructure upgrades in venues where wireless performance is business-critical.
5GHz DFS WiFi Channels: When to Use & Avoid in Enterprise
Learn how 5GHz DFS WiFi channels work, radar interference risks, CAC wait times, weather radar channels, and enterprise channel planning best practices.
Got questions about your specific setup?
Our team works with venue operators, IT managers, and network engineers across 80,000 venues. Book a 20-minute call and we will show you how others like you solved it.