Apartment Wi-Fi Congestion: Channels, Neighbors and Noise

Your phone shows three bars in the kitchen and a web page still takes eight seconds to open. Nothing is far away and nothing is broken. You are sharing the air with most of the building, and between 7 and 11 pm they all show up at once.

That is channel congestion — the one Wi-Fi problem where adding hardware usually makes things worse, and one of the few a free settings change genuinely fixes, provided you change the right setting.

Congestion costs you airtime, not signal#

Wi-Fi is half duplex and polite. Before any radio transmits it listens, and if it hears another Wi-Fi signal above roughly -82 dBm it waits. That rule keeps a channel usable when several networks share it, and it is also the mechanism that hands your bandwidth to strangers.

Your neighbor is not weakening your signal. What you have lost is airtime — the fraction of each second your devices are allowed to speak. If your network and three busy neighbors share a channel, you get roughly a quarter of the seconds available, no matter how many bars your phone draws.

There are two ways a neighbor takes your airtime, and they are not equally bad.

• Co-channel. They sit on exactly your channel. The radios hear each other, defer and take turns. Throughput drops and latency gets lumpy, but the link stays sane.
• Adjacent-channel. They partially overlap you — you are on channel 3, they are on 1 and 6. Their energy lands inside your channel but does not decode as a Wi-Fi preamble, so nothing defers. It raises your noise floor, corrupts frames mid-flight and forces retransmissions.

Partial overlap is worse than sharing. That one fact drives the next two sections. For the non-Wi-Fi half of the noise story — microwaves, cameras, USB 3 enclosures — see what actually interferes with Wi-Fi, and for what a raised noise floor does to a room, how dead zones are built.

How to see the noise floor and the traffic#

The numbers you want are not the ones on your phone's status bar.

• Your router's admin page is the best single source. Look for a Wireless Status, Site Survey or Channel Analysis screen; mid-range routers report channel utilization as a percentage and often a noise figure in dBm.
• Android analyzer apps list the channel and RSSI of every visible network. macOS has Wireless Diagnostics; Windows answers netsh wlan show networks mode=bssid.
• iOS does not let third-party apps scan for nearby networks at all, so an iPhone cannot do this job — see measurement apps versus planning apps.

Take the readings from the seat that matters, at the hour that hurts. A survey next to the router at 11 am describes a home you do not live in.

Those are rules of thumb, not thresholds any standard defines. The pattern matters more than the figure: strong signal, weak throughput, big swing by time of day. If yours is mostly an evening problem, run the diagnostic in why Wi-Fi gets worse in the evening first — congestion is only one of three suspects there.

Pick 1, 6 or 11, and nothing in between#

The 2.4 GHz band in the US has 11 channels spaced 5 MHz apart, but each one is 20 MHz wide, so they overlap almost entirely. Only 1, 6 and 11 are genuinely independent. Channel 4 is not a clever gap between the crowds — it is a radio that interferes with everyone on 1 and everyone on 6 while deferring to neither.

Here is the whole procedure.

1. Survey from your usual seat at around 9 pm, not from the router at noon.
2. Note every network stronger than about -85 dBm, with its channel and width. Anything weaker sits close enough to the noise floor to ignore.
3. Sort them into three buckets by which of 1, 6 and 11 they overlap. A neighbor on channel 3 counts against both 1 and 6; a 40 MHz network counts against everything it covers.
4. Score each bucket by strength, not by headcount. One neighbor at -50 dBm costs you more airtime than six at -78 dBm.
5. Pick the lightest bucket and set the channel fixed. Do the same on 5 GHz, choosing a channel no strong neighbor sits on.
6. Retest a day later in the same seat, at the same hour, doing the thing that annoyed you.

Two counterintuitive notes. If the lightest bucket still holds one very loud neighbor, take it anyway — co-channel contention with a radio you can hear beats half-overlapping one you cannot. And turn auto-channel off once you have chosen: many routers pick only at boot, so one that restarted at 4 am optimized itself for a sleeping building.

Channel width is the setting that matters more than the channel#

Width is the lever most people never touch, and in a dense building it usually beats a channel change.

The physics is simple: a wider channel collects proportionally more noise, so every doubling of width lifts your effective noise floor by about 3 dB. An 80 MHz link is therefore roughly 6 dB more fragile at the far end of the apartment than a 20 MHz one, and it steps on four times as many neighbors.

The trade is real throughput at the edge against headline throughput next to the router. Where you can hear fifteen networks on 5 GHz, dropping from 80 to 40 MHz frequently makes the far bedroom faster, because the link stops thrashing. Keep 2.4 GHz at 20 MHz permanently, and see 2.4 versus 5 versus 6 GHz for which band carries what.

DFS: the half of 5 GHz nobody is using#

Budget routers ship on UNII-1 (36 to 48) and UNII-3 (149 to 165): nine 20 MHz channels, only two clean 80 MHz slots for a whole building, which is exactly why they are jammed.

Channels 52 to 64 and 100 to 144 are DFS — dynamic frequency selection, shared with weather and aviation radar. They roughly triple your usable 5 GHz spectrum and in most residential buildings sit close to empty. The catches are real but manageable.

• Before transmitting, the router runs a Channel Availability Check — typically 60 seconds, and 600 seconds on channels 120 to 128 near the weather radar band. Your 5 GHz network is offline for that window.
• If radar is detected the router must vacate within 10 seconds and stay off that channel for 30 minutes. In practice your clients drop for a minute or so while it moves. Within a few miles of an airport or coastal radar this can happen weekly.
• Not every client supports DFS. Older streamers, printers, smart TVs and cheap IoT gear may never see the network.

Try channel 100 or 104 for a week. If nothing disconnects and the evening slowdown eases, you have found free spectrum.

When moving the router beats changing channels#

Channel work has a ceiling. You control which slice of spectrum you fight over, not how many people are in the fight. Signal strength is entirely yours, and it buys two things at once.

The first is margin: congestion works by squeezing your signal-to-noise ratio, and every dB of extra signal is a dB of noise you can now tolerate. The second is speed, which converts back into airtime. A laptop at -75 dBm on an 80 MHz channel may negotiate a link rate around 100 to 200 Mbps and deliver 40 to 60 Mbps of real throughput; at -55 dBm the same radio negotiates several hundred and delivers three to five times as much. That transfer now takes a fraction of the airtime, so a stronger signal shortens your own time in the queue.

Moving away from the party wall helps a third way: you hear less of your neighbors, and they hear less of you.

The moves that pay: off the floor, out of the media cabinet, away from the shared wall, toward the middle of the space you use. Where the router should actually sit covers the constraints, including a coax entry point that pins you to one corner.

What will not pay is an extender. A single-radio repeater retransmits every frame on the same channel, so it roughly doubles the airtime your household consumes in a band that is already oversubscribed. If congestion is the diagnosis, the shopping list is a flat Ethernet cable, a second access point on the end of it, or a 6 GHz-capable router — never a repeater. Our guide to which fix suits which layout covers the alternatives.

Testing router positions before you move anything#

Range Up does not scan for neighboring networks and does not measure noise — an analyzer app or your router's status page owns that half of the job. It owns the other half: how much signal margin a different router position would give you. Sketch the floor plan, set the material preset for the build, mark where the router sits, read the coverage room by room, then move the router to the shelf you were considering and compare.

Since fighting congestion is largely a matter of buying back SNR, that is worth knowing before you run cable clips along a rented wall. Try it on your layout and see whether twelve feet of repositioning beats a channel change.