Wi-Fi Dead Zones: How to Find and Fix Them With Network Testing

You have one room where the Wi-Fi is bad. Maybe it’s the back bedroom. Maybe it’s the garage. Maybe it’s the corner of the office where the printer lives. Every time you walk into that room, your phone says “two bars” and YouTube buffers. You’ve moved the router twice. You’ve bought a mesh extender. The dead zone is still there.

The frustrating truth: most attempts to fix Wi-Fi dead zones fail because the diagnosis is wrong. “Low signal” is a symptom, not a cause. You can have full signal bars in a room that delivers 30 Mbps, and you can have a faint signal in a room that delivers 500. To actually fix a dead zone, you need to measure throughput, not signal — and you need to test from the device that’s having the problem, not the laptop you happened to bring.

This post is the structured approach: how to map the dead zone with real numbers, identify whether it’s coverage, interference, or device-side, and pick the right fix.

Why “signal bars” lie

The bars on your phone are a coarse summary of RSSI — received signal strength indicator. Two problems with using them as a diagnosis:

1. They don’t measure noise. A room with strong signal and strong interference from a neighbour’s network on the same channel will report “full bars” and deliver terrible throughput. RSSI alone doesn’t tell you about competing traffic.
2. They’re measured differently by each phone. An iPhone’s “two bars” and a Samsung’s “two bars” are not the same dBm. Comparing across devices is meaningless.

Throughput, by contrast, is an objective measurement of how much data the link actually delivers. It captures coverage, interference, congestion, and band selection in a single number. That’s what you want.

The mapping pass

The investigation is structured: walk the property with the phone, run an iperf3 test in each room, save the result with the room name. After 10–15 minutes, you have a coverage map you can act on.

Setup:

1. Pick a server. A wired Mac, a Linux box, a NAS, or another phone in a known-good location. Plug it into Ethernet. Open the app in Server mode. Note the IP and port (5201 default).
2. Walk to the first room. Open the phone app in Client mode. Enter the server IP. Pick TCP. Tap Start.
3. Save the session with a meaningful name: “Lounge”, “Kitchen”, “Back bedroom”, “Garage”.
4. Repeat for every room the customer or you care about. Don’t skip the “fine” rooms — they’re your baseline.

By the end of the pass you have a table of rooms and throughput numbers. The pattern usually tells you everything.

Reading the map

What the numbers mean:

• Close to the AP, full speed. 600+ Mbps on Wi-Fi 5, 1+ Gbps on Wi-Fi 6/6E. If the room next to the AP is fine, the AP itself is fine.
• Distant room, low speed. 50–150 Mbps. Classic coverage problem — the signal is reaching the room but the link rate has dropped because of distance and walls.
• Distant room, near-zero speed but signal present. 5–20 Mbps. The phone associated with the AP but can barely communicate. Almost always interference or band-fallback to 2.4 GHz in a noisy environment.
• Distant room, no connection. Real dead zone. Phone associated, but the link drops or times out.
• Close to the AP, low speed. Unusual. Usually means CPU-bound router, a Wi-Fi chip on a single client device that doesn’t negotiate well, or aggressive throttling somewhere upstream.

The pattern that pops out — “fine everywhere except this one corner” — tells you the dead zone is geometric, not systemic. The opposite — “low everywhere, even right next to the router” — means the problem isn’t location, it’s the AP or the upstream link.

Distinguishing coverage from interference

A coverage problem and an interference problem look superficially similar — both manifest as “throughput drops as you walk away from the AP.” Two ways to tell them apart:

1. Time variability. Run the same test in the same room three times, spaced an hour apart. Coverage is steady — distance and walls don’t change. Interference fluctuates — a neighbour’s 4K stream comes and goes. If the spread between runs is more than ±30%, it’s interference.
2. 2.4 vs 5 vs 6 GHz. Run the test on each band separately (with per-band SSIDs as above). Pure coverage problems show 6 GHz dropping fastest, 5 GHz next, 2.4 GHz lasting longest. Interference problems show the opposite — 2.4 GHz crashes from neighbour networks, while 5 GHz and 6 GHz are clean.

The fix differs:

• Coverage: add a mesh node, move the AP, or run an Ethernet drop to a new AP closer to the dead zone.
• Interference: change channel, add a 6 GHz radio, replace nearby cordless equipment, or stop running the microwave during video calls (yes, this is real on 2.4 GHz).

Common dead-zone causes you can rule out fast

While mapping, look for these specifically:

• Brick or stone interior walls. Eat signal aggressively, especially at 5 and 6 GHz. A dead zone behind one stone wall is geometry, not equipment.
• Foil-backed insulation. A surprising number of newer homes have it. It’s effectively a Faraday cage. Mesh helps; moving the AP doesn’t.
• Concrete floor between AP and dead zone. Multi-story homes have this constantly. Wi-Fi attenuates faster going up/down than going sideways through wood walls.
• Mirrored or metallized glass. Smart mirrors, security glass, and some commercial windows reflect Wi-Fi.
• AC units, refrigerators, large metal appliances. Block signal. If the AP is on top of the fridge, the room on the other side is going to suffer.

If the dead zone correlates with one of these, the fix is almost always a second AP or a mesh node on the other side of the obstacle — not a different router on the same side.

Picking the right fix

Based on the map, the decision tree is:

• One bad room with otherwise clean coverage: add a single mesh satellite, or run Ethernet to a wall-plate AP in the bad room. Mesh is faster to set up; wired AP is more stable long-term.
• Whole floor underperforms: AP is in the wrong room. Move it to the geometric centre of the floor. If you can’t, add a second AP.
• Throughput is fine but specific bands drop: interference. Change channels, replace older cordless devices, or move to 6 GHz where allowed.
• Everything is low including right next to the AP: the AP itself is the bottleneck. Old hardware, CPU-bound, or running outdated firmware.

After the fix, redo the mapping pass. Save the new sessions with names like “Back bedroom after mesh”. The app’s history view shows old and new side by side. If you can’t see a clear delta, the fix didn’t work — try the next option, don’t assume.

The mesh-versus-wired-AP question

This comes up so often it deserves a paragraph. Mesh adds a satellite that backhauls to the main node, usually over Wi-Fi itself, sometimes over Ethernet. Wired AP adds an access point connected by cable to the router. Both can deliver good coverage. The numbers say:

• Mesh with wireless backhaul: half the available throughput goes to the backhaul. A 1 Gbps Wi-Fi 6 mesh delivers ~400–500 Mbps to clients on the satellite. Fine for streaming and browsing, mediocre for large file transfers.
• Mesh with wired backhaul: the satellite gets near-full speed because the backhaul is cable. Approaching wired-AP performance.
• Wired AP: the satellite is just another AP on the LAN. Full speed at the client, no backhaul tax.

If you have or can run Ethernet to the dead zone, a wired AP almost always beats mesh. If not, mesh with a satellite placed at the boundary of the dead zone (not deep inside it — at the edge) is the right call.

When the dead zone isn’t really a dead zone

A few last things to check before adding hardware:

• Is the client device the problem? Old phones with 2x2 Wi-Fi 5 chipsets cap out around 400 Mbps regardless of the AP. Test with two different devices in the same room. If one is fine and one isn’t, replace the device, not the network.
• Is the dead zone time-of-day correlated? Some “dead zones” are really congestion during peak hours from neighbour networks. The same room is fine at 03:00 and terrible at 20:00. Solve that with band/channel selection, not coverage hardware.
• Is anything in that room metallic and new? If the dead zone appeared after a recent change — new monitor, new file cabinet, new appliance — try moving the new object before adding APs.

Bottom line

A Wi-Fi dead zone is a measurable, fixable problem. The only requirement is real data, not vibes. Walk the property, save sessions per room, look at the pattern. The right fix becomes obvious — usually a wired AP, sometimes a mesh node, occasionally just a channel change. The wrong fix — buying more hardware without measuring first — is what most people do, and it’s why dead zones persist.