STEP 1Set up so you test only the wireless hop

Run the iperf3 server on a wired machine (connected to your router by Ethernet) and the client on the Wi-Fi device. That way the only variable is the wireless link itself.

The server must run on a computer — a PC or Mac, ideally on a wired Ethernet connection (running it over Wi-Fi would muddy the test). The client — the thing you test from — can be that computer, or simply a phone or tablet app (see Step 2).

Need that machine's IP? You can read it from your router's device list or your computer's network settings — but since the terminal's already open:

Use the address that starts 192.168… or 10… — that's what the client points at in the next step. (On a Mac connected by cable, en0 may return nothing; ifconfig | grep "inet " lists every interface's address.)

When you're done, stop the server with Ctrl-C in its terminal. iperf3 -s stays up and keeps accepting tests until you stop it, so you can run as many as you like first.

STEP 2Test from the Wi-Fi device

Use -O 2 to omit the first two seconds while TCP ramps up, and run both directions. Test in the room where you actually use the device.

Prefer your phone? Free iperf apps exist for iOS and Android (search "iperf") — they act as the client, so you can point one at your server's IP from Step 1 instead of typing commands. Handy, since the phone is the device actually sitting in the problem room. They're third-party apps and quality varies, so treat the numbers as indicative.

STEP 3Check stability with a UDP test

Wi-Fi problems often show up as jitter and packet loss rather than low average speed, and a UDP test reveals both. This is a stability check, not a speed test, so the rate is deliberately low — 10 Mbps that any link carries comfortably. That way the jitter and loss you see reflect link quality, not you overflowing it: UDP sends at whatever fixed rate you set whether the link can keep up or not, so flooding it would just manufacture loss and tell you nothing.

Loss above ~1% or jitter above a few ms at modest rates points to interference or weak signal.

STEP 4Interpret & improve

If the numbers disappoint: move the router off the floor and away from metal, prefer the 5 GHz band, change channel away from neighbours (Steps 5–6), and repeat the test room by room to map your coverage. If a far room stays poor, a wired access point beats any antenna tweak.

STEP 5Scan the airwaves around you

Poor Wi-Fi is often not your router's fault — it's your neighbours' routers shouting on the same channel. A free scanner app shows every network in range, its channel, and its signal strength, so you can see exactly where the congestion is.

Walk to the room with the worst performance and scan from there — congestion is about what that device hears, not what the router hears. Note which channels your strongest neighbours sit on.

STEP 6Adjust your channel if needed

Log in to your router's admin page (the address and password are usually on a sticker underneath) and find the wireless settings for each band.

After changing the channel, give devices a minute to reconnect, then re-run the iperf3 test from Step 2 in the same room. If throughput and jitter improved, the channel was the problem. If nothing changed, the issue is signal strength or building materials — go back to placement, or run a cable to a wired access point.

BANDS2.4 vs 5 GHz — which to use

It's a range-versus-speed trade. 2.4 GHz travels further and pushes through walls and floors better, but it's slower and crowded (only three non-overlapping channels, shared with Bluetooth, microwaves and every neighbour). 5 GHz is much faster with far more clear channels, but shorter range and more easily blocked by walls. So use 5 GHz for anything near the router or bandwidth-hungry (laptops, TVs, gaming, video calls), and 2.4 GHz for distant rooms and low-demand IoT gear where reach matters more than speed. Same network name on both bands lets devices choose the band automatically; separate names let you pin a device to one. Keep the names the same if you want it simple — but if something that should be on 5 GHz keeps clinging to the slower 2.4 GHz band (a smart TV near the router is the classic case), give the two bands separate names so you can connect that device straight to the 5 GHz one. This suits gear that stays put; for phones and laptops that move around, the shared name is usually better, so they can fall back to 2.4 GHz's longer reach when you're far from the router.

And for a device that stays put and streams — a smart TV especially — a wired connection usually beats Wi-Fi even when the cable is the slower link on paper. Many TVs ship with only a 100 Mbps Ethernet port, so a 5 GHz speed test may post a higher number — but streaming needs steady delivery, not peak speed: most commercial 4K streams need only 15–25 Mbps, depending on the service. The cable's far lower jitter and near-zero dropouts make it the better choice for a TV, headline figure aside.

POWERDon't just turn the power up

Raising the router's transmit power is tempting but usually backfires. Wi-Fi is a two-way conversation: a louder router still has to hear a phone or laptop replying on its much smaller antenna. Crank the router and you get a lopsided link — the device hears the router fine, but the router can't hear the device, so you get retries and dropouts, not more range (and there are legal power limits anyway, which vary by country). In multi-access-point setups you often want lower power, so devices hand off to the nearest one. The real fixes for coverage are placement, the right band, or another access point — not more power. Most consumer routers don't even expose the setting.