The Wi-Fi works in the living room but not in the bedroom. What can I do?

The router sits in the living room, where the socket is. In the living room everything is fine. In the bedroom the video stops, and in the bathroom at the end of the corridor the phone says it is connected but nothing loads.

The first thing to understand is that you are not fighting the router. You are fighting the walls, and walls have a price that can almost be calculated.

What a wall costs

The ITU, which is the United Nations body for telecommunications, publishes the electrical properties of building materials and the formulas for deriving attenuation from them. It does not publish a table of decibels per wall, and nearly all the tables you find floating about are invented or copied from who knows where. The ones below I worked out myself from the coefficients in recommendation P.2040, and I give them with this warning: they are the absorption inside the material, they don’t include what is lost bouncing off the two faces of the wall, so in reality you lose more.

Going through a typical thickness, at 2.4 GHz and at 5 GHz:

Obstacle	2.4 GHz	5 GHz
Brick wall, 20 cm	around 4.5 dB	around 5.1 dB
Concrete wall, 20 cm	around 13 dB	around 23 dB
Plasterboard partition, 10 cm	around 1.9 dB	around 3.8 dB
Solid wood door, 4 cm	around 0.6 dB	around 1.2 dB

Look at the brick row and compare it with the concrete one. That is the point that contradicts the thing everybody repeats, namely that 5 GHz gets through walls far worse. On brick the difference is almost nothing. On concrete it is nearly double.

And there is a detail that is even more counterintuitive. In the recommendation describing propagation inside buildings, the coefficient saying how much the signal decays with distance in a residential setting is 28 at 2.4 GHz, and at 5.2 GHz it is 30 in flats and 28 in detached houses. Practically the same. So it isn’t true that 5 GHz fades faster as you move away: it starts lower, because a higher frequency starts out about 6 decibels down, and it loses more only when crossing certain materials. Put like that it sounds like pedantry, but it changes the advice: if all that stands between you and the router is light partitions and distance, 5 GHz can reach perfectly well.

Floors are another story. From the same source, in a flat the loss from crossing a concrete floor slab is around 10 dB at 2.4 GHz and 13 at 5.2. In a detached house, where the slabs are lighter, 5 and 7. A floor costs more than a wall.

Measure before you buy

This is the moment where you need a number instead of an impression, and the number takes two minutes to get.

Open How good is my connection? in the living room, next to the router, and take a measurement. Then go to the bedroom and do it again. Then the worst spot in the house. The tool keeps the history on your own device, so by the end of the walk you have three or four rows to compare, each with the time next to it.

What you are looking for is not the absolute value, it is the drop. If you have 200 in the living room and 15 in the bedroom, you have a coverage problem. If in the bedroom you have 80 but response time doubles and stability gets worse, the signal is arriving but it is being interfered with, which is a different problem and often gets fixed by changing channel.

The bars on your phone, on the other hand, are no use at all, and there is a precise technical reason: the same signal strength means different things depending on how much noise there is around it. A decent signal in a block of flats full of networks is worth less than the same signal in an isolated house. The bars show strength and ignore noise.

The four solutions, with their price

Moving the router

It’s free and it is nearly always the first thing to do. Manufacturers agree on three points: as central in the house as possible, because against an outside wall half the signal goes to warm up the street; out of closed furniture, because a wooden door in front of the aerials is one more wall; and away from appliances and large metal objects.

On height I have to flag an oddity. The TP-Link page recommends keeping it between thirty and forty-five centimetres off the floor, and a little further down, on the same page, it shows two measurements in which taking it up to around a metre eighty improves the signal by 18 and by 22 decibels. The two things don’t sit together, and since it is the same manufacturer giving both, I would follow the measurement rather than the recommendation: high up, in the clear, central.

Then there is something that applies to almost every Italian home and appears in no product list: the router sits where the socket is, not where it ought to be. Moving the socket, or running a network cable to a better spot, is often the highest result-for-money job there is. Nobody recommends it because there is nothing to sell.

The repeater

A repeater receives the router’s signal and retransmits it. It works, but it has a cost that needs saying: if it uses the same radio to receive and to retransmit, the bandwidth on that branch halves at every hop.

Here I have to be precise about the source, because the figure circulates everywhere and almost nobody says where it comes from. The technical document I found is from Cisco, it concerns professional mesh networks, and it gives a numerical example: starting from 24 Mbit/s you drop to around 14 at the first hop, 9 at the second, 4 at the third. That document is talking about the link between devices sharing the same radio for transmission and reception, not exactly about a domestic single-radio repeater that is also serving the phones. The mechanism is the same, but I have not found any standards document saying “a repeater halves the bandwidth”, and I prefer to say so rather than pass off a deduction as a citation.

The second condition is simpler and plenty of people get it wrong: the repeater goes where the router’s signal is still good. If you put it in the dead spot, it retransmits an already poor signal, and you end up with a network that reaches everywhere without working anywhere.

Verdict: fine for one single dead spot, for things that don’t matter much, in a house where you don’t walk about while you are on the phone.

The mesh network

Here the model changes. Instead of one device that repeats, you have several points forming a single network, under one name, coordinated with each other. The Wi-Fi Alliance’s EasyMesh certification defines exactly that: a controller that decides and nodes that carry out and report back, plus the ability to suggest that a device move to the most suitable point, band or channel.

Two things are worth knowing before you spend.

The first: mesh does not repeal the physics of the previous section. If the nodes talk to each other on the same radio that is serving your devices, you pay the same toll as with the repeater. What changes things is a link between the nodes that doesn’t steal airtime: a network cable, a radio dedicated to that alone, or the electrical wiring. A mesh node connected by cable is the best setup there is in a home, and that is worth knowing before you choose the model.

The second concerns a frequent disappointment. You have bought the mesh and your phone stays hooked to the far node even when there is one two metres away. Your setup is not at fault: in the standard governing these suggestions, the network can propose that the device move, but the final decision stays with the device. And the mechanism that makes the handover genuinely instant is declared optional in the certification. So nobody guarantees you perfect roaming: it also depends on the phone in your hand.

Powerline

It carries data through the electrical wiring. It is the answer when the obstacle cannot be beaten by radio, typically a concrete floor slab or a load-bearing wall, and you cannot run a cable.

There are a lot of limits and it pays to know them all beforehand, because they are the reason some people love it and others take it back to the shop.

The number on the box is not the speed you will get: it is the rate of the electrical link. TP-Link, which sells them, states that the real result is typically 20 or 30 per cent of that figure. So on a box claiming 1000, expect between 200 and 300 in the good case.

Then: it has to go straight into the wall socket, because extension strips and surge protectors filter the signal out. Phone and laptop chargers, in nearby sockets, introduce interference. And on the question of phases the two manufacturers I read contradict each other: TP-Link writes that the two adapters have to be on the same circuit, NETGEAR writes that they work across breakers and that on different phases you get a noticeable loss, or else they don’t talk to each other at all. The cautious formulation is this: the same circuit is the favourable case, different circuits may work worse, different phases may not work. You try it, and if it doesn’t work you take it back.

One last concrete trap: entry-level models have a 100 Mbit/s network port, so they won’t go above that figure whatever the electrical wiring manages.

A new router won’t carry the signal to the bedroom

This section exists because it is the most expensive mistake you can make from reading the advertising.

Wi-Fi 6, 6E and 7 are not coverage technologies. None of the documents I have read, neither from the Wi-Fi Alliance nor from the manufacturers, credits them with more range. One of those documents says the opposite explicitly: wider channels reduce range, because the same power is spread over more spectrum. Another, discussing the more aggressive modulation in Wi-Fi 7, writes that to reach it the device has to be a few steps from the access point, because it needs a signal-to-noise ratio of around 42 decibels.

What these generations really improve is the behaviour when the signal is there: more devices at once, less waiting, lower and more predictable latency, more resistance to the neighbours’ networks. In a crowded block of flats that is a decent gain. In the bedroom, where the signal doesn’t reach, it changes nothing.

On 6 GHz a European clarification is needed, because American numbers get repeated everywhere. In Europe the portion freed up for Wi-Fi runs from 5945 to 6425 MHz, that is 480 MHz, and home devices fall into the low-power indoor category, with a maximum of 23 dBm. In 480 MHz there is room for a single 320 MHz channel, whereas where the freed band is wider there is room for three. The upper part of the band, in Europe, won’t be decided before the 2027 world radiocommunication conference.

One last thing, which measures just how much of a compromise Wi-Fi is. When in Italy you have to measure the quality of your connection in a way that counts against your provider, the official procedure asks you to connect the computer to the modem with a cable and switch the Wi-Fi off. The regulator, having to measure the line in earnest, leaves it out: not because it is useless, but because what you would be measuring would no longer be the line.

To understand what sits underneath

If the problem is instead that the line is slow everywhere, and not just in one room, the checks in order are in Slow Wi-Fi: six checks to make before calling your provider. And if the symptom is video calls breaking up while the test says the speed is excellent, the reason is a different one and I have explained it in The video call keeps breaking up, but the test says the line is fast.