Why Foil Behind a Router Helps One Band and Nulls the Other

This page contains affiliate links. If you buy through one we may earn a commission, at no cost to you.

It works, within a limit that is both smaller and stranger than people expect. A flat conducting sheet behind an antenna can add at most 6.02 decibels in the forward direction, and only if it stands exactly a quarter of a wavelength off the antenna. That is 30.7 millimetres at 2.4 gigahertz and 14.5 at 5. Those two numbers are irreconcilable, and worse than irreconcilable: the spacing that gives 2.4 gigahertz its full 6 decibels puts 5 gigahertz on the wrong side of a null and takes 8 away.

So on a dual band router, which is every router sold for the last decade, the foil is not a free improvement. It is a trade, and it trades in the wrong direction, because 5 gigahertz already starts 7 decibels behind on path loss.

The sheet does not reflect, it makes a copy of your antenna

The useful way to think about a conductor behind an antenna is not reflection but imaging. A perfectly conducting plane makes the field in front of it identical to what you would get from the real antenna plus a mirror copy of it, sitting the same distance behind the plane, with its sign reversed. Everything follows from adding those two together.

The forward field is one plus a copy that has travelled an extra distance of twice the spacing and had its sign flipped. Work the phase and the magnitude comes out as twice the absolute sine of two pi times the spacing divided by the wavelength, so the gain over having no sheet there at all is twenty times the base ten logarithm of that.

Two sanity checks fall straight out and both are useful. At a quarter wavelength the sine is one, the two add perfectly and you get 20 log of 2, which is 6.02 decibels. And at zero spacing the sine is zero, which is a complete null: a sheet of foil taped flat against an antenna does not reflect its signal forward, it shorts the antenna out. That second result is the one that catches people, because taping the foil on is the obvious way to do it and it is the one spacing guaranteed not to work.

So the whole method lives or dies on one measurement, and the measurement is in millimetres. Channel 6 sits at 2,437 megahertz, so its wavelength is the speed of light over that, which is 123.0 millimetres, and the quarter wave is 30.7. Channel 36 sits at 5,180 megahertz, wavelength 57.88 millimetres, quarter wave 14.5.

Why the two bands cannot be served together

Set the sheet at 30.8 millimetres for 2.4 gigahertz and work out what 5 gigahertz sees. Two pi times 30.8 over 57.88 is 3.344 radians, whose sine is minus 0.200, so the gain is 20 log of 0.400, which is minus 7.96 decibels. Higher up the band at 5,500 megahertz it recovers to minus 2.00.

And there is no escaping it by picking a different spacing, because the geometry is periodic and the two bands land in opposite halves of the cycle. For a 30.8 millimetre sheet the exact null falls where the wavelength is 61.6 millimetres, which is 4,867 megahertz, immediately below the bottom of the 5 gigahertz band. The next constructive peak is where the wavelength is 41.1 millimetres, which is 7,300 megahertz, above the top of the 6 gigahertz band. The entire 5 gigahertz allocation sits between a null and the next peak, which is the destructive half.

Which matters more than it sounds, because of the order the two bands start in. Free space loss goes as the square of frequency for fixed antennas, so 20 log of 5,500 over 2,437 is 7.08 decibels of penalty that 5 gigahertz carries before any wall or any reflector is involved. The band that needs help is the one the foil hurts, and the band the foil helps is the one that was already reaching further.

The one case in our catalogue where it is unambiguously right

There is a category of device for which all of the above collapses into a clean win, and it is worth naming because it is the honest answer to the question rather than a dismissal of it. Single band devices.

A ZUMIMALL 2K solar security camera with 360 degree coverage connects, as its record states, "over 2.4GHz WiFi through a built-in dual antenna", and its own drawbacks are explicit: it "Works only on 2.4GHz WiFi, not 5G". There is no second band to sacrifice. It is wire free and solar powered, so it is placed where the view is rather than where the signal is, which is exactly the situation where a few decibels decides whether it holds a connection at all.

So for a camera at the far corner of a garden, a sheet of foil or a biscuit tin lid on the far side of it from the house, standing 30 to 31 millimetres off the antenna, is a genuinely correct intervention with no trade against anything. Note which end to put it on: antenna gain is reciprocal, so a reflector helps a link in both directions equally whether you fit it at the router or at the camera, and the camera is the end that is usually further from everything and easier to get at. Our ZUMIMALL solar camera review has the detection features, which matter more day to day.

And one honest caveat on that record: it publishes a "built-in dual antenna" and no gain figure for it, so the starting point you are adding 6 decibels to is unknown. That is the norm in this category rather than a failing of this listing. Across our whole catalogue, antenna gain appears only as bare spec-table numbers on a couple of USB and PCIe adapters, where our USB WiFi adapter coverage records twin 5 dBi antennas on one dongle and a 6 dBi omnidirectional on another, with no reference point for either.

What the foil cannot touch, and it is usually the thing stopping you

Three limits, in order of how often they are the real problem.

First, distance beats gain and it is not close. Indoors, path loss goes as distance cubed or worse, because every wall and floor adds its own loss on top of the geometric spreading. With an exponent of 3, six decibels of gain extends your reach by ten to the power 6 over 30, which is 1.58, and at an exponent of 4 it is 1.41. Against that, halving the distance to the router is worth ten times the exponent times the log of two, which is 9.03 decibels at an exponent of 3 and 12.04 at 4. Moving the router to the middle of the house beats the best possible foil, by a margin, and it takes a minute.

Second, a single wall can eat the whole gain. Plasterboard runs around 3 decibels at 2.4 gigahertz, interior brick or blockwork 6 to 12, and reinforced concrete more, with 5 gigahertz worse again on all three. So arranging the geometry so the signal crosses one fewer masonry wall is worth more than any reflector, and that is a question of where the router sits rather than what is behind it.

Third, and this is the one nobody mentions: gain is reciprocal, so a reflector increases the interference arriving from the direction it is aimed at by exactly as much as it increases your own signal. If what is limiting you is a neighbour's access point on the same channel in roughly the same direction as your own device, the foil raises signal and interference together and your signal to interference ratio does not move at all. A directional antenna only helps congestion when the interference is coming from a direction you are not aiming at.

The device that solves the problem by moving instead

A SIMO Solis Edge 5G WiFi hotspot is in this article because of one feature that is the correct engineering answer to the whole question. Its record describes SignalScan, which "automatically scans and connects to the strongest mobile signal in the area", and names "RVs, campers, motorhomes and road trips" as the use case.

That is choosing the best available path rather than trying to beat a bad one, and it is worth far more decibels than any passive gain. The record also makes the sensible observation that "the thing you want is not raw speed but something that keeps finding a usable network without you intervening". For a 2.4 gigahertz device at the edge of a house, the equivalent move is to change channel and change position before touching the antenna.

Its own record is unusually straight about its gaps, and two are relevant here. It publishes two conflicting weights, "2.8 ounces and 0.68 pounds", which the record flags rather than picking the flattering one. And it publishes no antenna specification at all, which means the WiFi 6 radio inside it is as unquantified as everything else in the category. Our SIMO Solis Edge review has the data allowance arithmetic, which matters more than any of this, and our portable hotspot coverage has the category view.

The same geometry at a tenth of the frequency

One contrast makes the millimetre figures concrete. FRS radios transmit around 462.6 megahertz, so their wavelength is the speed of light over that, which is 648 millimetres. A quarter of that is 162 millimetres. A reflector for a walkie talkie has to stand a hand's width off the antenna, where one for 2.4 gigahertz stands roughly the width of two credit cards off it.

Which is why nobody builds reflectors for handhelds and everybody builds them for microwave links: at 2.4 gigahertz the hardware is pocket sized, and at 462 megahertz it is furniture. The physics is identical and the only thing that changed is the one number the geometry is measured in.

A Midland LXT630VP3 FRS walkie talkie three pack also happens to carry the best written range paragraph in our catalogue, and it deserves quoting because it is the same class of claim as a foil mod: Midland quotes up to 30 miles, and the listing itself qualifies it: longer range communication in open areas with little or no obstruction. FRS is line of sight. Hills, buildings and trees all cut it, and in ordinary terrain real usable range is a small fraction of the headline. Its own drawbacks keep that honest, stating that "The 30 mile range applies only to open, unobstructed ground". That is the right way to publish a propagation figure, and it is the lens to read any gain claim through. Our Midland LXT630VP3 review covers the weather alerting, which is its real feature.

The honest limit, and the free technique that settles it in two minutes

The limit is that the 6.02 decibels is a ceiling for a perfect conductor of infinite extent, and your biscuit tin lid is neither. A finite sheet lets energy diffract around its edges, so the real gain from a hand-sized piece of foil at 2.4 gigahertz, where the wavelength is 123 millimetres and the sheet is only a couple of wavelengths across, is meaningfully less than the ideal. Three to five decibels is a fair expectation and the drawing's 6 is an upper bound. The spacing sensitivity is real though, and it is the reason most attempts produce nothing: at 10 millimetres instead of 30.7 the gain is essentially zero, and at zero millimetres it is negative.

Also honest: the band argument above assumes both radios share the same antennas or sit close enough together that one sheet covers both, which is true of most consumer routers and not of all of them. If your router has separate external antennas per band you could in principle give each one its own spacing, which is the only configuration in which the trade goes away.

And then the free technique, which is better than every purchase mentioned here and which nobody does. Measure it. Every phone can report received signal strength in dBm, on Android through the WiFi details or a developer screen and on an iPhone through Apple's own AirPort Utility with its scanner enabled. Stand in the spot that is giving you trouble, write down the number, make your change, wait ten seconds and write it down again.

That turns the entire question from an argument into an experiment with a unit. If the foil is worth 4 decibels you will see 4 decibels. If it is worth nothing you will see nothing, and you will have learned that in two minutes rather than believing something for a year. And while you are holding the meter, try the two things the arithmetic says are larger: move the router two metres toward the middle of the house, and switch your 2.4 gigahertz radio between channels 1, 6 and 11. Those are the only three channels in the band that do not overlap each other, so they are the only three worth trying, and one of them is usually several decibels of interference quieter than the one you are on.