35

I'm thinking of buying an antenna for a router so I can extend my Wi-Fi. I have seen a few products like TP-Link TL-ANT2405C Indoor Desktop Omni-directional Antenna:

Enter image description here

The length of the wire is 130 cm. Is it OK if I increase the length of the wire or will it affect the performance of the router? What is the maximum cm of length that I can increase?

  • 9
    Surely this depends on the quality of the coaxial cable you purchase? – Burgi Jun 2 '16 at 21:42
  • 3
    Signal loss occurs along every foot of coaxial cable. The amount of loss is dependent upon the frequency of the signal, the length of the coaxial cable run and the quality of the cable used. Different types of Coaxial cable have different attenuation per foot of run so it depends on which type of cable you use. – DavidPostill Jun 2 '16 at 21:59
  • 2
    You're asking wrong question. The real one is: how to extend WiFi coverage? And the answer is "get a repeater". – Agent_L Jun 3 '16 at 15:18
  • 3
    This antenna is not meant for base stations. It's for desktop PCs, so you can use a WiFi card in the back, but move the antenna to the front. Basically to stop the PC from blocking itself. Routers are small enough that you can move the whole thing to best location. – Agent_L Jun 3 '16 at 15:21
  • 5
    @Agent_L At least in my experience, most WiFi repeaters are pretty terrible. You're way better off using additional wired access points, often that's so even if you have to wire them through the powerline. But I definitely agree about positioning routers or access points -- try to put them where you want them, extending the Ethernet to them if needed. – David Schwartz Jun 3 '16 at 19:37
48

There is no arbitrary limit. Any increase in cable length will reduce signal strength. (So will the connectors that you'll need to connect another length of cable to this one.) As Burgi and DavidPosthill said in the comments, how much it's reduced for a given length depends on the cable and the frequency.

A common relatively inexpensive cable for short runs of WiFi antennas is LMR100. At 2.4 GHz (the common WiFi band), 15 feet of LMR100 will result in signal loss of about 6 dB. That's equivalent to dropping power to just 25% of what it was. (Each 3 dB = 50% gain or loss in power)

With LMR400 cable, your loss would be only about 1 dB! (But that cable is more expensive, and also a lot less flexible = more difficult to install.)

The loss in dB is linear with the cable length. So if you use 30 feet of LMR100 cable, your loss will be 12 dB (i.e. your signal is now about 1/16 the power that it was). With 7.5 feet, loss will be only 3 dB (you lose half of your signal).

All of the above numbers are for the 2.4 GHz WiFi band. For 5 Ghz it will be much worse.

Don't even think about using RG59 (the older, thinner coax that used to be used for TV cable and antennas and commonly is seen with "F" or BNC connectors attached; it's not even the right impedance) or RG58 (the right impedance, but still very lossy at these frequencies). These cable types aren't rated at all for use above 1 GHz.

You can find data sheets (with signal loss graphs) and calculators for various types of microwave coax all over the web. Here's a calculator (found at a cable dealer) that covers a wide variety of table types.

And to convert dB to ratios (or back), try this (Note, since this is signal loss we're talking about, enter the dB number as a negative number before pressing "calculate". Note also that you want the power ratio, not voltage.)

One last tip: Don't try to assemble cables yourself. Buy cables with the right connectors already attached. Very minor-seeming mistakes with connector assembly can cause huge losses at these frequencies. And absolutely do not cut the connectors off and try to splice the coax. Might as well throw the antenna away at that point.

7

The short answer is:
Ideally you would have no cable. Every bit degrades the signal.
You get about 10 metres of cable before you need some kind of amplifier/booster/repeater, otherwise your signals about as good as dead.

  • 7
    Although this may sound silly at first glance ("how can you have no cable?!") it's entirely possible to do. You get a WiFi access point or that can survive outside and to which your preferred antenna can be attached directly... or with a one meter or so cable at most. Then you run Ethernet and power - use PoE, power over Ethernet, if you can - down the pole. The distance limits for twisted pair Ethernet are far longer than those for the WiFi signals. The TP cable will likely be a lot cheaper than low-loss coax too. – Jamie Hanrahan Jun 3 '16 at 1:24
  • 4
    -1 No references or explanation. – tymtam Jun 3 '16 at 3:27
  • @JamieHanrahan All I meant was; any signal will gradually degrade as it traverses (almost) any medium. The same thing happens to electricity as it traverses a battery cable; the voltage gradually drops, the further it travels. Every additional component provides another opportunity for signal loss. Ideally you would solder the antenna directly to the circuit board. (Or in this case screw the antenna onto the RP-SMC connector.) Using a cable has obvious benefits, it just happens to do so at the expense of signal strength and fidelity. It's a compromise; there's a trade-off. – voices Jun 3 '16 at 4:20
  • @Tymski I don't have any references or explanation. This isn't something that I read online, or in a book somewhere. It's from my actual life; my own observation and personal experience. – voices Jun 3 '16 at 4:50
  • 3
    Well, I do quibble with your "10 metres" as if that were a constant. – Jamie Hanrahan Jun 3 '16 at 6:34
4

When you're trying to get actual practical information about the product, first step is to take a look at the product datasheet. So on TP-LINK's downloads website for linked product, you'll be able to download the datasheet. One you do, you'll be able ti find the exact cable type used there. In this case, it's RG-174.

Unfortunately, they did not specify exact manufacturer of the cable, so we need to go and look for some generic answers. Usually, losses are specified either per meter or per 100 feet and are frequency dependent, so let's see if we can try that for this cable.

Go to Google, type RG-174 coaxial cable losses 2.4 GHz. One of the results is this handy table. Looking there, you can find that RG-174 has around 1.7 dB attenuation per 1 meter, giving around 2.2 dB loss for the 1.3 m piece that comes with the antenna. Do keep in mind that the numbers are not for the specific cable you have, but for category, so don't be surprised if you find slightly different values elsewhere.

  • That's interesting. The cable calculator I linked in my answer ( timesmicrowave.com/calculator/… ) says that RG174 isn't even rated for anything above 1500 MHz. But at 1500 Mhz it says that 15 feet would loss of 5.1 dB. Loss increases with frequency, so... looks like TP-LINK didn't make the best choice there. – Jamie Hanrahan Jun 3 '16 at 19:27
  • 1
    @Jamie Hanrahan Of course it's not a good choice when losses are the main criterion, but when judging the best choice, entire system design needs to be taken into account. With shipped 1.3 m (or let's say around 4.3 feet for the calculator), the losses are around 2 dB. Antenna has gain of 5 dBi. If we replace a say 0 dBi stock antenna with this, we'll still have a bit of improvement. Even more can be obtained if we manage to say place the antenna at a location where it won't be in a fade. – AndrejaKo Jun 3 '16 at 19:53
  • Extra gain can also come from proper separation of antennas inside of an array used in modern standards. Quite often, home routers are too small to have good enough separation, so spatial multiplexing and spatial redundancy don't really work as good as they could. With proper spacing, we could maybe get one or two decibells of improvement. Then, there's the bottom line: Would the average consumer be able to recognize that a more expensive antenna is more expensive due to reasonable quality cable and would this justify increased price? Given how many people use WiFi, I'd say not really. – AndrejaKo Jun 3 '16 at 19:55
  • @AndrejaKo Note that 0 dBi is virtually unattainable. Every real-life antenna has a non-isotropic radiation pattern because no real-life antenna is a point radiation source, and every real-life antenna system has some degree of resistive loss, if nowhere else then in the connectors (insertion loss) and feedline (resistive loss), the latter being basically what this question is asking about. (On lower frequencies not relevant to WiFi, you also have to consider potential capacitive and reactive losses. See shortened HF antennas. On non-matched loads, also losses due to reflections on the cable.) – a CVn Jun 4 '16 at 14:59
  • @Michael Kjörling I agree on the reality part! When I wrote 0 dBi, I was mostly thinking about average antenna gain. In such case, it's not actually all that uncommon to see antennas with 0 dBi or even lower average gain. See for example this antenna with 0.7 dBi average gain (this is actually a device SMD antenna). I was also oversimplifying things by thinking of router as a device with no internal feed-line losses, while in reality we might have 10 cm or so feed-line on the inside as well. – AndrejaKo Jun 4 '16 at 15:10
2

It depends very much on the cable type you buy. You need to buy a cable that is designed for the frequency range. Once you have found a suitable range of cables you need to chose the physical size, smaller cables will be easier to manage but lossier.

Personally I find LBC240 to be a reasonable compromise. It has a specified loss of 0.42 dB per meter at 2.5 Ghz which means that with 7m of cable you will lose about half your signal power. So a few meters of it is not too bad for your signal strength, and it's small enough to be reasonablly managable.

I would suggest getting cables made professionally, a badly crimed connector can lead to reliability problems and/or high signal loss.

In general I would try and locate the AP as close to the antenna as reasonablly practical but if a few meters of antenna cable are the difference between a crappy antenna site and a good antenna site then put the antenna in the good site and accept the cable losses. Losing half your signal in a cable sounds like a lot but it's nothing compared to the losses a bad antenna site can bring.

2

What is missing is the VSWR conditions in the setup. If the antenna/cable/receiver input is not a match then unacceptable losses mount. As a user, controlling these conditions is beyond most folks out there but may in fact, be the culprit in puzzling situations. The best I can offer is , if you must maintain maximum performance with minimum losses, consultation with professionals in RF is highly recommended.

http://www.antenna-theory.com/definitions/vswr.php

2

A much better idea is to move the router to a much higher location by extending the hard wiring and power then you still get maximum power to the aerial and little noticeable loss on ethernet cable

0

Another short answer: if the cable gets the antenna to a better place, it is worth every dB of signal loss.

Search for "Coax Line Loss Calculator" if you want a numerical answer: but trust me you don't need it. Anything under a few meters is going to be just fine.

Position the antenna, find the best spot, use the shortest cable that reaches. You can take a long cable and cut it down, using a crimper. But it's not worth the effort. Best to look at your router for a signal strength indicator, and at your PC end for the same. Move the antenna. Optimize both numbers.

  • Define "short". – Afshin Mehrabani Jun 3 '16 at 17:27
  • That depends on how bad the original location was. If the antenna from behind a "tower" PC that's shoved into a corner, yes, you're probably right... but then you don't need the 15 feet of cable this thing comes with, let alone more. – Jamie Hanrahan Jun 3 '16 at 19:32
  • @JamieHanrahan umm the antenna comes with a 130cm cable according to the question and the ebay listing. – plugwash Jun 4 '16 at 2:30
  • You're right. I'm not sure where "15 feet" came from! – Jamie Hanrahan Jun 4 '16 at 5:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.