3G Internet at a distance from the base station using a dish and a coffee can

Hello to all residents of Habr!

The essence of this article is to show another option to catch the E / M wave in the relative outback for the needs of Internet access.
If somewhere exactly this design is already flashed, you are welcome to link, it will be interesting to compare and communicate.
I will set out the material in a narrative form, I hope, I don’t torture anyone with this, any terms, if anyone does not know, and other details can be safely omitted, this will not change the essence of the post.

Formulation of the problem

The Internet is useful and necessary, with this, I think, difficult to argue. And if there are no problems in the city with the Internet, then there are certain problems in the village, where I often come for family reasons. There is a private wooden house a la “hut” (location: RB), 3 operators provide access from the Internet in my neighborhood, but the maximum is only EDGE technology. The nearest two base stations (BS) are about 3.5 and 7 km. Between the house and the BS forest, but not close. EDGE would be enough, but loading these stations from other neighboring villages, judging by the speed, is such that one can “hide in potatoes”.
The option to buy a ready-made solution was dropped immediately for reasons of sports interest.
Having been inspired on one of the forums by the successful experience of the guys who used the satellite dish (also known as “offset dish antenna”) and the 3G modem itself in focus, I decided to simulate something like this and twist it.
I must say right away that I wanted to throw the antenna with the modem up to the attic, so as not to spoil the exterior of the house (more on this below).

Shopping, modeling, measuring

The first thing I was looking for was a bigger mirror. I agreed with the toad on the Ukrainian mirror “Kharkiv” with a diameter of about 85 cm, because for those that I saw more than 1m, the price tag already soars nonlinearly.
On the map of the coverage area of ​​the operators, I found that the nearest point with 3G is about 21 km. The relief to the BS - not the mountains, but the forest, again, enough.
')
The first experience was in the winter. There was no bracket, so I tried to expose the antenna as I had to. Approximately put on the alleged BS with 3G, but no matter how I twisted the mirror itself and did not focus the modem in the focus area, nifiga, sorry, it did not work out. Although the width of the radiation pattern (DN) (if simpler, the "beam width") at 2.1 GHz is quite wide, this is not digital satellite TV on the Ku band, where try again to "get" into this satellite.
The reasons are as follows. The DN in the modem is calculated as isotropic if possible, so part of the power goes not to the mirror, but to the opposite side, and the antenna itself on the modem's board leaves little to be desired for its ability to do something weak, this is logical. The BS and the antenna are separated by a wooden roof shield, which also gives a certain weakening. On this topic I even came across interesting literature, mainly on the subject of wood drying. Quite entertaining, but with a bunch of electrodynamic formulas characterizing the dielectric properties of the “sticks”. From the useful, in the first place, this is what the attenuation coefficient depends on the orientation of the fibers in the board with respect to the polarization of the wave. Previously, somehow did not think, although it is also logical.
But even here there was no luck: the boards in the shield are all vertically arranged - the attenuation maximum.
In general, there was a desire to mess around, but there was no opportunity (in the attic with a laptop on the battery for a long time) and there was no health (temperature -> to the street => hands).

A little bit upset, I decided to act according to the classical scheme: a homemade feed - in the focus of the mirror, from a modem to the feed - a coaxial cable.
As for the irradiator, there were thoughts about what to take: a kind of rectangular waveguide, or round. Stopped on a regular tin can. In the west, it is called "Cantenna". Nothing original, there is enough experience on the Internet according to this miracle. Tried, like some, just to push the modem into the old jar at the point where the probe is usually placed in the waveguide and thus irradiate the mirror, was also unsuccessful.

As a result, having determined the diameter of the jar with the condition of the propagation of one mode at its frequency (~ 2.1 GHz) in a circular waveguide, I went to the store with a ruler (!) For the jar. Before that, I had to dig a little in the internet, at which frequencies we have UMTS deployed. Taki found the necessary 10 cm in diameter. I bought two, for which I received verbal lyuley from my relatives, although I bought, sort of, for my own. Coffee was not cheap.

In the appendage, I had to buy a SMA connector for a nut so that the jar could be disconnected from the modem.
I modeled the antenna system and irradiator on the computer separately, went to shred the can. Here you need to be pretty neat with the size: characteristics are sensitive to misses.

The figure below shows a model of a jar in which a attracted copper tube (a fat cylinder tapering to the bottom), soldered at the SMA connector pin, is located at random.



That's what happened with modeling (CST).
Gain based on simulation results: 22 dBi.
The following is a general view of the entire offset antenna, which is familiar to many, + 3D NAM.





The characteristics of the antenna itself were not measured, only the CWS on the SMA connector directly on the bank. The graph curves S11 (reflection coefficient in dB). The green curve is an experiment (in theory, you can also use it at frequencies of 3.4, 4.5, 5.5 GHz ;-)).


Pleasantly surprised that it coincided quite accurately. For me, the SWR is good: all the power from the aperture of the antenna goes to the cable (and vice versa).

But then the saddest thing began. At first, I wanted to use the one I bought once in the Huawei E303 store, soldering directly to it with a cable, since modem without external connector.
Having found a rather small connector on the board (U.FL, most likely), I tried to solder the end of the RG174 cable in its place, since We did not even try to connect the cable connector in our retail trade. The idea was, frankly, idiotic - there is little experience, although I used a soldering station. Covered the RF path of the modem - it became a sort of USB-microSD adapter. In the photo is still a tselehonek (photo for reference to someone, maybe interesting, I was searching for an E303 in the internet, I did not find it).



So to him: how much did I get to him with the transfer to modem mode, and not the network card, so that I could see the signal level in numbers (it even came to use linux). Well, somehow the numbers turned out.
After a while, kind people donated a Huawei e173 with a connector, but with a broken USB plug. I am glad that now I’m soldering off the old connector and soldering it to the new one, I find that the contact area on the board on the signal pair is broken, and under the microscope I could see only the point of the transition jumper (via). "Lucky." But the world, again, is not without good people: MGTP’s “hair” was soldered onto this jumper (surprisingly, the solder grabbed) and the connector was soldered. The modem is working. Happiness was full pants. Now I needed a CRC-9 connector. After a trip to the radio market in Minsk, the only thing I found was only the cable “CRC-9 - FME” (the same as in the photo below).
I had to rewire FME to SMA.

Testing

Finally, in the summer it is time to test what happened. In the garden, under the apple trees.
On the left on the photo is blue-red - a can of coffee, on it a modem is taped with adhesive tape, from the connector on the modem to the connector on the bank the above cable. On the right in the photo is another perspective. Over the mound of a solid forest.
Having turned the antenna, I found two points. In one, the signal is worse (the result of the MDMA program), at -103 dBm (the forest is closer, or the channel is worse, or the BS antenna shines weaker in my direction, did not understand), the second one has a better -93 dBm. I note that this is all from ground level! On the second photo you can see the direction to the second BS (a distance of a little more than 20 km, according to the non-monitoritel, for which a special thank you to them).


Speed ​​test from under the tree


By the way, it is very funny to talk on Skype in video mode, sitting under an apple tree.

It remains to pile it all in the attic. But it remains to solve another problem. We did not want to lead to the attic of 220V in order to install a router for powering up the modem. Since USB cables are not cheap (I needed> 7 m), I had to solder a USB extension cable using a twisted pair (I had a chance to buy SSTP 6 cat, I was terribly pleased, and used it for greater reliability, although it is not cheap either). For information on how to make a USB extension cable from a twisted pair, there is enough information on the network. Having drilled a hole in the ceiling, he lowered the cable into the house. Well, there as fantasy will allow, even Wi-Fi.
The signal in the attic floats from -87dBm to -93dBm. But this is uncritical.
On the left is a speed test in the attic of the entire system. And on the right, what the modem gave by itself in the attic of nearby BS (2G) (even ICQ slowed down).

Photos from the attic. On top of another antenna - DMV (DVB-T). The part of the photo, where the window is, is the shield mentioned above.



And the photo banks are a little closer. Anyone got a Nobel Prize for scotch, or not ?! If not, then this is a terrible mess!



I hope that information will be useful to someone for reflection and will bring civilization closer to your God-forgotten corner. Please constructive criticism in the studio. If you have questions, I will gladly answer.