The page of WiFi explanation (FAQ WiFi, tuto wifi, tutoriel wifi, blog wifi, file wifi, synthesis wifi, documentation wifi, guide wifi, ebook wifi, ibook wifi, wifi rural) of Yves MAGUER

Last setting with day: 09/09/2007
Creation date: 28/01/2003
Where is this page:
Translation is possible for others languages

I make studies of bonds (grid possible) or of covers WiFi or WiMax. Voir mes CV.
My rural partner: Numevia
My city partner (hotspots): Coffeespot

To consult me: (remove _POURIEL of the email).

To go directly to synopsis.

If one day, you succeeded in reading all this page (cheer!), capitalize your effort, and read then only them differences!

Software exists:
HTMLDiff (google is my friend, “keygen” also)(I want to pay 1 euro by utility, not more).

My page seems to be censured in China.

For the Toulousains it is by here with a proposal of network with a grid.

Description of the problem to be solved, resulting from the true life, I quote:

“Madam, Sir,
I live in a small hamlet (place says Mouillecrotte) which depends on the commune of BRESTOT in the Eure.
I also live to 9 km of the DSLAM of which I depend and with 80dB I am not eligible with the ADSL.
I live to 11 km of an antenna WIMAX and I thus made make on site a test of eligibility with the WIMAX (currently deployed in the Eure by Altitude TELECOM for “to cover” white areas of the ADSL)… which is revealed negative (I live in skirt of forest and am not thus not in visu direct with this antenna WIMAX).
In short I live where high banc is not available!
Please help me!

Cordially "

Here “Gruyere of high banc” ADSL (for Loiret) and I place in lower part the establishment of WiMax by Medialys:

Interesting, not? WiFi, it, go there or people want it.

One marché “without Community wire” emerge.

2 fingers of policy (05/2007): "1Mb/s pour tous"?? ... tous pour 1Mb/s !!

It is what a promise? "80_ of them [connected SME] will use them systems of the BLR"
It was in the year 2000, 7 years ago already…

Message for the happy holders of licenses WiMax with 1 euro (les areas: Aquitaine, Alsace,
Burgundy, Brittany, Corsica, Poitou-Charentes):

I invite you to stipulate obligation with the owner of the license, with to authorize its customers to resell their access, according to a price defined by them, the customers.
Ben, au case or WiMax would disappoint, one can play the chart (!) complementarity (include/understand WiMax+WiFi).
Why I say that? Because a WiMax pylon is sold 90.000 euros. I speak just about installation, not of refunding recurring of the licenses, which will fix, in the long term, it price of the subscription for the final customer.
The known as areas will be able to take part, admittedly, but how long?
Prudence is of setting:
“Them misant WiMax operators on the rural areas of the markets developed will expose themselves to low registers difficulties as services DSL will take a lead in these territories. “
Will FT “encircle” the WiMAX pylons?

"We knew that would fall to water. But nobody could to allow not to follow the doctrinaire approach"
All goes very well, madâÂâme Marchioness….it was in 2002!!!!

WiMax France (2007): “Le parcelling out of the licences does not allow an operator to reach a sufficient critical mass to be profitable." 

“Even if it is initially used for the installation of territory, le principal market of Wimax is the cities“, has however estimated Jean-Paul Rivière (Altitude Telecom).

The intervention des local government agencies in telecommunications. Report of work of the Committee of the Networks of Public Initiative. March 15, 2007.

WiFi against WiMAX according to Intel.

And to say that there is (still) qui of it croient in WiMAX as rural solution!! Perhaps they will be right in 2013, if Altitude Telecom poses indeed 3500 WiMax stations.
Hold! me ici is told (minute 52) that one needs 10.000 stations to cover France….I can in to conclude that Altitude Telecom will not solve digital divide with WiMax???

In Ukraine: “[] the viability shows of the use of BreezeMAX 3500 in the service road of zones urban very populated"

In Germany, "Purchasers [of licenses WiMAX] engaged to ensure cover of 25 _ of the communes their zones from here with 2011"

I put myself a question about mobile WiMAX (802.16e); how not vider quickly the battery of the mobile device (considering the power of emission necessary)?

La radio coverage is better of 20 _ into 2,5 GHz than into 3,5 GHz. Include/understand 20 _ better in WiFi than in WiMAX, for France. To cover zones of not very dense habitat in Mobile WiMAX with 3,5 GHz will be thus rather expensive.

If WiMax does not go, there will be always WiFi:

La France puts itself at the “mesh”.
“What one less knows about Wi-Fi, [...], it is that it allows point-to-point connections in line of sight bordering les 10 kilometers thanks to the antennas with profit " (source Pierre-Antoine Baubion) (I confirm, I did it. With 100mW (in the air), 10Km is the high limit.). The configuration which I have used was: An access point (AP) normal with the emission and a large antenna with la reception.
This means that it becomes possible to collect a AP well beyond its coverage area for portable PC or PDA.
They is the Parisian ones which will be content, grace with 400 hotspots installed from here the end of June 2007!

The world records are of 200Km with 300mW in the cable (=25dBm) (it is not thus not 3 times more than 100mW in the air, but much more), which fact a power in the air (EIRP) of 48Watts!! in France us naps limited to….100mW, i.e. 480 time less. (source)

Le record passes to 382Km (18/05/2007) with large antennas:

And with 304Km in 5GHz (802.11a) ., between Sardinia and Italy.

Here vidéos of France3 on WiFi installations for to serve “isolated” villages.

To bring Internet to those which do not have the ADSL:
The large ones networks…

A company (Alsatis) vous allows to equip in WiFi.
Nomotech équipe la Normandy in WiFi, by using fiberoptic and beams herziens.
A mariage fibre and WiFi (rather than WiMax) in the Rhone.

And of small networks…
Rural WiFi is in evolution, here solutions which function:
With Viviers-lès-Lavaur François Laperruque used one mesh network with OLSR (comparable with OSPF) and freifunk (site in German)(binary)
With Saint Cernin Sylvain Lacaze used the infrastructure mode with WDS
With Vaours  Didier Lebrun diffuse Internet in WiFi, via the satellite since 2003. Didier uses 4 ADSL bonds today, since it coupled its network with that du plate Cordais, as one can see it on this splendid chart, in time reality.
In Villeneuve-Tolosanne, Marc Quinton with used OLSR and sound clean firmware here a vue satellite with the bonds.
With Verdon Patrice Valade cross the valley of the Dordogne to bring it high-flow (in WDS) in a zone “whiter than white”.
With the Mayet of Montagne, Dominique proposes the subscription with 20 euros/month. La chart of cover of its network.
With Tamniès, Richard Bonnaffoux, propagates his network of relay out of relay.
With Plougonven, Marc replaced his satellite connection by one line Symmetrical France Telecom with flow guarantees. (SDSL, 2Mb/s).
Has Trébons on the fatty one, Jean and Robert made their network which serves 25 hearths (07/2006).
With Catllar and Los Masos, l' association WiFi Conflent crosses the valley of the Small fireclay cup to deliver Internet free, with a mesh network composed of “ Locusworld “house”.
With Saint Bresson, a chart of the network, provided by Pascal of Montpellier. (Quand the townsmen give the hand to the rural ones)
Mode WDS is used to make pass the bytes of network of a point WiFi access to another.
One can thus relay the WiFi network on the long ones distances and to benefit from the WiFi network at each point access.
Here un access point the commercial, which allows this function WDS. (80euros, 06/2006)
The difference with a repeater, is that the repeater allows to relay that only one apparatus WiFi (only one channel, the distant MAC address is repeated).

La list RAN (Rural Area Network) gather this kind of initiatives.
One has sympathetic meetings: Meet spring RAN  27 and May 28, 2006 with TAMNIES
To put faces on names, voici a photograph of group.

Among the last meetings, there was: Meetings wifi 23 and September 24, 2006 in the Verdon.

The last meeting was it
Saturday March 31 and Sunday April 1, 2007 with Spring Wi-Fi 2007in Auvergne
Here < C 0 > Report of these enriching and sympathetic nerves days.


Panorama solutions [CPL, WiFi, WiMAX, 3G (GSM), satellite] by 01net.

Solutions carried out by professionals:
In Ardeche
WiFi with a grid in Gordes in Luberon.
With Borce and Etsaut, they use a mixture of CPL (30MHz max) and WiFi to prolong a satellite solution.
The CPL " outdoor " (of the meter towards EDF) and “indoor” (of the comptor towards the catches sector) is in the course of normalisation (10/2005).
However the CPL could scramble the band AM as that seems to be the case with the U.S. (video of 18Mo, in English).
One speaks about EDFbox and
Typically, on a zone or there are 350 NRA (connection ADSL), it is necessary to equip 10.000 transformers EDF (with 3000 euro each one), for distribuer CPL with the same subscribers. Moreover, it is necessary to place one repeater CPL at the end of a few hundreds of meters to preserve a flow raised. In addition, it would seem that the steps administrative are long; 1 an, before the implementation (03/2007).

If France Telecom (FT) wants to make cernage (to install not-profitable high-flow to kill competition) on each network, they will have to spend of the thousands euros for each more quoted WiFi network high… pffff….: -)

“France telecom wants to make of Wi-Fi on the commune of 300 inhabitants or  I have deployed my transmitters !!!! “known as Dominique.
When I see that FT establishes concurrent solutions (in WiFi) precisely there or of the microphone-operators (see list RAN) try to survive, whereas there is place with side to serve those which do not have an access Internet high banc, I am only fairly astonished by the absence of FT in Wimax on the metropolis. The ARCEP me surprised ace!.ouiap, does it of it is FT who is not interested in this market… officially… for to make good figure.

Saint Cernin and Viviers are already encircled. FT even encircles maintaining communes front que the material is operational. Example: Wifi-quercy-south. Ca wants to say that you only have very little money to spend to make you encircle (to declare one association aiming at installing of WiFi mutualized) (more you you make known, more you will be encircled quickly). Super not? :-)

During this time, FT puts WiFi more rentable in the cities. They are not only… “Visiblement, Sprint intends to aim at the market of the nomadism".

FT sold a contract with some departments so that the distributors of their communes are equipped in high-flow before 2006.
It is the charter of the departments insulators innovating.
One is there; made the assessment.

But what some did not include/understand, it is that it acts to only equip those which peuvent to have the ADSL, therefore are excluded, the private individuals (or companies) who are too far from distributor !! (environ 8 _ of the population) Then in the statistics, guess how one counts? and well on count the communes equipped, and of ceux blow which is too moved away from the distributor, pass to oubliettes.

At the end of 2007, it will remain 620 000 Net surfers on the edge of the motorway of information (99_ covered population).
Me would have been lied? according to this source (HDRR) who relates to 11 areas there would be 800 thousand people concerned (in 2006!), which would make 1.600.000 people for France (22 areas).

Still a detail concerning WiFi flow in shift:
If you wait until new materials (bringing more of flow) appear to equip you, I think that you mislead a little, here why:
There are 2 types of connections

The first are done at short distances, thus for theoretical outputs raised. But because of reflexions, these flows are less. Evolutions technological current tend to erase this handicap (see MIMO, increase number of antennas). "débits realities oscillate between 40 and 50 Mbit/s inside the buildings and they exceed briskly 100 Mbit/s in space opened" (04/2007)

The seconds will thus not benefit from technologies aiming with to mitigate the effect of the echoes.
In addition, it seems rather difficult to me to install 3 (or 7) antennas, there or there is one of them today.

The current flows obtained in shift, are likely to return le current material perennial than that of the cities.
In conclusion, it is maintenant qu' it is necessary to invest on installations of countryside.

When the débit cities (FTTB or FTTH) will be larger, it will be enough to change - not antennas - but the cases (connected to the antennas) to convey the flows futures. The cost of the evolution will be thus minor.

The FTTA (100Mb/s in 42GHz) is an abuse language, its range (3Km) is very reduced by report/ratio in WiFi.

Objectives and strategy

I seek to know if one can make a waiter of authentification which is not centralized; if you have des brief replies, you can to write to me (to remove _POURIEL email).

 See the bonds which leave this page.

Where are my readers:

Locations of visitors to this page

Test your flow in the 2 directions.

The blog of WiFi.



Why this page?
That is it what it is WiFi?
Why one speaks about it now?
Why gradually legal in France?
WiFi for what to make?
Which are the competitors of WiFi (2.4GHz/11Mbit/s)?
Swelled, Is MHz, Mb/s it what?
Which is the bond between flow in bit/s and bytes transferred?
Which are the various standards of WiFi?


Y' has what has to take that which goes most quickly?
Why y' has 2 antennas on the AP?
Yes, but all that they is assumptions?
How to take the control of its material via Ethernet (RJ45 or WiFi) since a PC?
With what is used for to take the control of its material?
How to know if a WiFi connection will go, without material?
How to implement the software “Radio operator Mobile”?
Y does it have ancestors of WiFi in France?
Which is the price of the material to date?
How a link long distance is established?
How one takes omnidirectional a step expensive?
That is what can scramble us and which scrambles one?
How to do without the electric cable until the AP?
Then what to make? 4 solutions
Which solution to choose?


It is what a hub (repeater), a switch, a bridge (bridge), a router?
How to choose between fixed or re-used address IP?
Who buy IP addresses?
Addresses MAC, addresses IP re-used, how that goes?
Not of waiter DHCP on a bit?
The address Can IP, it be unspecified?
It is what a mask of sub-network (“subnet mask”)?
How to identify the computer which is connected to Internet?
To have is an address IP, that enough?
Les Can AP go on the roofs without problem?


Les is waves they bad for health?
How to detect the radio waves?
How to put a mini-PCI in a satellite toshiba?
How to add a connector on D-LINK DWL-650+?
With WiFi is what I would be pirated?
How to improve a WiFi connection?
One can to make a bad antenna “ricoré”?
It there T it has reflectors on the antennas grids SD27?
Why my compass does not work well?
But the lightning?
My antennas are well granted?
Le Does WiFi, that grow bigger well?
How to find itself there in connectors?
How to choose the cables?
End of a myth on WiFi, them walls, them doors and windows?
Who will gain, the pair coppered (the wire of the telephone) or the radio?
Which is the form of radiation of an antenna?
How to circumvent an obstacle?

Why this page? to synthesize and solidify information data by forums, the press, the Web, my courses, my experiments, it good direction.

That is what it is WiFi?

WiFi is means of communication by radio waves enters computers (in the broad sense). From a software point of view, it is the equivalent of a LAN (network) of company.

The protocol used is pressed on Ethernet and allows to make to pass protocol IP (Internet Protocol).

Why one in speaks now?

Because that becomes gradually legal in France. It was necessary initially that Jospin authorizes the radio communications encrypted (with 40 bits with the time) above the public domain. It was then necessary to legislate (withART or ARCEP) on the authorized power, for not to obstruct the Army Frenchwoman (the ministry for defense), former owner frequency.
And then in the USA and in the other countries of Europe, it is already legal, and the sales progress.

Why gradually legal in France?

Because this technology can concurrencer Telecom France, at the time or this one would occur (sentence from there timeless).
ART defends France Telecom: “One gives satisfaction with France Telecom immediately with detriment of the subscribers and one push back with later the decisions in favour of competition " according to Jacques Veyrat.
The limit in power concerns exit of the antennas, what is still very restrictive for point-to-point bonds (by opposition to bonds makes with an antenna omnidirectional). By definition the point-to-point bonds cannot to scramble that very little world since the waves are focused towards the other point. Legislation U.S. (or Canadian) holds some count, it is more tolerant for the bonds not at point.

ART wishes to study them connections point-to-point (thus supérieures with 100mW) on a case-by-case basis:
“The operator who wishes to deploy connections not with not, with ranges noncompatible with the limitations of power indicated in the tables, must to solicit with this effect near the Authority a autorisation of use of frequencies "

It should be noted that liaisons point-to-point WiFi, with less 100mW of PIRE, can be made on 6 with 8Km, contrary with the assertions of the ART, which envisaged 1Km. Voir comment ici .

On the other hand, if you modifiez the antenne (to put one of them more powerful) of your WiFi material, without modifying (to program) power of the apparatus (to lower it), your installation becomes illegal. Ne do not count on the scruples or competences of salesmen of material, for you to prevent some.

In March 2005, and to my knowledge (which is limited), it minimum of programmable power is of approximately 0dBm. For information, 0dBm corresponds to 1mW. The dBm can be negative, not Watts.

The limitations of power are of 0.01 Watt with 0.1 Watt (according to the channels) exit antenna for the 2.4 GHz in France.
In Canada, these limitations are of 4 Watts, in Germany of 500mWatts. Les English (Ofcom), evoke the possibility of authorizing 10W, it who nuierait doubtless in Wimax!!! Me, my opinion, it is that it is Paris which fixes the powers and which they is the rural ones who suffer from it….who spoke about delocalizations???

The 5GHz (802.11a) is to date interdit in France for one external use, but that could change.
Ben yes, that changed (02/2006), to see them powers authorized in fonction of frequencies.
Then I put myself the question: why the ARCEP authorizes the 5GHz with 1Watt or WiMax with 20Watts and WiFi with 0.1Watt?
I am not the only one to ask me this question:
“The needs for frequencies increase. If us let us continue to distribute them in an ineffective way, purely administrative and on a base nationale, the industry of the télécoms will not take its take-off. One needs an approach which is inspired more operation of the market ", souligne with the daily newspaper Vivian Reding. (police chief of the Commission European as regards télécoms)
I would even say “local market”, ViVian, in thinking in Wifistes rural!

The study of SagaTel for the ARCEP (2006)

Good, then let us positivons!!:
With 1W in 5GHz, which range do we have?

11Km, by report/ratio in WiFi 2.4GHz under the same conditions (8.55Km), we have gained….3Km!!
How that is it done? (if little?) Ben yes, with the increase in frequency, the loss in the air passed from -124dB with -134dB.
We lost 10dB because of the frequency and gained 10dB with the power.
With antennas well in height (half fresnel=0, 12m of release) there would be 24Km.
Under these conditions (dégagement of 23m), WiFi 2.4GHz would carry to 17Km.

<>Et then, sometimes I do not resist emulously of you to make divide what makes me die laughter (MDR), like this labsus in pharse phrases concerning WiFi5 (802.11a, 5GHz):
“The Wimax material will be intended for opératUeurs télécoms professional ". thank you Dominique.

One speaks to me here of WiFi with 60GHzto have a flow of 150Gb/s, super! But which is its range by improving the antennas?

While considering that the sight is direct (Fresnel with 0dB) one arrives at…500meters!! thus unsuited to rural WiFi.
To use mon utilitaire  it is necessary to recover the run-time engin which makes 20méga.

If you seek “WiFi” on the site of ART (ARCEP), you find to the 05/08/2004, 47 articles, no the synthèse up to date.

Personal message for the Prime Minister:
Would be necessary to limit the power of WiFi (100mW) at exit of cable of antenna rather than at exit of the antenna (EIRP).
Ca would enable us to use legally des antennas one-way.
If not, France of in bottom flow, eh quite elle will remain in bottom.
In my opinion, it is better that work is done in the Creuse one (nearshore) rather than in China or in Inde (offshore oil rig).
And now, that bears a name: homeshoring (telecommuting in residence)(less transport, less oil taxes… zut!).

I did not find in “small Larousse” of definition equivalent Radiated Isotropic Power (EIRP).

Then you known as “y' with the ADSL”, except that:

Dominique Bordered, communications director of France Telecom for the Limousin, insists on profitability necessary of installations: “We should reach 80 to 100 customers to hope to deaden over three years, the equipment of a telephone centre in ADSL. “ Knowing that in rural area 8 _ subscribers only are likely to take out a high subscription flow, calculations are quickly made: no connection for communes of less than 2.000 inhabitants. It is moreover the threshold that the operator fixed itself public. They were 32.079 with the last census " (on 36.000 communes in France).

My remark with me, it is that one should not take des figures on the practices of yesterday, to predict the use technologies of tomorrow.

I point out les European texts:

Texts of the European commission

- The recommendation of the European commission from March 20, 2003 inviting the Member States with to facilitate the use hertzian lans (R-LAN, also called w-LAN and Wi-Fi) for to reach at the public services.

I do not invite you to read them, because they are fact for to be read by those which write them, only. For people normal (like me), it is incomprehensible. Idem for the Constitition European.

You attendez Wimax? (me I wait to see the price of Wimax: 125 million euros)(hold, precisely, ART wants to nauseate them peasants (and telecommuters) because them licenses will be sold with the biddings (as for the UMTS) thus only the cities will be profitable).
Hold, ils opinions changed! :-)

O you, living deep France, you does not expect with it that the operators of WiMax start with to equip the zones les less profitable in first?? (you saw what it is passed for the ADSL, not?). 99 _ of the population " it is the objective, deep France it is the 1 _ remaining. After all, that only makes 600.000 voters… with less than the townsmen appreciate the holidays with the countryside?
Les forgotten amount with nearly 4 million individuals

Let us note that on 70 projects of communities referred by Arcep, that proposed by department the English Channel is only to have filled the reduction objective of the digital divide.

“But even with Wimax, it will remain to reabsorb a residue of zone white“according to Jean-Daniel Poisson with TDF.

WiMAX is the true alternative to the last kilometer“… Ha!!! It is for that they start by equipping them godforsaken holes (8000 inhabitants) kind .......Geneva, ouarf, mdr, forgiveness… it is not funny. Finally, one can prolong me by of WiFi…
"to profit of a service access to Wireless Internet in the large ones metropolitan cities", CQFD.

85_ places of access [Internet] come from a local initiative"

For the moment, only Altitude Telecom (Iliad/Free) have a license, and can sell " Pre-WiMax " at a reasonable price. Especially if one consider that one can cancel sound subscription France Telecom with 15euros/mois.
If ART is obstinated to sell them licenses WiMax with the biddings, without remettre in question that d'Altitude Telecom, this one will be in situation of monopoly, and will thus be able, in the long term to increase its tariffs. With less that Altitude Telecom be interested only in the companies…

In addition, si too little of licenses are vendues, it will risk se to produce an agreement between WiMax operators (with 2 or 3, that should not be too hard…it remain only Maxtel, Bolloré Télécom, HDRR and the license main road of Free), like that at summer the case for the operators GSM (see price of the SMS and the intervention of the government).

Hold, what I said to you: “Doit one to expect a concentration of the WiMax offer?"
“Last March [03/2007], the operator [Free] did not speak more than of one “experimentation" for 2007 [en white area].
Free hears to propose the nomadism (high banc without wire in the cities)."

It is odd, one had learned to me with a course of economy that if one was less than 5 competitors, one could to agree to cancel the effect of competition. (With 10_ of fine on the sales turnover, that is worth perhaps the blow of to continue like that?).

Ca makes too a long time that we pay l' tax telecom:
2003-1997= 6ans. 6ans * 3 euros * 12 months = 216 euros
It is à us them users, who should be returned this money, not in the State.
216 euros * 15 million users = 3240 million euros. The fine will sera of 534 million euros.
Profit of the operators: 3240 - 534 = 2706 million euros. Not beautiful life for them??
Thank you which?? thank you the ARCEP.
(Calculation is false because there was not 15 million users in 1997, but good… on another side, you saw lowering the price of telecom frankly mobile of 2003 with today??? me, not)(operators continue le racket on the international one : "Le benefit for the operators flirte thus with the 500_!") Le roaming a le BEC dans l'eau!
(source1, source2)

Citizen, is you ready to vote a tax with monopoly, which would consist to pay with the State for each company, X_ of the turnover market divided by the square of number companies of this market? The goal being of to return intolerable economically a monopoly, all in leaving to live the start-up.

There is more than 20 operators ADSL (in 2005). That is to say them owners who will sell of WiMax are animals (I doubt it) that is to say ART preach not for many users citizens. I do not have dreamed, it is well the government which indicated it chair ART?
Why the operators would not get along for not to go up biddings? Because it takes them time for to know.
Thus mutualisation of the pylons or points high, for GSM was not possible (economy of 20_). Damage, it is the user who pays.

WiFi, that feels good! especially since that them licenses is free, finally for the small operators! (05/2007) 
Heu….unless one should not pay?? (Exemple of calculation for the operators of radio local loop)
Yearly rental of provision: R = A * (DF/F) * (DS/S)
S = 550.000 Km² , F = 2450 MHz, DF = 13 * 20 MHz = 260 , DF/F = 0.1, has = 31709395,59
with DS = 1Km² for example, that made  R= 5,76 euros
Yearly rental of management: R = 533.572 € * (DS/S)
R = 0,97 euro

I cannot read the site of the ARCEP, que this texte means (20/02/2006)? qui is operator?
With my Freebox WiFi close to the window, I am an operator?

And then there is the problem of cyber-terrorism.

To choose a FAI in ADSL according to the number of customers who it quittent (churn).

WiFi for what to make?

1) for one mobile access Internet

with the travail, it is less expensive than of to cable and to maintain wire and catches in the walls. Ca does not make to work a person to declare a computer on one catch (that is called mixing) each time that one changes catch. One passes from a part to another (or one limit to one other) without losing connection Internet (that is called it “roaming”).
travailleurs nomads or mobiles are 15 _ more rentables that the others.

with maison, one can profit balcony, terrace without trailing a wire. As soon as one has more one computer, one avoids cables between those. The walls are obstacles, but le WiFi is propagated through the doors and them windows, and follows the corridors and the stair-wells. WiFi is propagated also by reflexion on the houses of opposite.
No the computer in the living room!. that the TV. Yes but source of media (Internet) is the modem and the computer, it is necessary then to send information (music/films) towards TV. To avoid wire in the corridor, there is WiFi. Here why the FAI propose 2 cases: a WiFi modem + a case “decoder” (WiFi towards TV/chaine Hi-fi). Cases of video offset are incompatible with WiFi (even frequency).

in the bus (experiments in Paris in progress), in l'plane (Lufthansa/Boeing with a satellite connection (arrêt operation in 2006) or rather for a380 AIRBUS) (AIRBUS continues in GSM), in the voiture (Ca is called “war driving” when one seeks just to find WiFi terminals), in subway,  the train, with the FAC.
A test took place on the TGV Paris/Bordeaux in November 2003.
The availability is planned for 09/2007 in 52 trains TGV.

Aux stations d'essence des autoroutes (A10). Chez Carrefour (A62).

With the terrace of one café, in public places very attended (they are called “hotspot”) such as les stations, aéroports (list with the 09/2006), the restaurants rapides, in the camping.

in the hôtels (at B&B it is free) (“service B&B is true success! We receive regularly returns users explaining us that they do not go any more but to B&B because WiFi is free there! “ known as Alexander Casassovici, President de WaveStorm),

In the cuisines restaurants, for gérer customer orders.

dans la rue.

Yes, then I say “in the street” but it is for to make “well”.
Hein? Ben yes, customers of the thing-BOXES (ADSL+WiFi modem) are not formed to place their thing-BOX of such kind that it diffuses in the street (so still it is the intention of these customers) (yes, there are money or service to be gained).
If such is your case, will know that it is necessary to place your thing-BOX behind a window for that it diffuses in the street.
This message also concerns and especially the users of FON, which risk être disappointed by the range of their Fonéra (Fon-BOX).
But FON received the message and proposes an antenna patch to fix on a pane, by means of suction cups.
One description of the fontenna, version to be stuck.
The fontenna (antenna FON) increases the range of will fonéra. Demonstration of fabcool:

basic antenna:

and with the fontenna:

Despite everything, one wardriving in Toulouse, during 1:30, with an antenna of 7dBi on the roof of the car, gives 4000 access points (04/2007). Here points outlines (thank you Stephan):

I consider it regrettable that there does not exist site to index and to factorize this kind of initiatives!!
Ca would be used for better telephoning in WiFi or with surfer free more easily…

WiFi free for 8 million habitants (in Mexico City).

On 50.000 hotspots (WiFi access) there are 6 of them _ which are free (at the beginning of 2005). But ça could to change, because the operators seek with to approach final customer. Eh yes, that which reads a publicity pays! (statistically). Imagine that Google puts publicities in the pages of acceuil of its hotspots (one calls that a captive gate). Ca y est, c'est fait. Rrrhhaa, the FAI do not have there thought ..... it is bundle!! ouf! :-)

As for the GSM, the multiplicity of the hotspots could to allow geolocalisation by WiFi.

Big Brother in company.

Summers you “Bill”, “Linus” or “alien” ?? The projet Mobile FON of Internet starts in Spain…
With tomizone, they have competitors (source, chart)

Une interview of the persons in charge of Ozone and FON, on LCI in February 2006.

2) pour to equip the campaigns

That it is for the telephones portables or for the access Internet high banc, policies (same Europeans) do not envisage to force private operators or public to cover 100 _ territory. N the other hand, l' State authorize WiFi to extend the bonds high flow towards isolated communes. A typical solution is to install one relatively expensive satellite connection, which is divided with 15 users by WiFi to deaden the unit on a few years.
Another solution is of relayer with of Wimax and/or of WiFi, the fiberoptics which pass on the motorways or les access ADSL (NRA) of the close villages. There exists possibilities ofaggregation band-width several suppliers access Internet (FAI). But it is possible to make the same thing “manually” for less expensive, while declaring plusieures footbridges (worms of the different FAI) on one even network. The objective is more reliability, than the profit of flow.

 le report/ratio at the National Assembly
a bon exemple by Didier Lebrun (attention at the prices, they drop)

Some personnes actives developed RAN (Rural Area Network), they are reachable sur one mailing-list.
The idea is to bring high banc until thatched cottages, by any means (satellite, prolongation ADSL, CPL, etc…).

Precedent RAN group logo Web
boxing ring
According to
Randomly List

3) pour payer moins

An access unlimited to the quarter of the price of market.

For 4 users with WiFi material, if it there in ONE has which have a connection towards Internet, it can to divide with the autres and to be made rembourser them 3/4 of what it pays. For the 3 others, they ne pay that 1/4 of prix of a connection Internet. (This is normally illegal, one needs a license operator telecom, but good…)
N the other hand, if the 4 users use en even time Internet for to download, each one will have only 1/4 of the flow.
Telegraphic connections (if it of a) take there precedence over WiFi connections. This can be prevented in putting a software who distributes the flows in way homogeneous (see QOS, Quality Of Service, NetLimiter (Windows), wondershaper (Linux)(explications), ).
The remote loading (of large files) uses connection Internet has 100_, therefore in this case the flow is divided.
But in the other cases: - consultation of pages Internet, - remote loading only one among the 4 users, one will not feel difference compared to a solution without division.
Software exists to program the hours to which one prefer to download (in the night for example).

Into WiFi, the flow of the AP is divided by the number users connected above.

In WiMax also, ADSL too. The number of users by sold flow, is called it rate of application.
In ADSL, a nice operator will sell the flow maximum, with 20 users, less nice, will sell it with 30 users.

The access Internet in France, costs 5 computers

That is to say 30euros the access ADSL (or 35 euros the naked ADSL), in one year that made 360euros. Today for this price, there is one normal computer. If one changes computer every 5 years, that fact that access Internet costs 5 computers. Le price of the access is one determining factor for the users.

La digital divide companies, extract of this article (08/2005):

Tariff means noted for a flow of 1 Mbits/s per month
A- Paris, great zones of businesses, equipped with deblocking with xSDL or BLR: 610 euros HT/mois
B- Communes of the first 10 urban areas in province: 660 euros HT /mois
C- Common peripherals of the 10 first urban areas métropolitainres and zones not dégroupées but with BLR: 710 euros HT /mois
D- Other dégroupées zones of the territory: 800 euros HT /mois
E- Territories having a single offer ADSL, generally via France Telecom: 930 euros HT/month
F- Other territories without cover xDSL: 2.000 euros HT /mois.
I point out that there is only a seul owner pair coppered, one company deprived: France Telecom…

4) pour to telephone

With the voice on WiFi (VoWi-Fi), prolongation of the téléphonie on Internet (ToIP)
Linksys and announce a solution of VoIP. Panorama VoIP in March 2006.
PDA GSM/GPRS and WiFi with Skype in March 2005.
Le development of VoIP fact fear with the mobile operators, (perte incomes on the roaming)… especially if one can to program its own telephone.
“The telephones Dual-mode [GSM and WiFi] are already manufactured by Motorola Nokia, Fujitsu, HP and NEC."
See also Milkfish (solution SIP)

Marriage WiMAX and MIMO for telephony.

Des companies of which employees needs have for internal mobility deploy the voice on Wi-Fi with small scales

50 _ mobile calls passed through the employees of companies are carried out at the office, although less expensive fixed telephones are at their disposal"

Project FON will use to telephone thanks to partnerships with operators GSM. (good fear adviser?) :-)
Here one chart geographical de la FON/Neuf cover.
Another with Ozone, Meteor Networks, Fon or the restaurants Mc Donald' S

And with Boingo.

Ca always did me frankly to laugh, drawn radio ranges with the compass (Ozone/Paris)…

""It network does not belong to anybody, “ explained Rafi Haladjian during the launching of the start-up Ozone".
If it were true, nobody would have sold it with Last nine cegetel!!!
To kill competition, the secrecy to bleed the consumer.

The Freebox V5 left, and the WiFi telephone who goes with also… (on can put besides full with things inside.Linux core 2.6)

Wi-Fi Alliance®and the CTIA mettent in place a programme to test the range of telephones WiFi.
(In French, program ne se translated not always by software).

François and all those of his WiFi network can to telephone for 1centime of euro the minute, towards fixed telephones in France, but without passing by subscriptions Internet (offers triple play: freebox, 9box, livebox, etc.)(~30 euros) and subscription FT (14 euros).

Transform an old cellphone into footbridge towards VoIP with Skype and IPdrum (and benefit from the prices “SkypeOut” on your calls of portable, in the coverage area of the old man portable).
One can call that one mobile WiFi, who added to the offer “triples play”, fait une offre..."quadruple play".

Triple play it is telephony, Internet, TV. Le débit for to have the TV drops, si on utilise le Mpeg-4 au lieu du Mpeg-2.

Inventel allows téléphoner with Internet, without computer, and with one very good quality of sound.
And to say that there are some who telephone 6 hours for $29, with skype in the avions [English]! (source)
For the train, one would need catches sector in the oars (to monopolize the catch “razor” in the toilets, as I have it made in the TGV, it is not well!)

Ca is likely to be socially interesting, if one can use resources provided by private individuals (WiFi terminals) for to telephone…
(until or the altruism will go?)(that points out time or one to me went in the neighbor who had the telephone, for télépnoner. One he gave money or one proposed some at least: -) courtesy? …)
There are then preoccupations with a posting, an outline of solution with “craiefiti".
I am not the only one to think thus, to see Ozone.

Soon the festival ofUMA:
The UMA has the aim of promoting the roaming enters networks of mobile telephony and WiFi network servants or access points public with the wireless networkings. The roaming, that please say that one passes from one network to the other in the same one communication.


While waiting for the UMA you can combiner advantages of mobile (GSM) and skype while calling your proper number (skypeIN) which will be redirected towards your corresponding (skypeOUT).

Here a first setting in work, at Nortel Networks.
Another at LG, a téléphone Dual-mode 3G/WiFi.

Telephones WiFi, first tests at Free and Last nines Telecom. (June 2006). Détails implementation.
Favours for France Telecom which envisages to make UMA, with its “LivePhone” (= " Unik”?).
The question that I am posed, it is: Which will be the control of the user on changement of tarif in the course of conversation?
It is well what I craignais….(06/2006)
Compete with between the PocketOne of Internet Club (become Neuf Cegetel) and Twin?

prix VoIP calls (since a machin_BOX) from France.

VoIP functions better if all packages IP borrow it even way, a solution with MPLS.

A company, or a small rural network (in WiFi) can use one waiter Asterisk to connect telephones traditional (analogical, FAX) oubien of the téléphones IP.

5) to make house automation

The product presented is only one interface towards a connection series (RS232), but this field could extend with the management of a house (house numerical) (power station of alarm (already in “without wire”), shutters and electric doors, video monitoring (them WiFi cameras already exist), simulation of presence, energy management, etc…).

projets more ambitieux consists in connecting plusieures technologies of transmission. Personally, that itches me since to discuss a long time on the telephone with my corresponding by listening to it on hatred Hi-fi, with a microphone in the middle of the part ;-)

For faire remote monitoring.

Which are the competitors of WiFi (2.4GHz/11Mbit/s)?

WiFi to the standard 802.11g made 2.4GHz/54Mbits/s.

The values of flow are theoretical, in practice that made less (count half).

The terminals relay of WiFi are the AP, their price is accessible with the private individual.

HiSwan (??)
Bluetooth (12Mbit/s) (carried 10m) (2.4GHz) (protocol L2CAP, pas IP native),
Wavelan (to 800 MHz) (missing),
HomeRF (2 Mbit/s) (disparu),
ZigBee (250 Kbps) (2.4GHz) objective low consumption and small range (to come)

Like telephone GSM, technologies following require terminals relay on pylons (expensive), therefore no FREE connections into point-to-point between private individuals
Hiperlan2 (5 GHz) (à to come) (I do not know the price terminals relay),
UMTS (384Kbit/s) (à to come) en test with Toulouse 102 euros (10 hours) or 15 euros of the heure (3 operators thus not of competition) (here of what to have an idea of the tariffs) (la Visio with 20euros/heure)
UMTS R5 (1Mbit/s) (à to come) (you have 100 billion euros to be refunded, finally if you use!!) pauvres operators (suite)
Flash OFDM (10 times the UMTS) (à to come)
Ultra Wide-band (UWB/WUSB) (400 Mbit/s) (à to come) (range 10m) be likely to replace Bluetooth
Wimax (à to come), it is like WiFi, but with Watts, of the pylons and the licenses moreover, public price and technological effervescence in less. But that seems to change with l'utilisation technologies OFDM+MIMO.
WiBro (to come) (for June 2006) (30Mbps) (2,3Ghz) (Wireless Broadband, standard 802.16e) it is of Wimax for South Korean.
If you move with 250km/h (or less), it there will have in 2006 (still not seen in 2007) le MBWA
BluWan (40,5-43,5 GHz) (12Gbits/s theoretical)

HSDPA (to come) (3Mbit/s) La standard envisages to reach 14 Mbit/s, with a bond going up going up to 320 kbit/s

802.16m (to come, not before 2009) (1Gbit/s into fixed and 100Mbit/s in mobile) with MIMO.

The LTE (Length Term Evolution) (à to come) in 2009/2010 should make 100Mb/s

One day, of the radio connections with 2Gbps (56GHz) with Gi-Fi (spectral effectiveness of 0.035 (28Hz by bit))?

source 01net

Versions owners (compatible with 11Mbit/s with other manufacturers) allow of WiFi 802.11b 22 or 44Mbit/s.
Idem with the 802.11g with 54Mbit/s, there exist versions owners faster: 108Mbit/s and 125Mbit/s. But I believed to understand that certain technologies used channel 6 “to dribble” on those of with side…

One comparison of the costs in energy electric fact of arising that WiFi is 25 times less greedy than WiMAX, which, is to him 2 times less than the 3G WCDMA.

Swelled, MHz, Mb/s it is what?

Hz: is the unit of the hertz. Ca measures a frequency, i.e. the number variations in one second.
Electricity, in France, is delivered with 50Hz.
b/s, bps, bit/s: mean all the same thing: bits a second. Ca measures one flow numerical informations (or data processing) exchanged. A flow measure a quantity for a given time.
bit: it is the unit of computers, that is worth 1 or 0. .
B/s: attention, there is a trap! It is about byte (English word) a second. byte means byte. A byte is a whole of 8 bits.
bps: attention, still a trap. Ca can be bauds a second.
baud (or data base): Ca becomes complicated. The computer knows that the bits which are worth 0 or 1, therefore a bit has 2 states possible (perhaps a day one will code 5 bits thus 32 possible states). Connections between computers can to handle tricks which have 4 states possible. Thus one trick can transport 2 bits.
The number of states is called the valence. The valence of a bit is worth 2. The valence of the trick of my example is worth 4.
And bauds then? The bauds are the tricks of my example.
To convert bauds into bits it is necessary to know how much bits are coded in a baud for the connection in question.
How one makes? Ben for example if one varies frequency (like the frequency modulation, the FM of the radio), one can decide that a variation of +1 code state 1 of the bit and that “not of variation " codes state 0 of the bit. In this case the valence equal 2.
But if one adds the two bits to code them in only one variation, one could have:
“not of variation” if the two bits are to 0,
variation of +1 so only bit 1 is to 1,
variation of +2 so only bit 2 is to 1,
variation of +3 if the two bits are to 1.
K, (K): Ca please say kilo, thus 1000 (if one speaks about Hz).
M: that please say méga, therefore 1000 times one kilo (if one speaks about Hz).
G: that please say giga, therefore 1000 times a méga (if one speaks about Hz).
if one speaks about Hz?
Ben yes, for the data processing specialists who measure bits, 1 kilo 1024 are worth (2^10). It is insane not? But like one is mistaken that of 2_, one lets make confusion.

Thus in my example one can code 2 bits during one time of the frequency (one period corresponds to one Hz).
The way of coding the bits makes it possible today to place until 256 states (= 8 bits) per period (0.125Hz by bit). But these systems are more sensitive to the disturbances and by consequent le flow reality goes down again in conditions of ground (practical flow different from theoretical output).

This is true so much so that one prefer to change system of coding when the connection is bad. It is what explains various possible flows of WiFi according to outdistance.

When a WiFi material tries to change system of coding, not badly of bits are lost. It is thus preferable to select (manually) flow right in lower part.

The spectral effectiveness (bit/sec/Hz/cellule) is division flow (in b/s) by the carrier frequency (2.4GHz for WiFi) and by the number of cells. A cell is one equipment WiFi, there is as much as connected users.
WiFi 802.11b has a maximum spectral effectiveness of: 5.500 000 (b/s)/2.450.000 000 (Hz)/1 = 0.00224
Ca wants to say that one transmits 445 times the same bit (445Hz by bit). It is much, and it is improvable perhaps with 802.11n (theoretical output 320 Mb/s for 2.4 or 5 GHz)(7.6 Hz by bit)
The spectral effectiveness gives an idea of “output” system of coding.

All this is true only in the condition of comparing system codings which have same bandage busy (BP).
Kézako? (that is what it is?): It is the diameter (of) pipe (X) used in parallel multiplied by (ur) the number. The band busy is then the equivalent of the sum of the sections of the pipes used. There you understand that more I have pipes in parallel, more the aggregate rate will be high.
A pipe is a channel. One channel une has width of bande (difference between frequency more high of the channel and the lowest frequency) which corresponds with the section of my pipe.
For information, the 802.11b has 3 (or 4, to see low) disjoined channels (on 13 available in France), whereas the 802.11a have 8 disjoined channels.

One cannot thus compare spectral effectiveness, that for the same BP. l' effectiveness should not be compared spectral of fiberoptic, who has “I do not know how much width of channel” time the number channels (160), with the spectral effectiveness of WiFi 802.11g who have ONLY 13 channels with 20MHz by channel…

To compare one would need a measurement which divides the effectiveness spectral by BP.

When one arrives at a spectral effectiveness near to 1, it will be necessary to worry about the antennas used. Indeed, les antennas with whorls (the omnidirectional ones) superimpose them adjacent waves (periods of the signal, the Hz, for one even bit). The antennas which have one dipole, will not have this problem (example; the family of the SD27).

Which is the bond between flow in bit/s and bytes transferred?

One takes as assumption that none system of transmission is not reliable. Thus one sends a whole of bits (a byte, or one package of bytes) and one checks that they are well arrived. For that one applies before the emission a calculation which gives one key and one sends also the key. With the reception, one still calculate key, it must be equal to the received key. If the bits have summer abysses during transport (different keys), one in inform the shipper and one them reforward. Information of checking and the bits abysses occupy one part of the flow, what explains the difference between the flow theoretical and flow measured.
The more the bits are abysses, the more the flow decreases.
There exist protocols which do not check the good one reception bytes: UDP (which has the same role as TCP).
One can check a byte with a bit additional, that be called the bit of parity.
One can check a package of bytes with bytes additional, that is called CRC or check sum.

Count 10 _ at least flow for control bits (for a protocol worthy of this name),
what makes roughly speaking a byte for 10 bits.


Which are the various standards of WiFi?

The 802.11 is nomalisé by the IEEE.
Le A, it B and it N concern electronics, the others are layers rather software.

type comment carrier frequency (MHz) theoretical output (Mbits/s) range with antenna (S) omnidirectionnelle (S) (meters) range with antenna (S) unidirectionnelle (S) (meters) treatment of the signal No. of channels
802.11A good for the city  5700 54 50   ofdm 12
802.11B most widespread. good for the countryside 2400 11 100 8.000 (legally, if not more) dsss 13
802.11c functions “bridge”. becomes 11d            
802.11d limitations France and Japan           13
802.11e quality of service (ping and flow minimal)            
802.11f roaming (change of access point without disconnection)            
802.11g B version “city” 2400 54 100 without obstacles
8.000 (legally, if not more) ofdm  
802.11h management of power against mutual jamming 5700       ofdm  
802.11i safety, replacement of WEP. The WPA is not retained (with TKIP).
The WPA2 is retenu, based on an encoding AES.
802.11j as (legislation has it different) 4900          
802.11k routing according to the quality of service, geolocalisation            
802.11ma Improvement definitions on the level of the physical layers (PHY) and access to the media (MAC)            
802.11n not before 2006 (use of technology MIMO). 2400 or 5700 320, 600 en 2007
higher than the 802.11g with reflexions
??? depends on the number of antennas (3 to 7)
802.11p Wireless Access in Vehicular Environments (WAVE). roaming fast            
802.11r roaming rapid directed continuous flows, streaming kind video, VoIP etc…            
802.11s networks with a grid ;-)
assessment in March 2006
802.11t Wireless Performance Prediction (WPP) methods of test, QoS.            
802.11u harmonize the communication between WiFi and networks external (telephony)            
802.11v management networks WiFi: monitoring, configuration and update (centralized)            
802.1x authentification on waiter            

Comments on the table:
range: with an one-way antenna (only one direction, 5° approximately) or omnidirectional (on 360°)
the number of channels depends on the laws on each country. Japan is passed from 3 to 14, others 8 or more. France passed from 3 to 13, but by keeping a EIRP limitation, i.e. one can legally use only the antennas omnidirectional of origin if one does not separate the emission from reception.
More there are channels less one has chances to be scrambled mutually in the same geographical area.
Parmis the 13 channels available in France, it would be possible to use 4 of them disjoints (which does not interfere between them; channels 1,4,8 and 11), if one takes into account filtering. (Passe-bande (in pi) constituted of 3 components before the exit antenna?)

The usual disjoined channels are: 1,6 and 11.
The more one is scrambled, the more the flow drops (automatically).
Treatment OFDM is more suited than DSSS to to eliminate it mutual jamming.

With the WiMAX, other standards arrive.

source Marc Olanie
source Marc Olanie
source ... securisation_802.11b.pdf


Y' has what has to take that which goes most quickly?

The increase in the carrier frequency has 4 effects who us interest:
1) the course of the radio waves approaches a line (like the light). Disadvantage: one does not circumvent any more obstacles. Diffraction becomes weaker. Diffraction it is when the object which receives the wave becomes small transmitter.
2) the waves reflected on walls are more strong. (phenomenon of echo: advantageous inside buildings, becomes scrambling if one receives the principal wave and sound echo.
3) it authorizes a flow (in méga bits by seconds (Mb/s)) higher.
4) the way traversed by the waves in the air, for same received power, is lower. Diffraction in the air, rain, snow.

It is for that which I say that of WiFi with one high frequency is better for the city or the interior of the buildings, then that WiFi with more low frequency is better for countryside.

I do not know what it is of capacity to cross walls. A priori, I would say that with the increase in frequency, the waves cross better (x-rays to do radios of human body). More the matter contains water or metal, more it stops the waves. One needs holes larger than la wavelength for letting well pass the waves.
The wavelength is calculated as follows: lambda = speed of light /frequency.
Example: 12cm = 300.000 km/2450 MHz.

Théorie source radio and calculations , la theory of the antennas summarized a little (in English)

I simulated comparisons (large a mercy in Bent Roger and its software “Radio Mobile") à:
144 MHz (frequencies of the radio operator amateurs), 2400 MHz, 5700 MHz.


Why y' has 2 antennas on the AP?

Everyone agrees to say that electronics selected the antenna which receives best, but…

A AP is an Access Not (Access point), i.e. one black box able to receive packages of bits of a shipper has and to send them to one recipient B, then that has is not in connection with B. One calls that the mode “infrastructure”, in opposition to the “ad hoc” mode which does not allow to make relay.
The antennas in question are omnidirectional.
That is to say the problem which consists in making an antenna who does not collect that a transmitter, at a given place, with an aim of being it less scrambled by the others.

The principle is “to work” these differences to reveal only the wave which comes from a direction data (our transmitter).

That is to say a part or there are a musical bottom, and two people in the corners.
I place a microphone (A) close to that which speaks, and one other microphone (B) with the mileu of the part. By electronics, I calculate the signal resulting from has less B. This signal gives me le speech of has without the musical bottom. It is like that one fact one-way microphones. Finally almost.
As the sound is propagated with 300m/s, it is necessary that I delay (for to superimpose) the signal of the microphone has before withdrawing to him the signal B. For 1m between has and B, I must delay has of 1/300 S, equalize 3.3 milli second (3.3ms).

With our two antennas it is the same thing, except that les waves move with 300.000Km/s. I suppose that it is that it mode “diversity”.
10cm => 0.33 10th-9s = 0.33 nano second =>3GHz

Antennas with 8 branches existent, but not yet for WiFi (though). It d'antennes acts adaptatives (intelligent). In this case one does one species of triangulation (by treatment of the signal) between the antennas and the point distant to filter the signals which are intended to him. Antennas must be isolated of a half wavelength (6cm for WiFi) (dephasing of 180°), this relates to only the omnidirectional aerials to make them equivalent to an antenna one-way at one moment T (see ZoneFlex from Ruckus Wireless). It would seem that the UMTS will not benefit from this system, already operational in other countries (Asian).

One from of thus deduced, that in urban environment (there or there is many echoes), modes “right antenna only” (or left), give less good performances that the mode diversity.

The treatment of the echo on walls is similar, put at leaves that the wave is shifted much in time, since it course of the larger distances. Thus one difference of way of 15 meters corresponds to one difference in time of 50 NS (20MHz).

One can also exploit the travel time wave to determine the distance to the AP, and thus to fight against the hacking.

Products with 7 antennes is in project at Netgear, with the “BeamFlex technology". The competitors Linksys (Cisco) and US Robotics await standardization 802.11n. Belkin proposes (March 2006) material incompatible with the future standard 802.11n.

Telephony 3G should use MIMO.

If technology is mixed MIMO and non-MIMO, one gains nevertheless in flow, but not as much that with all the apparatuses in MIMO.

source ... pdf
source Marc Olanie

Yes, but all that they is assumptions?

Ben not, not since I have checked! and fake.
Initially it is necessary to move away the thingummies wifi parce sufficiently whether they “sprinkle” or “radiate” with a little everywhere (I checked with paper aluminum to make mask), all is not armoured inside. I suspect them to sprinkle (also) by the electric cable (considering that the part transmitting is right beside food part, finally on my USR2249).
My AP continued to function whereas I had disconnected 2 antennas, the guy. I moved away them (one on the 2nd floor, the other with the DRC), once separated from 2 times 25cm of concrete, that calmed. I gave the antennas to carry out the tests

Then, “it” is said to me that the selected diversity mode better antenna?? Thus if I remove one of them, then that I gives and that I remove the other, (by keeping all environment identical (position, orientation, polarization, natural person in the beam of waves, etc.)) then mode diversity must function at least as well as one of the modes only one antenna?
Ben they is FALSE, it walks more badly, ya not photo! Whether it is the right-hand side or the left which I connect only.

Moreover you can test yourselves! (be methodical, rigorous and patient)(one draws the conclusions only on what is répétable).

Attention, the diversity mode is a mode of training, it fact of the tests. I thus made an average slipping on 8 measurements (several tens of seconds), and I concluded when measurement is stabilized. Contrary the modes only one antenna are stable as of the first measurements.
Ca seems logical, since it has many treatments (described above) to many measures, to make in mode diversity.

The diversity mode seems very “intelligent” since if I put an one-way antenna (on the right exit) in more of the 2 basic omnidirectional aerials, it is left there very well! Ca please say that it does not suppose the spacing enters les antennas like constant, but which it has an algorithm more complicated. By analogy with the microphones, it is a little like if the one-way one collected the signal of that which speaks (A) and the 2 omnidiectionnelles ones collected the background noise (B).

But the manufacturers, they say ...... Manufacturers protect their secrecies of fabriquation.
Moreover the tendency is to be protected moreover in more. At the beginning, the materials could be programmed (example of those which contain PRISM II) whereas my USR2249 (more recent) who contains a ACX100, is protected. The kit of development is communicated only to the partners of Texas Instruments. These partners engage (by contract) with not to diffuse the kit. unixiens write pilots, thank you with them!

But good, the last chips are credits, and allow “to treat” also network GSM. When it is known that certain solutions of payment (“protected”;)) ) on Internet, are based on codes communicated by GSM, you see what I want to say…

And then when a manufacturer (Agere) arrives at results (technique MIMO: 162Mbps in 5GHz) , it releases a little knowledge.


How to take the control of its material via Ethernet (RJ45 or WiFi) since a PC?

charts PCMCIA with a chip PRISM II
[PrismII based cards: limited functionality]
Addtron AWP-100
Ambicom WL1100 PC
Belkin F5D6020
Compaq WL100
Dell TrueMobile 1150 Series
D-Link DWL-650, DRC-650, DCF-650W
Farallon SkyLINE
HyperLink Wireless
LA4111 Spectrum24 Wireless LAN
Urgent Linksys WPC11 Wireless
Netgear MA401
Efficient SpeedStream SS1021
ZCOMAX AirRunner/XI=300 source at the beginning of 2003.

To choose its material, here list materials compatible with NetStumbler, a tool for scanner/to seek access points.

Le WRT54G (Linksys/Cisco) turns under Linux (y' in has which very included/understood!), and thus can for this reason, to be charged with what one wants dedans. Par exemple SveaSoft ou OpenWRT. USRobotics leur emboite le pas, but with only one antenna: - (
Here are others:
Sveasoft (17 euros), DD-WRT Projec, HyperWRT, Wifi-box project, OpenWRT (for the pro of Linux).
source des URL above.

The version 5 of WRT54G, did not make it possible to change the software (firmware), but maintenant it is possible, without modification of the material.
The customer is right always, thus, Linksys/Cisco having included/understood what made the success of the WRT54G, it leaves WRT54GL (L like Free?)  :)

With the case or the WRT54G a rare food product, the RAN devriendrait group, and more particularly François Laperruque, propose one alternative (for the moment in 802.11a with 5GHz); the chart “WRAP " and its software (voyage Linux) and the charts WiFi (miniPCI) which go inside.

And as the RAN group is very active, Marc Dilasser, it Breton, gives us one description of Buffalo WHR G54S who is an alternative to WRT54G.

One another solution(?) is the FRITZ! Box who has in more one modem ADSL and the voice on IP (VOIP)(SIP)(Fonisée). It is a “box” multi-operators, if you want.
The problem it is that the antenna is not detachable!! have! damage, not good for the RAN.
166 euros/unit oupss!

For a portable PC, best is to have a connector external on the chart PCMCIA, exemple.

With what is used for to take the control of its material?

It would seem that the electronics of the AP or the WiFi charts that is to say the same one. On the other hand the prices differ, and them functions too!
In short, there are two types of material: the 802.11a and the 802.11b.
The 802.11c types, D, E, F, G, H, I, 1x, are only software who can be placed in microcomputers dedicated to these tasks (microcontroleurs are said) and consequently locked up in a case.
Another solution consists in programming it material since your computer (the PC). When the logiciciels are available AND that the material allows this programming, one can to choose the functions which one wants to make with WiFi.
For example:
- to adjust the power of the emission part (to minimize les interferences) [; -) with FT&ART],
- to post the force (dBm) signal received (that avoid being equipped of a portable PC with pcmcia and netstumbler),
- to connect WiFi to the LAN (network filiare) (with a catch (RJ45) and a cable which joined the charts network of others computers),
- repeater (to be used just as relay to connect 2 others wifists),
- router (to choose towards whom to send a package received which ne is not intended to me),
- waiter of IP addresses (to allocate these addresses dynamically (DHCP)),
- encoding,
- waiter of authentification.

The problem is that if one wants to move away the antenna from the PC, to put it on the roof for example, the power sent and received by the material WiFi will be lost in the cable (one “pigtail calls”). And the PC is not put on the roof.

How to know if a WiFi connection will go, without material?

For outside, the software “Radio operator Mobile” allows to simulate:
- emitted power,
- losses due to the cables and connectors with the emission,
- profit of the sending antenna (plain or omni directional),
- losses due to the relief (and other characteristics) of ground,
- losses due to the distance in the air,
- profit of the reception antenna (linked or omni directional),
- losses due to the cables and connectors to reception,
- the threshold of sensitivity of the receiver.

Alas, I did not find information on the buildings and the trees (information “sensitive” for soldiers). Thus an impossible connection according to this simulation, will be really impossible. On the other hand a possible connection ne will be really that if there is no obstruction by buildings or trees. In this case, only the tests will be able to answer with the question. If one sees (with the eyes) the distant point, it is inevitably not the best case, but ya the strong ones chances so that it goes.

Here what can produce the software “Radio operator Mobile”:

The green bonds authorize a flow of 11Mb/s theoretical (or 54Mb/s), the yellow bonds authorize 1Mb/s.

For the interior of a building, the Wiplan software of at Sygmum (the company do not exist any more), allowed to simulate the cover of each AP:

But this software is not available any more because it belongs henceforth to the INSA of Lyon, as Guillaume explains it to me Of La Rock, a programmer of Wiplan
On the other hand, other software exists:

- “SITE PLANNER " of wireless valley ( ( ) with The USA which has just been bought (2006) by Motorola,

- “WINPROP” of Awe-communications in Germany. Example of result:

There are some which treats the wave propagation into 3 dimensions: SimuEM 3D.

How to implement the software “Radio operator Mobile”?

Software “Radio Mobile " use data of altitude to simulate connections radio.
From maporama and of one postal address, you will find the coordinates in degrees/minutes/seconds, to locate a WiFi station in the software.
You will find the data of altitude for your need on
I used the format dted0 with the precision “Military Specification”:

Desired Product Level = DTED level 0 (30 arc second)
Navigate using = Starting Center Not
Starting Center Not, to seize the coordinates
to click “roam DMS” => new page, to go in bottom, to correct the coordinates to have multiples of degrees,
to click “raster to export” => new page, to click on the file ZIP, which makes 80Ko by zone of 1 degree square.
To decompress by keeping the repertories.

Y it has ancestors of WiFi in France?
Yes, of the networks radio operator amateurs already in place (144MHz for 9.6Kb/s), we give the example:

“Within the framework of excavations in the Russian basement until 100m of depth, the Russian scientists found vestiges of copper wire which went back to approximately 1000 years. Les Russians concluded from it that their ancestors laid out already it there is 1000 years old of a network of copper wire.
The Americans, to make good measure, have also proceeded with excavations in their basement until one depth of 2000m. They found glass fibre remainders there.
It proved that they were approximately 2000 years old.
The Americans concluded from it that their ancestors laid out already 2000 years ago of a glass fibre network numerical. And that, 1000 years before the Russians!
One week later, in Belgium one published it official statement according to: Following excavations in the Belgian basement until a depth of 5000m, the Belgian scientists do not have anything found whole.

They concluded from it that the Former Belgians laid out already it there is 5000 years old of a Wifi network. “: -)

Which is the price of the material to date?
there are many sites!


How a link long distance is established?
One remplace the antenna of origine by - is an antenna the commercial (SD27 24dBi, with 65euros), - is an antenna manufactured oneself (limps of ricorée 11dBi) (antenna strainer). One can also put a key USB in one limps of ricoré, but on ne does not know the profit.
clés USB (which makes 15dBm instead of 17dBm for the AP) exist with external connector, to connect an antenna at fort there profit:

If you have chance, your key USB contains perhaps one connector U.FL (inside), Comme that one:

and here with a pigtail U.FL (towards RP-SMA undoubtedly) connected above:

This connector seems to be a MS156, rather than one U.FL (need for a point of adhesive).
Attention with the mechanical brittleness of assembly!!

The WiFi material of origin is delivered with one (of) antenna omnidirectional (right stem which “radiates” all around it, with 360 degrees), to respect them laws on power output in an unspecified direction.
By replacing this antenna by one-way (which only in only one direction) one emits increases (one concentrates) power since the wave is not sent any more on 360°, but on 18° (approximately) only, therefore 20 times more.

The P.I.R.E is the radiated isotropic power equivalent. It is the maximum of what is measured around (360°) of one antenna. Thus for the same power sent by the cable (pigtail), an one-way antenna will have EIRP 20 times higher with omnidirectional. Les dBi precisely measure this faculty to pass the waves in air.

Hold, for example here are some photographs of my antenna ricoré, connected to my AP (a USR 2249):

and when I would like to put 3 ricorées (to make a network with a grid), I would use one coupler 3 voies, oubien I would manufacture it myself (considering price of $60).

One uses a coupler to have a adaptation of impedance (50 ohms) and not too many losses.

After being itself occupied antennas, it is necessary (to try) to regulate certain parameters related to the travel time (longer) waves, to obtain good flows
For the moment (09/2006) it is not possible that by changing it firmware (software) of the WRT54G and en carrying out UNIX orders.
(thank you Didier).

If one does not fear the illegality, one can also put one WiFi amplifier on each side of the connection.

How one takes omnidirectional a step expensive?
with gros cable hyper frequency do-it-yourselfing.

With a little more work, on can make 10dBi with 7 whorls, oubien a 14dBi (9 whorls) at a cost of almost 10euros (without to count madeleine and the end of chocolate; -)).

That is what can scramble us and which scrambles one?
The apparatuses which are on the same frequency or at frequencies harmonics. A harmonic is a multiple frequency of one another frequency (of which the division of largest by more small an entirety gives).
with 2.4GHz, there is (I always do not have checked these information):
the microwawe ovens (I have checked),
the missiles crotales or radars of the Army Frenchwoman,
transmitters (or cameras) for to off-set it video (television),
some radars automobiles which, in WiFi, close to the road, send their information to the police officers hidden further.

The tool to observe if the frequency is free be called one analyzer of spectrum, but it costs an arm and a leg! By against there exist equivalents not expensive (150 euros when even, 09/2006), to see Wi-Spy. One version antenna external of Wi-Spy.

A beautiful morning the network of Sylvain did not function any more, afterwards investigations, it determined that sound network was scrambled.
Stephan, Sylvain and myself let us have decided to seek it culprit, with a WiSpy key which was made directional thanks to one limps of ricoré. Here thing: Then we checked directivity by putting a AP on channel 2 with one “beacon interval " of 1ms (at bottom thus):

Variation of the width (of the intensity in fact) of gross column on the left must with the fact that I turned the antenna in another direction that of a. On channels 8 with 11, one a normal use of WiFi (beacon interval of 100ms) sees.

Then we left to fish… and we found one beautiful LINE: -) a frequency with 2434MHz which sprinkled until bell-tower of the church:

After a psychological pressure exerted by Sylvain on owner of the hovel, it is proven that our jammer consisted of a camera with video offset, which is not perhaps not approved in France, if it comes the USA (200mW).
The apparatus génait also a WiFi printer at another neighbor.
As an information, my “wavemeter” does not have anything found, since the power was not enough raised for him, then that this same power scrambles WiFi which works with of low powers (- 60dBm).

A solution (temporary?) consisted with to change channel on the network of Sylvain. If the parasitor persists, it will remain to determine if it exceeds the French standards (with to see with the ANF, 1500euros of fine). If it is not the case it does not remain any more which has to pass in 802.11a oubien to put fine netting to make obstacle. La bandage 2.4GHz is free with all that implies!!

To have an idea of the quantity of use of the waves radio, look by example among Belgians.

How to do without the electric cable until the AP?

The POE consist has to use unutilised wire of a cable network (connector RJ45) to bring the food to a. Wire unutilised (Bi-directional dated) correspond to a use for telephony (ISDN). If you directly connect the AP to the PC it is necessary to use a cable crossed (also called “cable modem”) which connects:
connections “tranceive” of with dimensions has with the “receive one” with dimensions B,
connections “receive” of with dimensions has with the “tranceive one” with dimensions B.

It is thus judicious to use the twisted pair (7-8) for to bring the food. “Returns of measurement” (judicious) can be used with terminals 4 and 5, for to check that the tension with the good value at the boundaries of a.

Caution!!! I was made have, when I have cabled that, that did not go. The indicator “power” ignited but the AP did not function (the indicator dated does not ignite).

Explanation: my Ethernet cable made 25m (one has right with 60m). When I put the food of an end (with dimensions PC), other boils (with dimensions AP) I lost 1.2volt. Like the food deliver 4.7v (in load), there remain nothing any more but 4.7 - 1.2 = 3.5v, and there, the AP go more.

Then what to make? 4 solutions:

1 - to put the computer in the attic, with one cable Ethernet of 5m: Ben let us see!!! it is owl as solution (them owls live in the attics, not me).

2 - to use the 2nd twisted pair to put it in parallel with the first, as that I would lose 1.2/2 = 0.6v, therefore the AP will be fed into 4.7 - 0.6 = 4.1v, I do not have tested, that perhaps go (for 25m).

3 - to change the food for another which deliver 4.7 (or 5) + 1.2 = 5.9 (or 6.2) volts. Like the consumed current do not vary, it is not the sorrow to seek to control tension with 4.7v with the 2nd twisted pair.

4 - to use the POE only at one minimal distance. If the goal is not to put a 220v under the rain, the useful part of POE can be summarized at distance AP/attic. In this case length POE is equal to: length of mast + 1m (to go to the shelter of the rain). For a mast of 4m, made Ca (25/(4+1))= 5 time less voltage drop, i.e. 1.2/5 = 0.24volt. There, it is sure that it goes, since there remain 4.7 - 0.24 = 4.46volts. To 4.2v it goes, to 3.8v it goes more.

One can notice that there remain 25 - 5 = 20m of pair twisted who can be used to bring the 220v until the food (it is more aesthetic to have 1 wire which trails, rather than 2 (according to my wife). The risk is to bring parasites (50Hz) on wire of data, but like the current is weak there is chances that it goes.

Calculation of the current on the 220v: puissance= 5volts * 0.8 amps = 4watts. A bad food should have (with the conk) one output of 0.7, therefore power absorptive on the 220v will be of 4/0.7 = 5.7watts. For 220v, that made a current of: 5.7w/ 220v = 26mA (milli Ampères) which does not induce parasites on wire of data.

The cablage 10/100Base-T and the cablage 100Base-T4.

Which solution to choose?

It is you which see!! Personally, I chose number 4 (minimal distance = 5m); it function. The terminal voltage of the AP is (in load) of 4.2v. For parasites, y' does not have any, according to my measurements which are (with it day) duration of the “ping”. I have values on average with 3ms. The players in network will appreciate!!
The “ping” is an order which I type on line, i.e. - in a window “terminal” for the unixiens, - in a window DOS for the windosiens. Syntax is:
with X to give address IP of the material wifi distant.
The order ping is used to know if a machine is present on Internet network.

Caution: the firewall (avoid-fires) can stop les orders “ping” (ICMP echo)


For the continuation, @ wants to say address.


It is what a hub (repeater), a switch, a bridge (bridge), a router?

All this equipment provides the function “modem cable " or cables cross.

the hub connects all electrically,
the bridge filters the @ip according to @ mac,
the switch is a bridge which moreover separates the types from networks materials (concept of bit),
the router separates the fields d'@ IP. The router must to be configured to function.

The MAC addresses it is of Ethernet, IP addresses it is of the Internet.

The MAC address is closely related to serial number of the material Ethernet network. I point out that WiFi is initially of Ethernet. For certain materials, it is possible of to change l'@MAC, therefore mistrust on safety based on filtering @MAC!!

The address IP (xx.xx.xx.xx) can be fixed or re-used (y' does not have any enough for everyone on Internet). L'@ IP is coded on 4 bytes, that made 2^ (4*8) = 4.294.967 296 (4 billion) addresses possible, but that is not enough.


How to choose between fixed or re-used address IP?

If you bought one @ fixes, it is single with world. By against those which one not bought fixed d'@ IP, and which in choose one randomly, can take of them one which has summer bought by someone else, and can, one day, to enter in conflict with this one. Of or interest to ask one of them re-used.
For example, the supplier of access Internet (FAI) gives some one different to each time one connects oneself with a modem 56K, or connection ADSL crosses every 24:00 to force reallocation of one @ IP.

Who buy IP addresses?

The FAI, the companies, that avoid to them to have time of calculation and of the flow network dedicated to part of re-use.

Addresses MAC, addresses IP re-used, how that goes?

The MAC address is coded on 6 bytes, that fact 48bits. 2^48 = 281.474.976 710.656 (280 thousand billion) possible addresses. This makes it possible to allot a single number to each material network.
One indicates by “bit" the whole of computers which are electrically connected (with a cable modem (2 ordi) or a hub (several ordi)) by Ethernet.

When one puts this material in the computer, this one goes to charge with finding one @ IP corresponding to the @MAC ( opposite correspondence (stored in a table) is made by protocol ARP). For that, it will ask all them computers same bit, by a request which speaks with everyone (broadcast in English). The computer with which one programmed this function will answer (in broadcast) the computer which has fact the request, by a value d'@IP. This fonction be called DHCP.
When a material wants to speak with a @IP, it will use count ARP to find l'@MAC, and to send information to him.

The computer which answers provides the function waiter DHCP, the ordi who request is customer DHCP.
Attention, if there are several computers which answer, it is fastest which gives l'@IP!!
The normal configuration is: Only one waiter DHCP by bit, or at all.

Not of waiter DHCP on a bit?

Function DHCP is hierarchical. Ca please say that one can to propagate this request (on different bits) (it is said that the protocol is “routable”) until finding one waiter DHCP.
The equipment which sees 2 bits must have 2 @MAC? bingo! it is one switch or a router.

Today the AP can make customer DHCP and waiter DHCP, or to have a fixed @IP. The AP sees 2 bits, the telegraphic network (with the catch RJ45) and the wireless networking (WiFi).

I observed that with the DHCP, I had disconnections WiFi, which was solved when I passed in IP statics (@IP allotted manually). This solution is not certainly not the good one. My material: WRT54G and DWL520G, éventellement with a USR2249 out of repeater.

ex_machina” provided an explanation which I like well:

“For the disconnection every 10 minutes: it is in general with the fact that the PC
is in “automatic” address IP, in other words the PC goes to ask an address IP
with waiter DHCP of the modem.

This waiter “lends to him” an address IP for a lease of 10 minutes.

Beyond that, the PC must Re-ask a new address IP.

To do that ports 67 should be left & 68 open in UDP (that is regulated in
your firewall) because if not the PC cannot require/receive sound addresses IP
and thus it cannot be connected any more to Internet => disconnection.

Under Win98, duration of the lease obtained post yourself with “winipcfg”.

The address Can IP, it be unspecified?

It @ held there: (“localhost”, which wants to say “myself”) with with with
who are interpreted by materials network to ensure function particular. For example, by convention, one give for one @ IP of a router.

@ the IP are selected in ranges of values, which are programmed in waiter DHCP.

@ the IP are sold by an organization (internic). To make economies, those which bought @ IP make use of it for to be identified on Internet, and, “mask” one sub-network behind these @ IP. Thus the sub-network can to reach Internet. Ca is called the NAT. I would speak again about it.

Yes, but I did not buy d'@IP me! , then?
Ben the supplier of access to Internet (FAI), him, bought @ IP.

It is what a mask of sub-network (“subnet mask”)?

It is something which expresses the size of under network and which is the range of its @ IP. Bytes (or bits) of the mask who are to 0, correspond to @ possible of under network, them bytes (or bits) to 1 correspond to what is not pas le under network (outside, Internet). The 1 is on the left, them 0 on the right
Example: the mask, expresses that under network can have 254 computers (the 0 and the 255 are reserved, thus 256-2=254)

There exists another notation of the mask, added with @ the IP: ww.xx.yy.zz/N. Here, NR means the number of bits with 1.
For the mask, NR is worth 24.

One applies the mask to l'@ destination, for to determine towards or to send the packages of bits.

The interest of the mask is to identify if one @ IP of destination is under network or not. If the destination is in under network (even bit), one sends it directly with the ordi recipient, if not one sends the request towards the ordi which is connected with Internet (footbridge).

How to identify the computer which is connected to Internet?

One affects one to him @ IP reserved which is made for that, this ordi is called footbridge or “gateway”. For example

To have one addresses IP, that is enough?

Not, to post this page, you did not type d'@ IP (well that functions), but “HTTP: //yves.maguer ......”, there is thus a correspondence which is made. This function be called DNS, it is ensured by a waiter DNS. Le waiter DHCP can also configure the values of DNS.

If a machine is present (reachable) on network with the order ping (said “pinguer”) by its @IP but not by its name, then there is a problem of DNS.

There are more qualified people who explain all that better that me.
Explanations with captures of screen (Windows 2000).

 Routers between FAI are localised in Paris. This increases the ping in province and limit unnecessarily it flow with Paris and especially the flow province-province.

Les Can AP go on the roofs without problem?

NOT, the AP are conceived to be used in interior (with this day). There are 2 problems with to solve: the bad weather (rain, wind) and the temperature. With the temperature, my AP, a USR2249, broke down, there was in the history: “AP card die ".
After placehaving placed it in front of a ventilator, it east went back to to function. I thus realized one turn of refroidissement placed on the roof, for he. Ca it also protect from the rain. I do not have fear for cold, because it consumes 4W and that is enough with to protect from cold when it functions.
The short period of omnidirectional aerial which exceeds large tube, is enough to sprinkle all the house (2 stages).
With this installation, the heat wave does not have it gene!

Good, and then I have a second USR2249, much more apprehensive, which goes badly even to 30°. I explain:
While launching navigator Internet to go to config. of the AP, in the mitre “status”, it sometimes posted there
Wireless MAC: 00-00-00-00-00-00

And there, it does not make any more connection with another material wifi. (the luminous witness of the medium does not flicker of way variable).

Solution: (do not laugh), you disconnect it to you put with the regrigerator during half an hour, you replug and it remarche until one cuts to him the 220v.

The AP WRT54G seem to be sensitive cold wet. Locked up in a box seals of electrician (example: Case PLEXO Legrand 920 62, 220 X 170 X 86 cm, approximately 20 euros), with at side or below the AP, one heating, regulating system of temperature (between 10 and 20°C) (thermo plunger of aquarium in driving position with a diode in series on food 220V (1N4003 with 1N4007), lamp of terrarium with one thermostat, heating cord with thermostat), it seems to me that would have to go. Attention external cables (especially in 220v) do not have to move under the influence of the wind.

According to the thermo isolation of the box, I think that a heating from 10 controlled Watt with 10°C would be enough. The AP heat already of 4 Watts constantly.

If not, certain AP conceived to function in outside start to appear at DLink. D'autres ici.

Les waves they is bad for health?

, it is clear! On the other hand it should be known that it vient some du cosmos since the ground (and the human one) exist.
Thus all is question of quantity. Quantity received (power) is conversely proportional to the square of the distance (surface of a sphere according to the ray: S=4*PI*r²). With equal surface, the more one is near, the more one collects waves.

I applied following calculation: radiation receipt = Power of the source/(4 * pi * distance * distance). For powers considered following:
relay television: 1000Watts,
transmitters FM: 1000Watts,
satellite: 1000Watts,
relay GSM: 100Watts,
telephone GSM: 2Watts,
microwawe oven (power which escapes): 0,2Watt
WiFi apparatus: 0,1Watt,
WiMax: 20Watts.


The fichier is disponible, if you want to test values.

(Is GSM dangerous? Try to know where I call from?)

Consequently, I do not believe in validity of this study who would like to reassure us. Because I doubt that this one measured with 1cm (distance antenna/brain) it radiation of each one of our telephones portables. Miss of independence? (Of the testers of cellphones have have problems of memory (memory in the short run; or I have parked my car?)).
Download the passage TV of Richard Forget at Laurent Ruquier  (“It black file of the portable ", that made 32méga, but that the blow is worth!! that makes me think of the share media which disputed the warming ( half) whereas no scientific study it disputed. (See “Une truth which disturbs“, Al Gore).
When adorable Virginia asks Richard Forget if it WiFi it is similar, it answers “yes”. I him rétorquerais that one is not stuck the access point to the brain… (see table comparative above).
Harmfulness comes especially des telephones themselves, less of the antennas which are on roofs.
The proof, this study which shows that that break them bits of DNA.

the WiMax terminals make 20 Watts of EIRP,
the transmitters FM make 1000 Watts,
relays GSM make 100 Watts (and there are people with whom one has nothing asked - study in Marseilles, qui live just opposite),
The relays GSM become UMTS relays without respect of procedures.
the satellites make 1000 Watts (but they are loins, with 36000km),
telephones GSM make 1 to 2 Watts (and one them stick to cerveau) (use the auricles or the textos). More it telephone is far from relay GSM (in campagne) more it emits extremely ( power is adjusted).
the wifi makes 0,1 Watt,
my microwawe oven (with same frequency) disturbs my connections wifi in the house, therefore the furnace “sprinkles” also the house, in spite of its shielding. Old man or nine, it is similar. Therefore I have him affected a power (0,2Watt for the escapes) higher than WiFi.
I do not have information on the téléphones without wire that one place systematically with the ear. However, consider that the terminal (bases) with which they dialogue, is close (10m). Thus the need for puissance to join the terminal (bases) is faible (compared to the telephone GSM, for which the terminal (relay GSM) is with 100m or 1Km or more).

So that a WiFi connection functions, it is enough that received power is of -60dBm, i.e. 0.000001 MW (1E-09 on the graph), or thousandths of micro Watt.

The ARCEP with ordered a study in Supelec on power of WiFi; result WiFi is well in dessous of authorized powers (decree n° 2002-775 from May 3, 2002). Here the study (12/2006).

A cellphone under continuous operation posed on eggs out of incubator, produces 6 times more death at chicks (considering on TV).
This stronger sensitivity for the organizations in the course of development (our children), seem to be confirmed here.

It would seem that the ondes GSM disorientate the bees...

A chicken stockbreeder, saw his mortality to increase suddenly with the appearance of an antenna GSM with proximity of sound breeding.

It is notorious that children of people who work in the medium of ultra high frequencies, are often GIRLS.

Influence on the eyes of the waves emitted by cellphones:
- The cornea exposed during a time prolonged with this radiation presents macroscopic damage affecting quality of the vision. The damage increases with accumulation exposures and reached a culminating point at the end of some days (192 cycles of exposure). As of suspension of the exposure, the damage tends to regress.
- On the microscopic level, other types of damage are noted. Small bubbles are created on the surface of the cornea. They are directly related to the beam microwaves and are pas la consequence of heat generated in the cornea. However, contrary to those caused by the increase in heat, these damage is irreversible.

In any case, as long as one will not have, in France, of information exhaustive on health individuals, one will not be able to extrapolate what occurs for 62 million people in some month or years, with consequences more serious for thousands of people 40 years afterwards. See asbestos (100 000 died to come), radioactivity, etc…
One beginning of statistics with Cmap.
Unless the statistiques ones from the individus themselves do not come? (

One tenant who known as to suffer from electro-over-sensitiveness bound to the antenna-relays, in justice its financial backer attacks.

And the role of the frequency? Ben, more one bring closer to the frequency of resonance of water (70 _ of our body) more one heats our body. This is typic WiFi telephones.

Dilemma, one is today with 50V/m for the mobile phones. According to Denis Zmirou, (ex scientific director of the AFSSE) with 0.6V/m one could more to have modern communications (include/understand with a flow bytes of disont (that engages only me)… 1méga bits by second).

I benefit from it to recall that the réseaux ones with a grid use less power (to voir Femtocell) that “dorsal” structures, since one does not state that with closest (more easy to join).

And I give a layer from there: antennes directionnelles, precisely, sprinkle only that towards which they point. Thus each one is not pas sprinkled by its own material, but by that of the other. I exaggerate a little because there are lobes parasites, but in the principle it is that!

A proposition of loi (several parties) is in hand (09/2005).

There is WiFi in the hospitals...

How to detect the radio waves?

For 20 euros I am bought a multimeter who makes measurement of capacity. On gauge PF (pico Farad), in placing a wire of 3cm on an entry, I detect it radiation with 1cm:

I have say “I detect”, I do not have say”I measure".
Here the field intensity measuring device (or field-meter, or wavemeter) that I manufactured (thank you F1SXO to have provided me the diode IN21C)

The value read is of 100mV to a few centimetres antenna omni of my USR2249 or my WRT54G (with a “beacon interval " of 1ms).
With un meter, at equal distance from the 2 antennas (the maximum), it remain 10mV.

How to put the mini-PCI of D-LINK DWL520+ in a toshiba satellite?


The same open one:

One tears off (or one unsolders) the cable side mini-PCI. Under PC, one finds 2 wire (white and black) with connectors. These wire make will 800microns diameter, and they are the coaxial ones which go on a connector U.FL:

The white wire is cut and one welds it in the place of that which one removed. Heuu… it is not with carried everyone, parkinsonian, to abstain from!! (I acknowledge being with limit of what I can make in the field) I point out that it is necessary to weld it shielding (with the mass, one of the 4 holes in square) and the heart of coaxial (with the central hole).
Here what that gives:

One will take care to fix the black wire (with adhesive tape) to avoid risks. This white wire is connected to an antenna inserted in the frame of the PC. It gives very good performances, better than a pcmcia.

I very did not include/understand but there are charts which have already connectors U.FL, all is then simpler!

How to add a connector on D-LINK DWL-650+?

D-LINK DWL-650+ seen of top and lower part:

The chart opens while cutting the plastic with the cutter, on the edges. Then, 2 metal stems on sides slide slightly and the 2 hulls release.
I placed connector (female SMA at to weld) (more running is the female RP-SMA) with horse on the circuit printed, in manner to use the plan of mass to connect the mass and to ensure la mechanical rigidity of the connector. I have connected the central stud connector (on other side of the circuit printed) with the hole which is in the middle of the formed antenna of 2 spirals.
To weld the mass, it is necessary to scrape green varnish on 1cm square so that all the length of the 2 studs of the connector, is welded (better mechanical resistance). Here it detail side welding of the central stud, I used a leg of component like wire:


The same principle was applied on a DWL-G650, except that they welded directly the cable on the chart.
Others put petites omnidirectional directly on the chart.

Me I would advise rather, for reasons of resistance mechanics, to do that:

by putting a wire for the red feature, by welding the connector at the blue points and while crossing, possibly, with the feature purple.

With WiFi is what I would be pirated?

We are 6 billion on ground, some pourcents has a computer on Internet, among these, 1 out of 10 thousand is malicious pirate which want you evil. This made that there are a few tens of attacks per day on your computer if it is always connected to Internet. Good.
You put of WiFi in your computer, to date, you go to find some other computers on WiFi. Ben me I you say that there are few chances that your “neighbor” is one unpleasant, malicious pirate. Thus to connect itself to Internet is much more dangerous than to make a WiFi network.
And risks it for the attacker is not the same one, if it is to 10 minutes with foot, that if it is at 10 hours of plane.

For information, if you have one repertory divided under Windows, you give the means to the attacker (given) to erase your hard drive, via Internet. A fire wall (firewall) is a protection. You can put the fire wall for you to protect WiFi network.
I use “tiny personal firewall 2.0.13”.

In addition, if you use antennas one-way, le pirate will have is to be very near you (because of the parasitic radiation around your antenna), is to be in the beam of the waves (which will be with plusieures tens of meters height and which fact some meters of ray), is to be very close to the point distant.

On another side, those which are interested in new technologies are of a curious spirit. And with it titrate, those which could make a connection wifi with you, have one atypical profile…

Some sofware ( EtherPEG (MAC) Driftnet (LINUX/SOLARIS) ) allow to voir directly with the screen, the images which pass on a bond WiFi (if it is not protected). Other possibilities under Windows.

There exist CD bootables (Knoopix/Unix) which include tools for testing yourself secutity of your network.
A total solution at Immunity. (08/2006)

The pirates (or apprentice-pirates) attack the maillon le more faible of a system. Today it is it WiFi (not for technical reason, but by ignorance utilisateurs, because majority of the points access does not have or little protection).
There exist also failles WiFi in MAC (Apple)… not? … if, if… (correction Extreme Airport)
The pilot Linux MadWiFi is vulnerable to the attacks (04/2007)

The filtrage of MAC addresses does not protect you, since that one can simulate (one “spoofer says”) any addresses MAC with certain materials. And like in addition, the address MAC of your AP (and those of the computers which connect above) is diffused by the AP, the pirate lays out of all elements to be made pass for one computer of your network.

The WEP is not worth any more large chose today, because it is enough the quelques ones minutes of listening (secondes to the 08/2006) to break a key of 128bits, with one flow of data sufficient. Les outils d'attaque (wardriving) se nomment kismet, airsnort, aircrack weplab, wepattack, et certains combinent une attaque brute et une attaque par dictionnaire. Voici une liste de logiciels.
au not to deserve the nickname sometimes of Weak Encryption Protocol" ;-)

Certain people believe themselves in the shelter pirates in ne not diffusing (broadcast) the name of their network (the SSID). It is a weak protection, since it is enough to attaque by dictionary on SSID to find it.
In the same way, the fact of not diffusing the SSID, do not hide your network of the pirates. The pirate sees that you emit, and he sees that you hide your SSID. (provocation?)

attaques the passives (mode monitoring) do nothing but listen to your AP, it thus does not have there strictly none trace your side, of these attacks. I personally capté a AP normal (without modification of the antennas) placé on one balcony, with 10Km.
By definition, these attacks do not do anything on your machine, they aim only to to capture your data.

The attaques active ones, oblige le pirate to be emitted (for example for to make allot one @ IP according to its @ MAC). Thus, DHCP can help the pirate.
These attacks can corrupt your data or your system, but they are detectable.

I know people, whose pleasure is to go to see in repertories shared (by Windows) of Net surfers (thus even without WiFi) which does not put any protection.

And then a key of 8 bits (a character) breaks into 255 tests (attacks rough), therefore if you put a password too much short, it will be always fragile.
In the same way, if you choose a word of passes made up current words, an attack “by dictionary” will find easily this word of passe, some is your mechanism of encryption.
Against this, use characters generated by chance. La difficulty is to memorize them!!
Keys RSA of 1024 bits should be used until 2010 and keys RSA of 2048 bits until 2030, to preserve a good protection.

Let me be delirious a little.: I calculate a value N=P*Q (P and Q two prime numbers of at least 150 digits)
To find P and Q, knowing NR, is aujoud' today difficult.
If a key (NR) is that which I calculated, I know P and Q instantaneously.
Probability that an unspecified key, is that that I calculated is weak, but it exists.
Now, with a large machine (or several distributed) and a base of data in memory flash, I calculate des billion values NR and I memorize P and Q.
Probability that I find a key instantaneously (research in my data base on criterion length or figures of left of NR) increases…
Especially if it is known that one should not use P or Q with less 150 digits under penalty of brittleness with methods traditional…
My database will be consultable on line for one moderate … hein summons?? that made fear? not, I am delirious!! (05/2007)

Microsoft found a solution (WCN) key USB. It is why ZyXel propose material with a catch USB.

The Wi-Fi Alliance propose (at the beginning of 2007) it WPS, for to simplify the setting in work of encoding in Wi-Fi.

For that which can touch the material physically, seizure of a simple code PINE or even support on a button.
For that which has access only by the waves, gross key of the WPA2.

Another solution, with the test (03/2007), with SIM chart, at SFR with Defense.

I quote (in 2003):

All in all, per hour of today, a wireless networking in 802.11b presents only little of safety and little to be easily pirated, this is why networks without wire are still a danger to the companies. Of course, it is possible to add a safety higher around with in particular the use of VPN/Ipsec, of the standard 802.1x, Firewall, but well little companies are aware of the dangers that they take by installing this kind of material and do not take thus not these precautions.

Fortunately a “substitute” of the WEP is being studied, it acts of the WPA (WiFi Protected Access). Its operation is based on one system of exchange of keys dynamic, renewed every 10 KB of data, this process is called TKIP¨ (Temporal Key Integrity Protocol). According to the manufacturers, it will be possible of upgrader Firmware of your material 802.11b to integrate the WPA. This new process should improve appreciably the safety of the wireless networkings, only flat, the algorithm used remains unchanged. source, to see 802.11i for the evolution of the WPA.
Various apparatuses supportants le WPA2.
A mise up to date of the WPA2 for Windows XP (The patch: KB893357).

The WPA2 is today (03/2005) the best protection in WiFi, good. L'Alliance Wi-Fi makes compulsory the WPA2.
With la grand' mass of the pirates and hackers, (22nd European Hacker Conference at the end of 2005 with Berlin, 22C3, Chaos Communication Congress) 3845 attacks did not come with boils of the WiFi network of Aruba, and that during 4 hours. A first!! Idem en mesh.

WPA2 = 802.11i = protocol CCMP + authentification 802.1x (or key divided) + coding AES
WPA = protocol TKIP + authentification 802.1x (or divided key) + coding RC4
WEP = protocol WEP + optional authentification + coding RC4

But, it would come you to the idea from you to connect to Internet without fire wall or antivirus?? not (I hope for it).
You know that today a “contaminated” computer by a virus, spyware, (and other tricks in “.ware ") can diffuse (without WiFi) with its author what you have typed with keyboard, for example? (the worm is in apple)(les banks react).

Thus for WiFi, as Internet, meilleure safety is that which you mettez in place yourselves (WiFiRadis).
Do not offer to the pirates WiFi, the protected part (LAN)(by fire wall) of Internet.
I thus share the point of view of ozone, which crypt not sound network WiFi in the big cities. However, their access Internet is constituted of a dorsal which does not seem accessible to private individual (5GHz). See attack “man-in-the-middle”, le pirate is made pass for equipment which normally you connect to Internet.

To avoid that, CISCO, put in its AP and theirs customers of functions of monitoring (WLSE) who listen to and defer the arrival again WiFi material in the coverage area of those. This function of listening also allows to locate (with a minimum of apparatuses equipped) the materials non-WiFi, with an aim of removing them interferences.
Thanks to a centralized administration, one can then adjust powers and to affect the channels to optimize radio coverage.

Attention with the traps, when you surfez with your PC WiFi portable in the street, of badly intentioneds people install “hotspots hooligans" (points pirate accesses) which imitates them truths, but which enables them to collect your data confidential of connection.
Thus do not give your Bleue card number for you to connect to a hotspot, buy a chart rather prepaid.

To hide its data, it there the VPN has (growth of 23 _ ventes in 2004) (idem in 2006)

The point of sight of Befti (Brigade of investigations into the frauds with information technologies)(05/2006)

the VPN, la solution for the safety of WiFi?

the VPN it is what?

Google propose VPN for its hotspots WiFi.

a page of entreprise on WiFi.

a réseau semi-public, convey by Internet; completely hermetic Internet has

a “upgrade” for IPSec on OS W9x and NT4

With OAPI, Trapezoid wants to centralize the management of the access points. According to me, if safety is not with go, it is the door opened with virus WiFi (WCS). Faults and patchs at Wintel (Windows+Intel) on the wireless telegraphy. (charts PRO/Wireless 2100, 2200BG, 2915ABG and 3945ABG) and their software of administration PROset. See the software Bad_Seed.

La safety gives reason to the purists, thus OpenBSD (UNIX), which does not use pilots whose only binary ones are available, is not concerned with this kind of fault.

If you want to sort the nice ones users of your WiFi terminal, bad, it is enough for you to their donner means to join to you.
You can do it while giving as nom on your WiFi terminal
(field SSID) , your email.
Thus, any user of your terminal which will not have sought with you to join, will be equipped with bad intentions.

It is necessary for you to activate the option which consists in diffusing this name (broadcast SSID = yes).
The frequency to which is diffused this name (beacon interval) is programmable (of 1ms at one minute, in general).
A too low value will collapse your flow.
A too high value will return to you statistically less visible occasional users, while being visible by your neighbor, who is to him longer with the listening of your terminal.

Under these conditions vous can observer, with AirSnare (downloadable here) (which uses if necessary Ethereal)
les users who connect themselves to your access point (some that is to say their OS (Operating System)), and their to send messages (if you and they, use Windows NT, 2000 and XP (order “Net send”)), oubien to prevent the guards, vigils, by email automatically.

Pour aller plus loin, vous pouvez localiser le pirate
(by triangulation). You must be equipped with an antenna directional and to collect it since 2 geographical points. With each point, you raise the azimuth of the WiFi connection of the pirate and coordinates GPS of your position. By chart geographical or by software, you deduce the coordinates from them GPS of the pirate. The 2 points should not be in the same axis that it pirate, ideally the 2 azimuths are perpendicular.
If the pirate uses an one-way antenna, it you will be necessary to approach you him until you find one azimuth from which the angle is sufficiently different compared to first. Indeed, a directional antenna always emits one little around, one calls it that of the parasitic lobes. If it is in a building, you can change the polarization of the antenna, to locate the stage more easily.
The more directional your antenna is, the more the localization is easy.

A manufacturer (Intel) exploited the travel time of the radio wave for to determine the distance to one or more AP.
In the case of an only AP, one can thus thus use the distance with the AP to prohibit the access to a.
In the case of several fixed AP, one can inform each AP on its own physical coordinates (coordinates GPS).
When a mobile computer is connected to the AP, it will be able to deduce its physical coordinates by triangulation. According to la precision of the system, it becomes possible then for the system of security management, not to authorize connection of mobile, that for certain geographical areas data.

One article on the safety of WiMAX (01/2007)

How to improve a WiFi connection?

A WiFi bond is a bond bidirectional (by opposition for example, with a radio operator bond between one transmitter TV and your TV).
It is necessary thus that the connection between you and the other is done, AND, the connection in the other direction too.
Here a diagram of a connection in only one direction:

The units in MW and dBm are absolute, them units in dB and dBi is relative. One can add one unit in dBm, with plusieures units in dBi or dB together.

An antenna has strictly same properties in emission that in reception. Only one measurement characterizes amplification with the emission (profit) as with the reception (profit or “sensitivity” according to some) it is about measurement of the dB or dBi (profit in dBi = profit in dB + the 2.15) (salesmen pushed us with to use the highest unit; -)
What I call a “large antenna” has a profit raised, a “small antenna” has a weak profit.

The legality of the installation is measured in “AIR”, on the public highway.

In this diagram, 1 is the transmitter, 2 the receiver. The profits are green, the losses oranges.

PUI1: power (in MW or dBm) of the material WiFi (AP ~50mW, chart NCV, PCMCIA ~30 with 200mW, key USB, case USB, etc…),
AMP1: power of a possible amplifier (dB),
CON1, CON2: losses (in dB) in the connectors of each side,
CAB1, CAB2: losses (in dB) in the cables of each side,
ANT1, ANT2: profits (in dBi or dB) of the antennas,
AIR: they is the losses (in dB) in the air or the obstacles,
SEU2: threshold (in dBm) of the receiving material, varies according to quality (disont -94dBm for a flow of 1Mb/s for material running),

MAR1>2: it is the margin of remaining power. If it is positive, the connection goes in the direction 1 towards 2.

There, one made half of the way. For that the Wifi bond function, it that connection 2 towards 1 walk too.
One realizes that there is a part common to the 2 connections. It acts of: CON1, CON2, CAB1, CAB2, ANT1, ANT2, AIR.
Ca please say that while acting on these elements one acts on TWO radio connections (1 towards 2 and 2 towards 1).
Ca please as say, as while acting on elements PUI1, PUI2, AMP1, AMP2, SEU1, SEU2, one acts that on half of problem.

In the same way, the use of one WiFi preamplifier allows to remove the losses in CAB2.
Operation is the same one as that of préamplis for TV or the satellite, one raises the signal there or it is it more extremely, therefore close to the antenna.
The difference (and the price) between a préampli WiFi and those caused, is that the préampli WiFi must to commutate very quickly for to let pass the signal vers the antenne when the receiver becomes transmitting.
When the receiver becomes transmitting, losses in CAB2 (which become CAB1) can be compensated in increasing power of transmitter PUI2.
To decide use of a préampli, compare its price with the price difference of the cable (good or less good) for length which you envisage to use.
In short, if you want to put one cable very long, put a préampli. You will be able, then, to allow you to use a cable of ordinary quality, i.e. with more dB with the meter, thus less expensive.

Attention with the abuses language, certains call amplificateur, an amplifier of WiFi range. These apparatuses make simply repeater. One makes twice more distance with 3 apparatuses (that of the medium makes relay (DWL-G710)) that with two. It is all.
Many of other AP have already this function for a long time…

In short, the assessment is:
XX = PUI1 + AMP1 + SEU2,
YY = PUI2 + AMP2 + SEU1,
ZZ = minimum of XX and YY,
Total MARGIN = ZZ + CON1 + CON2 + CAB1 + CAB2 + ANT1 + ANT2 + AIR

According to the value of “total margin”, there will be a flow of 1Mb/s, 2Mb/s, 5Mb/s, etc…

It is this minimum which does not seem to be assimilated by the WiMax operators, like the known as one Stephan, without tweezers; -)

Now that you very included/understood, I muddle you!

Joker!!! For the materials which have 2 exits of antenna:
If one has access to the low layers of software which controls the hardware, one can use an exit for the transmission and the other for the reception. And there, you known as “why make me”?

And there I answer you to try to respect the law (current). Explanation:

If I emit slightly (ANT1) (limit of 100mW), but that I have a large reception antenna (ANT2), I would collect better that with (small) the antennas of origin (ANT1 and ANT2).
Thus on each side of the bond, if I have an antenna small for the emission (the antenna of origin) and large for the reception, I could further carry that with the antennas (small) of origin, while respecting the law: -)

Of course, one will carry less far than if there is one antenna large in emission and large in reception. But I have when even made 10km (see page Toulouse) with antennas of origin of my AP usr2249.
The disadvantage, it is that costs 2 large antennas with the place of one, to make the same range.

Sylvain implemented this solution on 6Km:

Second joker!!! For materials on which one can program the power of exit:
This idea (of _SebF, D. Lebrun) is so good that I have swell not to have found it.
To respect the law and to use only one antenna in emission and in reception, it is enough to program a power of emission weak.

Example: my material has a power (PUI1) of 17dBm, with the antenna of origin (3dBi) I emit with 17 + 3 = 20dBi (=100mW of EIRP).
I put there an antenna SD27 of 24dBi instead of the antenna of origin from 3 dBi.
To return to 20dBm, I must remove the surplus of power of the antenna (24 - 3 = 21) with my power of emission 17dBm.
Thus I must emit to 17 - 21 = -4dBm = 0.4mW.

Here some examples of range theoretical for 2 antennas identical:

One can to make a bad antenna “ricoré”?

Yes, of course! It is very easy, it is enough to consult certain Internet sites (Paris-Without-Wire for example) to find bad councils which will allow to manufacture the bad ones “ricoré”: - (

Ricorés of Paris-Without-Wire make 8dBi, one good ricoré makes 11dBI.
To measure the dBi, one compares (with netstumbler for example) measurements with those carried out on an antenna trade, whose profit is known. In Toulouse one compared in a SD15 (13dBi). See also it video l' antenna ricoré for the null ones" ;-)

Ok, let us see rather how to make good:

For one ricoré, Lg/4 = 44mm and Lo/4 = 31 Misters.

With to leave of these diagrams it is necessary to interpret outdistance “Lo/4” (31mm) as being defined between boils antenna and:
- is it edge (in bottom) of limps (if the connector does not exceed) (see photograph of my ricoré in the paragraph How a link long distance is established?),
- is it shielding metal of connector (if the connector exceeds inside limps) (photo below).

The 31mm correspond to the central part of the driver (the heart) who is not surrounded by metal, and which, by consequent, lets escape the radio waves. It is the partie radiant.

The dimension of 31mm is critical, it must be realized with +-0.5mm.
This east coast related to the selected channel. 31mm correspond to channel of the medium.
This coast also makes it possible to adjust the impedance of the antenna. The impedance must make 50 ohm so that the antenna propagates all the energy sent by the WiFi material (of which impedance of exit makes also 50 ohm, it is standardized).
Look at l'explication in small “the Antennas - lines of transmission”, “the ROS”. (This site is a bible for me!).
For the same reason, you must use cable 50 ohm (the green cables make 75 ohm).
The impedance of the cables is determined by 1) the report/ratio enters - the diameter external of the cable exchange, and - the diameter interior of the cable which surrounds it (shielding, braid), and 2) it coefficient of swiftness of material insulator.

“Lg” is directly related to D, it diameter of limp. For one ricoré, D = 100mm and Lg = 175mm.

Also note that the distance “Lg/4” (44mm) is defined compared to the bottom (with interior) limps.

One can also cross limps it to respect the distance “3/4 Lg” (131mm) but it is less serious.
“It” was also said to me that 1dB was gained if one removed (with a can-opener) the small interior disc with dimensions lid.
There is no problem to leave the lid, it do not heat if one even puts it at the microwaves ( frequency with 2.4GHz).
One can check the coasts (31mm and 44mm) with fabulous instrument below, which I named it “gabaricoré”:

The “caroparty” consists in meeting around beers, music and limp of “caro” or “ricoré” for to manufacture antennas.
Thus not of “caroparty” without “gabaricoré”!!

There is T it reflectors on antennas grids SD27?

Certain antennas grids (SD27) are delivered without reflectors, 3dB then is lost. This one has as dimensions 105 X 77mm.
This sheet contains 2 foldings. There are thus 3 parts of 105 X 27, 105 X 23,105 X 27. The angle made 30° approximately.
Once folded the coast of 77mm becomes 65mm.
There are slits of 4mm broad to thread the ends of figure of the SD27 envisaged for this purpose.
Holes with 13mm and 76mm of the edge are at the end of the slits, for to make pass the plastic ends before threading them in slits.
An end of zinc goutière, cut with scissors, bored with the drilling machine, folded with a board, kitchen table and one screw clamp.
And here is the trick:

Why my compass does not work well?

The needle of the compass is aligned with the magnetic field terrestrial. Thus in full nature ya not of problem.
We use the compass beside one mast of antenna, one antenna. Those are out of metal and they deviate them magnetic fields and distort the indication of the compass.

To solve it problem, there is a method:
I place myself at least a meter of the mast (and of any metal part) between the point that I aim and the mast of antenna.
I point the required angle and I locate at the horizon what is in the corresponding angle.
I arrange my compass and I place myself behind mast. I turn my antenna in direction of the visual reference mark with the horizon.
And here!

If you do not believe me, you check: you places behind the mast with 1 measure, you aim at the visual reference mark and you approach the mast. With 10cm of the mast, the indication of the angle moved of 10°. Perhaps that that does not make it if one aim at north, I checked for an angle of 45°.


<>But the lightning?

What, the lightning? It is already fallen on your antenna TV? Not, Ben on mine either ..... good day! ;-)
Easy to say when one lives downtown.
The lightning makes more damage in the countryside than in city, perhaps because of the cables which are more exposed, or perhaps parce that there are more lightning conductors downtown than in shift… I do not know.

There exists
lightning protector (or surge protectors) for WiFi which semblent to be effective in the absence of impact direct (when the lightning falls on the WiFi antenna).
I quote Didier Lebrun:
“Has Vaour, I observed that the addition of lightning protectors fact of losing about 1 dB in summer. In winter, I remove them lightning protectors, because I noticed that they block completely the signal when it freezes and accentuate the losses per time cold and wet. “

One disconnects the computer to avoid them overpressures on the sector. Ca it is the radical solution, but the onduleurs seem to be effective against which had overpressures with the lightning.

In shift one disconnects also the wire of telephone.

I await experience feedbacks on the use antennas in short-circuit continu (like the SD27 (0.7 ohm)), compared to the antennas in circuit opened continu (ricoré and the omnidirectional ones (infinite continuous impedance)).
My subjacent idea is that antennas in short-circuit continu has more chances of protéger the WiFi material, since the power due to the lightning can be on directly in mass.

Following the meeting (very sympathetic) at Jean the 28/07/2006, it was to me confirmed that antennas “short-circuit” protect the WiFi apparatuses which are behind (in general of the WRT54).
The day consisted with, in one the first time, to repair WRT54, by removing the component (the diversity switch) which has roasted by the lightning or the loads electrostatic (ESD).
In the second time, one protected WRT54 while placing one coil (22 nanoHenry) on the exit. The coil causes of court-circuiter low frequency currents and thus with fortiori, loads electrostatic which could accumulate.
manipulations sont delicate, but accessible to the people meticulous persons.

To this occasion I placed one short-circuit on an antenna omnidirectionnelle (of chez IRISAGRI), so that it is protected like SDxx.
Since I had brought my TOSmètre, it was easy to check conservation of the impédence with 2.4GHz!
Here thus the result

Of course, we confirmed that its performances initial were preserved, by a comparative measurement in dBm (netstumbler).

Here interior of the head of the SD27, thank you Julien for Grenoble.

Yes, I know that appears strange it short-circuit… there is even that prevents from sleeping; -)
, Make you with the idea go that the signal high frequency propagate on the surface of the drivers…

It is enough to look at the lightning coming to make its provisions!


My antennas are well granted?

While measuring (software WLANexpert, thank you “black Milan”) the provided current by the exit of the material wifi, one can deduce the Rate of Waves Stationary (TOS or ROS (Report/ratio of Standing waves) or VSWR).
This relates to only the materials which measure this current. Les materials containing PRISM2 allow it (
usr 2450).

Another implementation of WLANexpert.

Another way of to measure the ROS, it is of to make a tosmeter WiFi (ROS-meter wifi, tosmetre wifi) (band SHF).
I have
carried out this one (thank you Erwin), the BAT62-03W is available at Radiospares.
From the PCB, I printed and stuck on copper. I have then marked with a screw and a hammer points remarkable (corners). Then I took off it paper, it me remained the marks then on copper. I crossed to cutter to remove copper.

I welded components CMS (Component Assembled on the Surface):

You will notice the 100 Ohm (marked 101,10 with 1 zero multiplicateur=100) welded in parallel for to make 50 Ohm. One all that in one puts limps, and here:

My wrt54g emits according to a variable power.
A first solution consists in putting it parameter “SSID broadcast " with ONE, and the parameter “beacon interval " to 1 (ms) (minimum).
A second solution consists with to transfer a large file to force the emission to the maximum.
I validated the operation of the tosmeter, in placing in front ricoré iron plate, then by removing it: that influences indeed on the ROS (normal, one re-examines the waves with the shipper) (value of 16 master key with 30mV and value of 400 master key with 900mV). The court circuit on the exit gives a maximum value of 1.7volt then that under normal functioning one with 400mV (with my ricoré).
It now remains to me to calibrate it with - is antennas of trade, - is references standard ROS in 50 Ohm.

Hard, hard calibration!!
Thank you in Laurent Mennecart (the purist of HF) to have me provides standards 50 Ohm Radiall.
Initially I twisted the wire which go to voltmeter (that avoids to distort measurements when one puts the hand on wire).
With the standard, the value must make 0mV, it was not the case. With 2 resistances 100 Ohm CMS in parallel (welded in a standard connector NR), I had something well near to 0. Morality? the CMS in parallel do not make 50 Ohm with 2.4GHz!!
I thus tried to have 0mV with the standard.

For that, I welded a leg of component on a point medium (the branch of reference) to which I welded a disc. I then folded back the disc to bring it closer of a plan of mass. Super, I approached the 0mV, but without to reach. I arrived at 13mV minimum.

I thus measured;
- the reference gives 13mV,
- a SD27 gives 40mV,
- my ricoré 50mV gives,
- my 50 Ohm CMS give 60mV,
- adapters of the type NR towards BNC with nothing with the end give 100mV,
- same adapters with a stopper 50 Ohm of Ethernet network coaxial, 60 give mV,
- an open circuit gives 220mV,
- a short-circuit gives 240mV,

<>Attention them “stoppers BNC” well-known of (old man?) administrators network do not make 50 Ohm with 2.4GHz!! (see above)
Also think that in diversity mode, if an antenna (that which one measure) walk badly, the apparatus will make use of the other (finally it is what I noted while disconnecting second).
I can thus assemble the values of the ROS according to the tension read with the millivoltmeter:

Good, then one should not panic, considering the percentages of losses of power correspondents to the values of ROS:
4 _ lost for ROS= 1,5,
10 _ lost for ROS= 2,
25 _ lost for ROS= 3,
50 _ lost for ROS=6.
A ROS of 1,5 is the limit that fix themselves, in general, them manufacturers of antennas. (VSWR <1.5 for all the channels)

By principle (bridge of Wien, I believe) only 1 quarter of power of entry, passed to exit (loss of 6dB).

WiFi, that grow bigger well?

Ben that depends, the anarchistic wifi grows bigger badly (with the OJ of Athens, “It is then necessary to transmit the image to the newspapers…“, they drew from the cables in the night for to circumvent the saturation of the waves).

Of another side, it 802.11s, or it 802.15 allow to extend the cover of the network with each arrival of a new user. It is what one envisages of to make in Toulouse, but with a mixture of fixed stations and stations mobile.

Our intuitions seem to be confirmed by this article of Jim Thompson, following this debate.

Then I put myself the question of the bond between number of jumps and the number of connected nodes. As I do not have found the formula to calculate it, I started with hand, for a network with a grid with 3 bonds by node. Ca makes a standard structure “nest of bee ". I colored with 3 colors, on the basis of point central, black:

Attention, put aside the unrealistic case or the ground is dish, that there is no obstacle and or the nodes are distributed geometrically, this drawing is not one geographical representation, but one schéma bonds, similar to those that I dessinais for Toulouse, when one was not not many.

According to the number of jumps, the number of nodes evolve/move as follows: (série1)
0,3,9,18,30, etc.
Sometimes one falls down on a point already colored. The number of already colored points evolves/moves as follows: (série2)
0,0,0,3,3,6,6,9,9,12,12,15,…, after I have decree!
Good, then I found that the difference enters values of série1, increased as follows: (série3)
3,6,9,12,15,18, etc…, it is nevertheless more sympathetic nerve! , there are +3 between each!
I thus coded a program which enabled me to continue (I am large lazy), here is the result:

I coded (red curve) X with the square (X ** 2) just to compare.
The white curve (of the number of connected nodes) evolve/move more quickly that the square.
It corresponds (with a margin of few things) with  there = 1.51 * X ** 2 .
Following an email of Jean-Jacques Gaudoz, the exact formula is: there = 3/2x² + 3/2x is  there = 1.5* X * (X + 1)
Which become les then limits of a mesh network of it type?
The number of jumps, will determine the travel time of one not with the other of the network (lasted of “ping”).
If it is considered that the WiFi networks which distribute one geostationary connection of satellite, have a ping with Internet of 500ms minimum, and that for the play, or the voice on line, it is a limit, then I will take the same value, 500ms.
Ca will change when constellations of 48 satellites are launched in low orbit with 400Km of altitude (instead of 36.000Km), but one is not there yet…

In addition, the AP (of today) bring a delay of disont… 5ms by node.
Before calculating, it should be seen that to join “with dimensions” network with the other, the double should be traversed, number of jumps necessary to go from the medium (black spot) until one “with dimensions”.

Thus 500ms divided by 5ms makes 100. One divides by 2 for to go from one end to another of the network, which makes 50 jumps.
One sees (!) on the curve that for 50 jumps, one can connect 3825 nodes.

It is of course one higth limit, since a structure was taken of network almost ideal and a maximum ping.

Obviously, it would be ridiculous to give a connection Internet single at the central point, because:
- the first three bonds would be saturated,
- the flows available by user too would be weak.

On the other hand, to distribute Internet in several points of it network seem to me to be a good project.
In addition, one could connect the central points (black spot) by a “dorsal” (makes a second network of it (with a grid?)) to constitute a network of networks with a grid. The dorsal then inevitably would consist of bonds more long (goégraphiquement) that those of the basic mesh network.

Bonds geographically longer, can to be obtained, for example, by the connections towards Internet (footbridges). It is the solution adopted by “Linux PC Services " (LPS). One can then choose the way more short (of many jumps) between the mesh network only, or, the mesh network/footbridge Internet 1/ footbridge Internet 2/the mesh network. LPS did it choice to pass systematically by Internet if the way in mesh network fact more than 4 jumps. I call this solution of “pseudo-mesh”, since one does not pass by the network mesh in ALL the cases, to join 2 elements of the network mesh.
LPS uses AODV which “flooded” its entourage; it is thus logical to seek to leave the network with a grid (by footbridges) as soon as possible, not to saturate it.
M@dBox of LPS is equipped (04/2005) with antennas omnidirectional, which slows down its use with countryside.

I do not know the detail of solution of Luceor which it, uses OLSR.

Thomas, in Pomeys, provides us the duration of “ping” in real-time, on its network.

How to find itself there in connectors?

Here a correspondance between the WiFi material (your router (AP) or chart NCV or PCMCIA) and its connector of antenna.

To indicate a connector there is in first it type:
type NR, type SMA, type TNC, type F, type MCX, type MMCX, type UHF, type BNC (not of threading but 2 pins), and it there in has others (photo exit of

Some are connectors (to be welded) dedicated to the test of end of chain, example, it
Hirose ms 156. It is not conceived to be handled more than 100 times.
connecteur hirose ms_156

It contains a contact which cut the bond towards the antenna, when it is connected.
It does not have machine element allowing to maintain connection, in short, it is not a true connector.

To indicate a connector there is then the sex:

If threading is interior, it is one male (M), if not it is a female (F).

Yes, I know, it is odd!

Then, it is necessary to determine if it is “normal” or “reversed” (RP, for “reverse polarity” in English)
For one male, if the center of the connector is one point, it is one normal.
For one female, if the center of connector is one hole, it is one normal.
For one male, if the center of the connector is one hole, it is one reversed.
For one female, if the center of connector is one point, it is one reversed.
To complicate, when it is normal, one does not say nothing.

Then, there are connectors which are assembled at the end of one cable (to cable), others which fix on a sheet (with to weld, frame), others which are welded onto printed circuit (to weld onto circuit printed).

For the same connector with to cable, there can be several possible diameters of cables, it then is given type of cable compatible.

Certain connectors have one not screw opposite of the time direction (Reverse Threaded)(RT).

Here, to indicate a connector, it is necessary to know all it who is in red.
Some examples follow:

Connector U.FL to weld:

male type NR (to be cabled)


type NR female (to be welded)

male type BNC (to be cabled)

female type BNC (adapter)

male type RP-TNC

female type RP-TNC

female type TNC (adapter)

male type RP-SMA (to be cabled)


female type RP-SMA

How to choose the cables?

A pigtail it is a cable with a connector of each end.

For the cables, the RG58 is recommended for 400MHz maxi (thus not good for WiFi), the RG142 (until 15GHz) has summer retained by it manufacturer pigtails.
What counts, it is the number of dB per meter of cable, with 2400MHz.
When I see all these pigtails proposed without it being indicated their loss in dB at the WiFi frequency, that irritates me….

Example the RG58= 1dB/m, LMR400=0.2dB/m.

Here some équivalences diameters of cables:
diameter 10mm: RG8/213 ‚ WBC400 ‚ LMR400‚ Altelicon CA-400, Belden 9913,7810
diameter 6mm: RG58/141/142‚ Hyperlink CA195RW ‚WBC195 ‚LMR195 ‚Belden 8240/8219
diameter?: WBC200 ‚LMR200, Altelicon CA-200, Belden 7807
diameter 2.6mm: RG316174/188 ‚ WBC100 ‚LMR100A, Belden 7805

A table of correspondence between connector and cables compatible.

End of a myth on WiFi, them walls, them doors and windows?

Then here, at home I have a AP with one WiFi antenna of origin on the roof (2 antennas in diversity mode, one is coupled with one-way). I have one house with a stage. The ceilings of the DRC and the first are composed of concrete 25cm (massive).
With the DRC, I equipped my AP (as a customer mode) with one antenna one-way (one ricoré). I sought by or passed waves.

While directing ricoré towards the AP of the roof (with through the ceilings, in straight line), I have a bad signal.
While directing ricoré towards the door of the corridor, I have a good signal, idem towards the windows.
When natural persons put themselves in encadrure of carry, the signal is less good.

Thus WiFi do not cross the walls (or ceilings). WiFi follows the doors, the corridors, them stair-wells, is reflected on the houses of with side, them trees in summer. See “Weakening (attenuation) relative of radio signal WiFi according to the nature of the obstacles (materials)".
The site chronomicro gives us values:

Material Level of attenuation
Not smoked thick pane 1-2 dB
Plating/carries out of wood 2-3 dB
Plaster partition 3-5 dB
Box of office 3-5 dB
Partition in verre6 dB 6-8 dB
Water (bottle/fountain/aquarium) 6-8 dB
Brick wall 8 dB
Ceramics tiles 6-10 dB
Stored paper 10 dB
Armor-plated glass 10-15 dB
Concrete wall 10-15 dB
Metal mirrors/preparings Total reflexion

There exist software for to simulate the radio coverage in interior.
software permettent to simulate cover of each AP:

Les anti-WiFi paintings against the hacking, are thus inefficaces, as long as one stopped not the waves which pass by windows.
This can be made thanks to the coatings thermics (transparencies!) who are used for better isolating the panes against les losses of heat.

For to generalize, it is enough to imagine house under the following conditions:
It grows dark, the AP is a source of light, there is bouillard, and the thick walls are mirroirs, the doors (not metal) are out of glass.
The problem to receive the AP is the following: “Is what you see the light? “; -)

The analogy with light is also valid to compare an antenna omni and an one-way antenna. You have a bulb of headlight of car (antenna omni), when you place it in the headlight of the car (antenna plain), the light carries much further.

Still another thing, as long as I there am, as much mode “diversity“is not used for nothing for a bond long distance (>1Km), as much it improve frankly the range in interior. Thus before to compare materials, check well that these materials allow the mode “diversity”. It is necessary in particular that these materials have 2 antennas. A fortiori, exclude the materials which have only one antenna: some AP, cases USB, charts NCV.
Apart from this criterion, I did not note difference of performances notable enters the current materials (except perhaps the threshold of reception).
If you choose AP, you can put one of them in infrastructure and it (or them) other (S) as a customer mode.
All the charts PCMCIA which I opened had 2 antennas. I do not know if all exploited the mode diversity, but that would be nevertheless astonishing that they do not exploit the 2 antennas!
The large manufacturers who propose to off-set them antennas, cases sell which contain the 2 antennas with to off-set, to use the diversity mode.

The principle diversity applies only on the receiving device at one moment T. Thus for that your bond either effective, it is necessary that each apparatus or equipped with 2 antennas. See the chapter “How to improve a WiFi connection?" if necessary.

And why not
3 antennas for each apparatus, as on WRT54GX, for to further increase the flow and/or the range?

My point of view is confirmed by this:
We recommend thesis first MIMO-based products to residential users. There is significantly better throughput and arrange performance than conventional products in every box we tested. MIMO-based has better approach than adding third-party, high-profit antennas but activates repeaters - it' S less complex and, even with the higher prices, usually less expensive. “

I translate; one gains more if there are more antennas, that if one puts antennas at strong profit, or repeaters. It is less expensive and less complicated to implement.
Thus while waiting, start with 2 antennas of each side of bond:)

To still improve the profit you can to use a “flatenna”:

to transform an omnidirectional aerial into antenna one-way (directive).

Arnaud (Alf92) nous propose to make it oneself. one speaks then “about antenna chocapic”….after the “antennas ricoré”, one can say that WiFi is morning; -)

Problem, I think that one cannot assemble 2 reflectors on the AP which have 2 (or 3) antennas, because of the obstruction mechanics reflectors…

Who will gain, the pair coppered (the wire of the telephone) or the radio?

MIMO + OFDM = 1Gb/s with 5GHz (the ADSL does not have which has well to hold!)
montée in flow of the networks radio operator Wi-Fi se continues (07/2005)
Radio techologies make it possible to increase the flow further (03/2007):
WHDI on the tape of the 5 Ghz emits, and can transmit to 3 Gbps while using the band-width of the 40 MHz "
What misses with this information, it is the number of channels (of 40MHz) used.

With my first course of electricity, one learned to me that the current always circulate in a loop (wire outward journey and wire return).
In addition, a loop constitutes a coil, and 2 wire coast at coast, constitute a capacity.
The capacity (condenser) is enemy increase of frequency and the increase in frequency allows the increase in flow.

“The problems which our customers encounter Freebox are dependent with too long or too short lines, or disturbed, even parasitized by a condenser placed at the time in telephone socket by the operator history”, indicate Michael Boukobza, managing director of Free
If you find a condenser (a capacity) in one of your catches telephones mural, remove it!
It is the black “animal” on the photograph:

Thus I am not astonished only the solutions telegraphic high banc are not proposed that has of less in less privileged (check) (the telegraphic flow “56Kb/s” is accessible to ALL la French population), whereas them solutions radio for an equivalent flow, they see range to increase.

The fiberoptic until chaque subscriber is very expensive:

se amount with several tens of billion euros"
Count yourselves, with 2 euros per meter and per annum. , and up to 12 euros with hiring…

Especially that FT would like to adapt the monopole fibre (option PON) , for ramasser under as with the good old man temps (at the time of the monopoly of telephone) (it is similar in Germany).

To buy 1Km of coppered pair, to buy posts, to rent one excavator oubien to put antennas on its roof?
(copper is expensive (07/2006)… value of overhead lines the SNCF too!).
And noise of the power picks to make the trenches in cities? And them stoppers of circulation because of work? (WiFi, it is l'ecology! ;-)

The wire resist of course short distances (less 100m), as follows:
"In matter of 10GE, wiring copper in twisted pair on a distance from 100 meters, as for the lower flows of Ethernet, goes quickly to become reality. This type of deployment will be officially validated with ratification of the standard IEEE 802.3an (10GBase-T), envisaged for the semione. It will have to exploit a wiring of category 7 (standard not ratified, but which made its appearance in the offers des cable-laying ships), which makes it possible to reach a frequency of transmission from 600 MHz and requires four pairs of copper, it who force to replace the traditional RJ-45 by a new type of connector ", source.
The 10GE corresponds to 10Gbits/s, on Ethernet.
In this case, a coppered pair is equivalent with a WiFi channel. 4 pairs with 600MHz are equivalent to 2.4GHz, but on 100m only.

One sees the curves of course below that the améliorations brought to the ADSL are done on flows and NOT on distances:

WiFi allows moreover long distances…

See: ADSL : eligibility with more than 5000 meters of a NRA

Which is the form of the radiation of an antenna?

One calls that a “owner of antenna” or diagram of radiation.
Antennas omnidirectional (tube), in driving position, send the waves with the same power all around them, on the horizontal level passing by the antenna (perpendicular (cross-country race-whodunnit) with the axis of the antenna (made up of one or more dipoles end to end).

Here what that gives if you look at the antenna over (cross-country race-whodunnit):

It is what is shown by the yellow curve. To compare them the shapes of antennas, one places the point corresponding to the maximum profit on the circle most external of the squaring, with 0 degrees.

For an antenna linkeddirectional, les other points (for each angle) are thus more close to center, and draw a geographical curve for which one receives maximum power.
The profit of the antenna is the maximum power.

Example of an Yagi aerial (like that of the TV herzienne (rake)):

If one place, not perpendicular to the dipole, but dans the same plan (Co-whodunnit, rise) (one is placed at side of the antenna), there exists another diagram of radiation (of 0° (in high) with 180° (in bottom)).
Example of a omni 10dB (even plane, Co-whodunnit):

You see that if you place the antenna in height too much, you will not collect the signal below, for this antenna. Others antennas “sprinkle” also below, but with the detriment profit.

Example of the antenna roasts SD15 (even plane that the dipole, Co-whodunnit):

Diagram of cette antenna patch (9dBi) in copolar:

and in crosspolar:

Example of the antenna roasts SD27 (even plane that the dipole, Co-whodunnit):

Although the antennas presented have different values of profit, the squaring of the diagram is the same one (only the scale changes), what makes it possible to compare only the forms (they are standardized curves).

Example of the SD27 in cross-country race-whodunnit (plan perpendiculaire au dipole):

I turned the maximum point of 90°, so that that resembles it that it would be seen whether one were beside the antenna, on the level of ground.
It is seen that the beam opposite the antenna is broader in the perpendicular plan (cross-country race-whodunnit) that in the same plan (Co-whodunnit).
One can thus conclude, that for the SD27, the beam opposite the antenna forms a rectangle. Imagine a stick of freeze (Eskimo), you almost cut it to the end, that made a section of form rectangular.
If you turn the antenna to make so that the dipole is vertical (like a omni), then the covered zone will be one rectangle flat.
In this case the azimuth (angle compared to north) will be easy to regulate, whereas the slope (high/low) of the antenna, is more delicate to regulate.
Contrary, if the dipole is horizontal, the adjustment of the azimuth will be more delicate to regulate. On the other hand you will profit more of the reflexions on the ground. Especially if this one has the shape of an half-ellipse (see zone of Fresnel) (for example enters sides of a valley).

These 2 ways of diffusing the waves (that is called “polarization”, vertical or horizontal) must be considered according to context: to pass between buildings, to use them reflexions of ground, etc.

pas is not needed to mix both types of polarisation, if not you will have very bad results.
Des appareils peuvent programmer la polarisation dynamiquement (PowerStation2 Ubiquity).

To be complete, it is necessary to speak about the antennas propellers (helicoid) whose dipole is not a right wire but a whorl.
In this case the diagrams Co-whodunnit and cross-country race-whodunnit are identical.
In addition, the direction in which one winds the whorl, makes that it an antenna wound in the same direction is necessary in face, under sorrow not to receive large thing.
Polarization hélicoïdale is incompatible with vertical or horizontal polarizations.

When one does not have the diagrams of radiation of antennas, one finds like characteristic, the opening. It is the angle around the maximum point power, for which the antenna emits half of its power (loss of 3 dB).

To show these owners from these photos, I created one program which posts the palette. One selects la color corresponding with the curve, and then the program generates the points in a fichier (*.ant) readable by “radio mobile ".

How to circumvent an obstacle?

More the obstacle is close to one of the two points of the connection, more it obstructs the connection, but nothing prevents to put it WiFi relay in the medium.

One relais credit can be fed into solar, and to undergo 15 days of misty time, like says it Dominique Guardiola Falco de Quinode (in Auvergne).
relais actif fact more than to circumvent an obstacle, it double the range.

Dominique used 1 AP, in repeater, to decrease consumption electric compared to a solution with 2 a.
Mekari proposes une solution all in one. source. It is planned for the sun of California? or it is awaited climate warming?

In addition, I would be curious to know the maximum duration (for France, for example) of the absence of sun AND of vent, for to determine the autonomy of combination solar wind, while using a battery 100Ah.

In the absence of power supply (sector, wind, solar) (a AP consumes 5Watts), there exists the principle of the antenne passivate WiFi. I validated his principle on ground. It acts of 2 antennas (of which one at least is directing) which is connected by a cable.
Here the principle which I checked (who must to also go for WiMax, or of GSM 900MHz and GSM 1800MHz):

Each antenna being directed towards each one of points to be joined:

Source des photographs.
One sees here a relay made with limp “pringles” but any other antenna is appropriate.
On the level of the assessment of the connection, nothing is changed, if not that with place to have a great zone of Fresnel, one has 2 small, end with end.

There are people who wonder how that can go, without energy source, here my answer:
I went to a living room radio operator amateurs, I have listened to the radio in a helmet, with a station with crystal, without no source of electricity.
The volume of the sound is low, certainly.
From or the energy comes which produces the sound in helmet? ..... of the emetor.
Thus the antenna converts the radio waves into electricity, this one is thus the source to re-emit on second antenna.

On the other hand thanks to this installation, you have all them chances to have 2 well cleared zones of Fresnel, instead of one only evil released. A zone of Fresnel well released (value gives notorious) 6dB of plus that a connection on the ground on a flat ground (in close-cropped-mound).

To include/understand, it is necessary to know what is a zone of Fresnel. It is the useful part of the air which transmits the radio signal.
It with the form of a “grain rice":

In this example, the useful part is decreased by the ground (with left), it is the tear line. You come from to include/understand why the antennas are often on pylons: -)

And here an good example of natural pylon:

According to Sylvain, le wind does not obstruct the connection.
Attention, the resin leaves badly to the machine to wash…

Here now another bond which corresponds to a distance shorter:

It is what I call a bond “in close-cropped-mound”. Ce bond will go hardly better that above, whereas la distance is of half smaller.

These profiles were obtained with “radio operator mobile”.

All partie of the ground which tangent the grain of riz (zone of Fresnel) améliore propagation. Thus to put the antennas highest possible the best propagation does not give. ;-) Of another side, one becomes depend on the vegetation and of constructions…

Thus our passive relay WiFi must enable us to increase surface grains of rice, therefore to increase the margin in dB, therefore to increase the flow.
It is better two goods small grains of rice, that large bad. (saying of the day: -))

But if the straight line uniting the antennas (one calls it line of sight, or English LOS) is blocked, all is not lost:

The zone of Fresnel (between the top of the obstacle and the part higher grain of rice) (here in green) makes it possible to pass the waves, but with more losses than if the line of sight were released.
To obtain best the résulat, it is necessary to aim, with each antenna, above the obstacle.
The height to be aimed depends on the distance between antennas, but it would seem that a height should be aimed at corresponding to ray of the ellipse of Fresnel. For example, for a distance enters antennas of 3Km, that made 10m with the top of the obstacle.
On the other hand, la zone of Fresnel being an ellipse, if 2 points of the connection are close (here out of chestnut) the connection will not go, since the grain of rice is completely blocked.

It is a large banana better, that a small sausage. (second saying of the day: -))

Below my page is in work…

AP (Access Not) (Access point)
OFDM (orthogonal frequency division multiplexing)
GAF (Automatic Management of the Frequencies)
PBCC (code of protection per redundancy)
FHSS (Frequency Hopping Spread Spectrum).
DSSS (Direct Sequence Spread Spectrum)
WEP (Wired Are equivalent Privacy).
MAC (Media Access Control)
DHCP (Dynamic Host Protocol Configuration)
WPA (WiFi Protected Access)
TKIP (Temporal Key Integrity Protocol)
FAI (Supplier of Access Internet)
BSS (BASIC Service Set)
ES (Extended Service Set)
CSMA (Judicious Carrier Multiple Access)
CRC32 (Cyclical Redundancy Check on 32 bits)
MIC (Integrity Message Codes)
EAP (Extensible Protocol Authentification)
RADIUS (Remote Authentification Dial-In Using Service)
TLS (Transport To bush-hammer Security)
LEAP (Extensible Light Protocol Authentification)
WECA (Wireless Ethernet Compatibility Alliance)
DOS (denial off service)
AES (Advanced Encryption Standard)
UMA (Mobile Unlicensed Access)
HSDPA (High Speed Downlink Packet Access)
WISP (Wireless Internet Service Provider)
Pulsers (Pervasive Ultra-wideband Low Spectral Energy Radio operator System)
Agora (Access Guaranteed and Optimized for Areas and Installation territories)
QOS (Quality Off Service)
TBF (Token Bucket Filter)
MPU (Minimun Packet Links)
MTU (Maximum Transfer Links)
SFQ (Stochastic Fairness Queueing)
MIB (MailInBlack) (authentification of email): -)
RAN (Rural Area Network)
WDS (Wireless System distribution) or (Wireless Domain Server) at Cisco
OSPF (Open Shortest Path First)
OAPI (Open Access Not Initiative)
ARP (Address Resolution Protocol)
WPS (Wi-Fi Positioning System)
WPS (Wi-Fi Protected Setup)
TOA (Time Off Arrival)
PLT (Precision Technology Hiring)
WCN (Windows Connect Now)
SOFDMA (Orthogonal Scalable Frequency Division Multiple Access)
WAPI (Wireless Authentication and Privacy Infrastructure)
WLSE (Wireless LAN Solution Engine) (CISCO name)
SIP (Session Protocol Initiation) replacing of H323
FTTA (Fiber Through The Air)
GIX (Total Internet eXchange) or Node of exchange Internet between operators

To increase the range of FreeBox:

I publish these photographs (the “modest one” and “bricolo”) to be used as support with improvements of these modifs.

In the round of bottom, it is necessary to cut the vertical track (or to remove the resistance which is under the heart of cable).
Thus, it would not be necessary to cut the tracks towards “ANT1” and “ANT2” (round in top).

The “large pies” of the medium, it is for to weld mass (on the left it was a test).

Good, then my note:
- there is no more that only one antenna instead of 2, therefore one loses the diversity mode,
- me, I would have rather added connectors (RP-SMA to be welded) on the level of “ANT1” and “ANT2” for reasons of mechanical robustness (and use of the mode diversity),
- then I would have tested with and without cut of the tracks on “ANT1” and “ANT2”.
Here. :)

LiveBox Sagem version:

Odd, that they cabled only one exit (AUX and not MAIN).

LiveBox Inventel version according to “Sparky”):