solar powered wi fi community node

I became interested in community based 802.11 networks after I realized I was living in a hotspot funded initially by a local apartment building owner and built by the local wireless users group.

Having geek tendencies, and my own business motives I decided to join the enthusiastic volunteers in the wireless users group and do some “legitimate” hacking (by which I mean build something cool - it seems too often in the popular mass media hacking and hacks are associated with breaking into stuff).

Business and Consumer Motives

As a capitalist consumer, the idea of getting something for nothing is simply irresistible.

In the case of a capitalist (i.e. an apartment building owner), spending a hundred dollars a month to provide a broadband pipe or putting up a thousand plus for a node on a public network is a great public relations move (in other words advertisement - it might be though of as a big corporation like BP or Ford sponsoring an alternative energy vehicle, it benefits the company in terms of great public exposure) that is a tax deductible expense since it is a service enhancement for tenants.

Because in this case volunteers donate their labor for free, it is win win situation for an apartment building owner, tenants and the neighborhood. I have calculated that by building a community based network and distributing the signal using 802.11a back hauls, and using 802.11b omni broadcast units, the cost per user in many cases is at least an order of magnitude less than commercial services charge individual consumers per month (this is because a broadband line can be shared among many users in a neighborhood).

As a consumer (in this case a resident of a neighborhood), the benefit of free wireless broadband access is not having to worry about spending money on what might be considered a luxury item. For individuals on a tight budget not having to spend thirty to forty dollars a month on broadband access means they can save that money for other items. Benefits of having access to free WiFi are vaguely defined, but one could easily envision a scenario where disadvantaged individuals could use the web to learn about new technology (which could broaden their range of possible jobs) and allow an individual to work their way up the economic ladder.

Geek Motives

As a part time geek, the benefit of helping build a free wireless broadband network is ya get to learn about and build cool shit!!!

solar power 802.11 free node hotspot

My first design contribution to the network was this solar powered AP, which was built at a cost of just over $1400 in parts (the break down is something like $500 for the [Soekris] computer which includes the 802.11a backhaul and 802.11b omni, about $500 for the solar panel, batteries and charger, and about $400 for the battery box and adjustable stand).

802.11a provides the back haul

I made a few mistakes during the building process but now that lessons have been learned it should go faster the next time I have to get all the parts.... By the way thanks to everyone who helped with setup (Michael, Matt, Bill, Lee, Jim, etc., etc., etc.,).

A solar installation was selected for this location for several reasons: first was cost effectiveness - if an electrician was used to run conduit to the roof it would have cost at least as much if not more (for parts and labor) as the solar setup, second the stand alone solar setup was a trial design that could be used in remote regions, underdeveloped countries, or heaven forbid in urban areas after a blackout cause by terrorists attacking transmission lines or power plants, third “Geek Chic” need I say more?

FYI power comes from a 65 watt solar panel, two six volt 225 AmpHour batteries wired in series which should power the setup for several weeks if the solar panel dies (batteries should last 10 years and the solar panel 20).

Beta Solar Node

To extend the coverage of the network I built and setup (with lots of help from sdwug members) a second (hence the name beta) node. Solar power was (again) used for a custom metrix rig because all I had to do was set it up on the roof and not worry about having an electrician run a power line from the basement.

the beta solar node, another free hotspot for the network

Because of lessons learned from the first solar rig, I scaled back the size of the panel (a unisolar 32 watt), and tried two 12 volt 18 AmpHour gel cell batteries wired in parallel.

Note some long term solar users may have concerns about the least little shadow on the solar panel will defeat the output almost completely (see photo above) to which I'd like to say, Unisolar panels are somewhat more shade tolerant than crystalline panels (such as Siemens, Kyocera, Solarex) because it is fitted with individual Bypass Diodes between cells.

For this node, I ordered parts from metrix, which included the weather tight box and a Soekris 4526 (which is much smaller than the 4511 boards we had been using). I have not totaled up all the costs yet for this set up, but I'd guess it again came in around $1400 for parts (a bit more for the metrix kit, a bit lower for the power system because a smaller panel and battery were used).

FYI in the process of getting parts for this particular setup, I had a chance of contacting Matt Westervelt (founder of SeattleWireless.NET, and Metrix.NET) and asked him if he had any experience with others trying solar rigs. Basically he said nope because up north there is not a constant source of sunshine (unlike San Diego). So my advise is if you are going to try solar power for a WiFi set up, do some calculations first to see if solar power is viable at a reasonable cost.

Murphy's Law
(no sun, not enuf power)

When I built the second node I tried using a smaller battery and solar panel. This combination worked a few months, that is until I was thrown a curve by mother nature. This past Oct 2004, it actually rained for more than a week here in San Diego and the node died because it did not have enuf reserve to keep on running.

When the week long series of storms ended, I expected the node to pop back up, but after a few days of sunshine I realized things were not working out as planned. I climbed on the roof and discovered that the battery was charging slowly but there was not enuf power to run the AP during the day and at night the AP was drawing down the battery reserve, opps!!!!!

My first idea was to try and build a "low voltage cutoff circuit" so that when the battery runs below a preset voltage the battery stops sending power to the AP (thus preventing a night time drain). However, I figured this is one more possible thing to go wrong so in keeping with the KISS design philosophy I have decided to try switching solar panels. In other words I think small panel, large battey combo or big panel, small battery combo will work this winter... STAY TUNED!!!!

Murphy's Law II

Switching the solar panels did keep the AP powered during the winter months. But as is always the case other issues came up - like the AP crashing. The metrix boxes in the network seemed to have issues (one on pink palace and one on my second solar node) so it was replaced by a WRAP box which I bought for the bankers hill project that I've been meaning to get up and running. Anyway a few days after the Metrix was switched out for a Netgate box, Michael noticed the node crashed. Seems its possible to login remotely (i.e. the 'a' radio back haul was working), but in the logs there was a "wi0 seek error" which first hits the 802.11b side of the prism card, and eventually the whole box crashes and needs to be power cycled. The fix seems to be to replace all the 802.11b radios with a/b/g radios - at least until we can get a distro with a fixed prism driver (FYI in later versions of m0n0walls are stable, but they'd don't support the atheros cards we need for the 802.11a backhauls).

One other item going back to the command line interface of Pebble isn't an option cause after playing around with both its like night and day (the graphical interface is so so so much simpler to use).

relay switch used as an AP reset

Climbing on a roof takes time and effort putting up ladders and all, so this is the simplest way I could think of to reset the AP from the ground.

I had been thinking of building a remote switch (based on a 56 bit encrypted garage door opener) to reset the system, but ruled that out because it that idea was complex (there was also the issue of misplacing the remote), and opted to use a simple wired relay to reset the solar node.

Going Online
(free hotspot vs. commercial services)

The major problem I discovered with developing public networks is trying to find an ISP that is willing to let a signal be shared. When I talked to a Cox cable representative to ask if I could use a cable connection for another pipe for the network I was asked the question “Why should we let you re-broadcast our signal?“ I tried showing the demographic difference and need between Golden Hill/Sherman Heights (92102 - lots of poor less-educated mexicans without web access) and La Jolla (92037 - basically rich white folks with web access). But Cox being what it is (a typical corporate business without a soul and a TOS agreement with lots and lots of fine print) only seems to care about their short term bottom line (and not helping out a geek run community service network that has great long term potential to help an under served demographic and develop technology). Bottom line from what I was able to gather with informal discussions is, Cox has bucks to advertise HBO, Showtime and other crap on TV to poor Mexicans, but does not have plans to help out under served folks get net access (the one exception being at boys/girls clubs and various community centers where Cox can control giving away bandwidth and try to buy some community good will).

Unlike many ISP's (i.e. COX and SBC), Speakeasy.NET and DSLextreme.COM allow customers to run servers (web, mail, etc.) and SHARE Internet connections in multiple locations (i.e. a community network). The reason I tried called COX is because DSL is not an option for the current geographic end of the network (because at over 15,000 feet it is too far from the CO), ideally there should be connections at both ends so that if one goes down users can still connect.

Upon doing some checking I learned from a friend (in order to protect his identity just call him Mr. X) that cable companies and phone companies over the past few years spent something like 70 billion plus on infrastructure during the dot com bubble era, and need to keep prices high to pay off the high costs of over building bandwidth infrastructure (some of which is not even being used). I also was told that each time a telecom company rolls a truck to a customers site the cost is something like $300 (which includes stuff like labor costs, insurance, marketing, cost of the truck, etc.) and that is why for example it seems companies give generous incentives for users to self connect DSL and cable modems (in other words a telecom company makes lots more money when users don't use on site tech support).

Although a community WiFi based network can be build at a very low cost, the problem is a public network is not manned 24/7 and is subject to occasional outages - but hey ya can't beat the FREE price for end users. IMHO free hot spots complement commercial services in that community networks provide a valuable service to everyone especially individuals without economic means, but if an individual is dependent upon the web for business or personal communication they would be very likely to subscribe to a for profit commercial operator that can provide guaranteed higher-speed service with greater security and privacy.

FYI it is a well known open secret among geeks that WiFi communications is unsecure. Although people may wish to think that surfing the net web is a private affair, all users should be made aware that unencrypted data passing over public access WiFi hot spots can be intercepted by any unscrupulous individual or organization.

Legal Questions about HOTSPOTS
and public networks

As with everything else, there is always going to be some kind of legal bureaucracy (i.e. bullshit), that will try and dictate how something like a public network operates. So far there has not been specific case law about hotspots (as far as I know), but it might be argued that a public network could be considered just like an ordinary tool, or automobile. In other words it is the responsibility of the operator of a product for their own actions.

The legal precedent I point to is with Napster's grave still fresh, the Recording Industry Association Of America (RIAA) have scooped up their pitchforks and torches and are marching on their next target: Grokster, Kazaa and MusicCity (Morpheus). The courts ruled on these services that the software only allows users to connect with each other for open exchange - like a file transfer. One EFF lawyer argued, "You don't hold a crowbar manufacturer responsible if its product is used to break into a house."

IMHO operators and owners of public networks should be considered just like crowbar producers or auto manufacturers (in that we just provide a good) that consumers use. Can you imagine a how long a public DA would have a job if they argued in court that since a Ford POS (piece of shit) SUV which was used in a bank robbery, that Ford be held responsibility for the crime?

I have read it is not illegal to take any signal out of the air, it is however illegal to decrypt a signal. That is why HBO ended up having to scramble their signals. They were sueing private satellite dish owners and manufacturers for copyright infringement. The US supreme court held that if it was not encrypted, it was indeed public domain. Secondly, the FCC has determined certain channels to be public use the 2.4 gig range used by WiFi is among those.

WiFi Security?

It is a well known open secret among geeks that WiFi communications is unsecure. When web surfing from any public hot spot, user URLs and other information are brodcast in PLAIN TEXT!

For the next expansion of the network which I will put in the Bankers Hill area, I am planning to emphasise education content so people can not only use the service, but that they can learn about the social implications of the technology. What I want to do is drill in the point to typical users (who usually don't know anything about technology), that there are tradeoffs in using an open network and one of them is possible privacy loss by blackhats who may want to steal an ID (by sniffing traffic packets), etc... so I plan on using a somewhat graphic login page.

Environmental Motives, and
Bridging the Digital Divide

As an individual concerned that humans are destroying the environment for short term gains, I see solar technology as one possible way to achieve economic and environmental sustainability while enhancing quality of life. With these nodes I have demonstrated solar power is a superior alternative in some applications to traditional AC power (even in a major US city), that not only is cost effective but extremely reliable. The other serendipitous application of solar powered nodes is potentially addressing the problem of the digital divide.

In todays world, society is split between the digital haves and the digital have nots. Basically I imagine the digital divide as a gap people need to cross in order to use online information and communication tools. Basically rich educated people face little or no gap because they can buy access, poor un-educated people face a big gap because they do not have the resources needed to build digital infrastructure which is required to access the web.

The web as we know it is transforming how people communicate, gather information and purchase goods and services. If nothing is actively done help people get online then the information-poor will become more and more impoverished (relatively speaking) because government bodies, community organizations, corporations and other entities are increasing turning to the web as the venue for communications and commerce.

In order to bridge the digital divide the digital have nots must over come two major digital infrastructure barriers. The first barrier is access or ownership of a computer and the second barrier is paying for internet access. Both these barriers are surmountable and can be done so with community based programs that are environmentally sustainable as well as being cost effective!!!!

Number of Internet users worldwide

in millions
Users as percent
 of total world users 
Users as percent
of population

World total407100.06.8
Middle East20.51.1
Latin America174.23.2
United States and Canada16741.053.9

Sources: User estimates are for November 2000 from Population estimates are for 2001 and were drawn from

The traditional method to connect to the internet has been dial up with some monthly plans starting under ten bucks per month. For those individuals who desire broadband service to download movie trailers or just surf the web faster they can connect to the web using ISDN, DSL, CABLE or Satellite for a bit more money. However inexpensive monthly dial up service is, there are still a great many individuals and social groups who are not yet connected.

people don't want to pay for poor people

The nationwide poll, which surveyed more than 1,000 people with Internet access, found that 72 percent of respondents support government efforts to make high-speed Internet access universally available. But backing for the activist policy has stalled on the question of who pays--and how much.

By providing free broadband web access in an urban area where many tech savvy volunteers can develope and document how to setup an inexpensive 802.11 mesh network (like the GoldenHillFreeWeb project) it is possible for other groups (willing to invest sweat equity) to duplicate a similar network anywhere in the world. Community wireless networks (we have shown) are a viable alternative to commercial services and would be very cost effective in low income public housing complexes, in rural undeserved areas such as Indian reservations where poverty is pandemic, as well as less developed countries where digital infrastructure is lacking.

The second major barrier to web access is ownership of computers and my solution to this problem is to recycle unwanted computers for disadvantaged groups. Because large urban areas generate lots of high tech trash such as PCs (with very high environmental costs), my solution would be some kind of educational program that would teach individuals without financial means and technical skills how to recycle old computers. By reusing old computers for use by the digital have nots, two major benefits are less impact on the environment and it could teach undeserved groups a valuable technical skill set, specifically how use a computer.

Since the consumer life span for computers is getting much shorter, it makes sense to recycle unwanted computers using open source bloat free operating systems (which work just as well as commercial products) as web terminals. I realize this mindset goes against the computer industry tide which has a vested interest in changing its designs so that people buy new stuff regardless of need. The fact is for many home users who just write letters and surf the web it is a waste of money to buy a faster more powerful computer every few years just to get a daily web fix. In many instances commercials use brainwashed to make people think they need the latest products, weather it be computers, cars, clothing, etc.

Technology (like solar powered WiFi) itself is not a silver bullet that will solve all social and economic problems, and there will always be some people who are not interested in connecting or see a need for getting on line. But for users living on limited incomes and willing to invest some sweat equity in the form of learning how to effectively use technology, recycled computers combined with free solar powered hotspots (like those found in a cooperative community built public network) have the potential to bridge the digital divide in a very cost effective way.

September 16, 2005
Afghan Solar Village

On their own dime, part of SoCalFreeNet (Michael and Drew) are off to Washington DC to meet with other NGO's and government officials, to kick around ideas and demo a home brew solar powered WiFi VOIP setup that might be used to teach Afghan students about one potential, low cost, last kilometer solution. After that its off to the gulf coast (aka the red neck rivera) to meet up with Seth and Matt to help set up WiFi for those unfortunate souls devastated by Hurricane Katrina with Part-15.ORG. I would have liked to tag along on that trip, but since I just got back from Burning Man and will be taking off for Asia next month, figure I better take care of business here in San Diego first.

I really like the concept of an Afghan solar village with WiFi and VOIP, but looking at the big picture I think an educational program based solely on solar WiFi is kind of like putting the cart before the horse. The problems being the low literacy rate of the general population and environmental issues of "limited natural fresh water resources; inadequate supplies of potable water; soil degradation; overgrazing; deforestation (much of the remaining forests are being cut down for fuel and building materials); desertification; air and water pollution" according to the CIA World Factbook on Afghanistan.

Given these facts, and a sushi dinner conversation with an anthropology grad student last night who spent some time in neighboring Pakistan studying the interaction of culture and technology, I have come to the conclusion that it would be difficult for people in country to see or reap the immediate benefit of a solar powered digital infrastructure. Instead I'd position solar powered digital infrastructure as the end goal of an educational series about useful solar power applications.

Since the first concern of most people over in Afghanistan IMHO would be looking for ways to improve their day to day lives I'd start off with an introduction to solar powered stoves, a solar water purification project and the Zeer Pot which is a Nigerian invention that keeps food fresh without electricity. This would get people in country accustomed to the idea that solar powered devices could be helpful in everyday life! Direct benefits of low tech solar technology would be to lessen the impact on the environment and improve public health.

The next step I'd suggest is introduce people to a basic solar panel then show it could be used to provide DC power for devices like LED lights and portable radios. FYI two billion plus people in the world are not connected to the grid and individual solar panels are a great way to provide basic power for basic luxuries in the third world. Down in Baja which is the nearest analog of a society without wide spread modern digital infrastructure, they use solar panels for pumping water and powering short wave radios for basic communications between distant villages and farming communities. I'd assume people in Afghanistan would use solar panels for similar applications.

A few personal observations, down in baja I've asked people over the course of several different road trips in different locations if they wanted to use the web, and the answer more often than not is we don't see the need for that kind of technology and these people have contacts with Americans and are less than 500 miles from the San Diego. My gut feeling is solar powered digital communications is not the first want or need for many people who live in the third world.

For the biggest bang for the buck I think if the US government backed a global program to export low tech solar solutions and produced cheap solar cells here in the USA for domestic and export use to help poor bastards in third world countries, it would be a positive thing for the US labor market and US economy. Besides that it would be a lasting way to buy the good will of 2 billion people who don't even have basic power. In addition if combined with a global recycling program for the batteries used in distributive solar systems, it would be a major plus for the global environment.

Rants now out of the way, if I was going to a demo solar powered communications device, I'd build something that could power the singe board computers we have be using 99% of the time. This is what I consider the sweet spot in terms of cost and benefits. It is possible to build some thing that would work 99.9% of the time or better but the cost would be much much higher. I'd also avoid any WiFi toy ideas out right cause in the real world these things wouldn't have a duty rating anywhere near 10%, but costs almost as much as the setups I've put together. In addition if ya demo anything that only works under idea conditions, people in country might not be able to see the potential usefulness of the device and dismiss the idea out right!

First off I'd use a Unisolar 64 Watt Solar Panel, because it has more than enough power for 802.11 devices and can quickly charge batteries. These cells listed on the web range in price from $325 to $500. Smaller solar panels are less expensive, but they do not meet the cost vs performance criteria I'm seeking (i.e. a system capable of powering a hot spot 99% of the time). One thing to think about when looking at solar panels, even though single crystal silicon cells are more efficient and thus have smaller form factors for equivalent power when compared to amorphous cells (i.e. the Unisolar line of products which are encapsulated in UV stabilized polymers), IMHO amorphous cells are more appropriate for WiFi applications because they cost less and are more robust (i.e. no glass can be broken).

Even though gel cells cost more and hold less charge than a flooded cell, I'd select gel cell batteries, because they can be transported on airline flights, something to think about if ya have to take a flight into Kabul and gel cells do not need annual maintenance. Its been over a year since I've installed two 18 Ah gel cells wired up in parallel to power one of the solar nodes I've built, and so far no problems (unfortunately the same can't be said for the WiFi hardware/software side).

Since helpful geeks would be in short supply in Afghanistan, or for that matter in any third world country, I would select commercial WiFi solutions that have been proven to work and have simple user interfaces. I am biased toward commercial WiFi equipment because hopefully most of the kinks would had been worked out and the higher initial cost would be more than off set by less headaches and long term costs trying to keep the network going. Basically I am considering the fact that if the equipment used by Afghan villagers or for that matter anyone else is too hard to use or breaks down, then the equipment becomes nothing more than high tech trash!

In place of the water meter box (which I've been using for the battery box), I'd try using some kind of surplus ammo box to mount the solar controller and batteries. In Afghanistan I'm sure it would be possible to get surplus ammo boxes in which to place the gel cells, FYI normal retail cost for an 18 Ah cell is about $50 bucks each at an electronics store, but matt someone in the SD wireless group might be able to get them at half that price.

An off the shelf solar controller is necessary to prevent overcharging the battery (note to maximize equipment life span excessive ware should be avoided). To further protect the equipment I'd also include a low voltage cut off circuit to prevent the total discharge of a battery (like what happened here in s.d. last winter, before I switched the solar panels) and I'd include an led volt meter circuit to give an operator a rough idea of the battery voltage. All these simple circuits could be had for less than $50 bucks (FYI an off the shelf solar controller - Morningstar's SunGuardController cost about $30 bucks).

A stand made out of 1 5/8" fence railings and industrial connectors, IMHO is well worth the higher overall costs because its possible to quickly optimize the angle of the solar panels in any remote location (the whole solar power stand can be up and going in about 30 minutes). Down in baja I've seen solar panels just laid against a building, which works but that also increases the chances of breaking equipment. Since the solar panel efficiency is dependent upon the incident angle of the sun striking the panel, the stand makes it possible to fix the panel in an optimum position while serving addition duties like being a stable mast on which to mount comm gear, antennas and a battery box.

802.11a provides the back haul

This solar powered 802.11 access point was built out of about $1500 worth of material that can stand up to the elements and can be quickly setup (on any uneven surface).

Bottom line given over a year of trials with WiFi solar rigs here in San Diego, and info about Afghanistan (see solar, Lat and Long info below), for a thousand bucks in material give or take a few hundred I'd say it would be possible to build a damn rugged solar power station that could run basic 802.11 equipment (or basic LED lights and a radio) year round in Afghanistan.

Ali Khel33.57N69.43E
Baba, Koh-i34.30N67.00E
Chagai Hills29.30N64.00E
Chah Gay Hills29.30N64.00E
Chahar Burjak30.15N62.00E
Diwal Kol34.23N67.52E
Dughi Kala32.20N62.50E
Haidar Khel33.58N68.38E
Khojak Pass30.51N66.34E
Khyber Pass34.10N71.08E
Lash-e Joveyn31.45N61.30E
Lora31.35N66.32 E
Maidan Khula33.36N36.50E
Margow Dasht-e30.40N62.30E
MAfghanistan Qal'eh32.20N64.50E
Nau Qala34.05N68.05E
Nawar, Dasht-I33.52N68.0E
Qala yangi34.20N66.30E
Safid Kuh34.45N63.0E
Sangar Sarai34.27N70.35E
Sar-e Pol36.20N65.50E
Spin Buldak31.1N66.25E
Spin Buldak31.1N66.25E

Note that the latitude of San Diego at 32 degrees north is similar to some places in Afghanistan. FYI according to the ADB, the solar radiation in Afghanistan averages about 6.5 kilowatt-hours per square meter per day, and the sky is sunny for about 300 days a year.

Last thoughts.... inspired by a third world WiFi system I thought A mil spec hand crank DC generator would be a nice solution to provide a backup means to recharge batteries for the solar AP, but at $4000 bucks (Linda Sassar phone (585) 244-5830, sales at I realize this is beyond the financial means of the Afghan project. So a quick and dirty solution might be to adapt an old 14 volt battery powered drill with a hand crank on the chuck or some how connect the drill motor to a bike to generate power (I need to work on this idea further to make it user friendly). FYI a diode placed in series with the jumper wires from the battery terminals on the drill is needed to convert a drill into a generator (the function of the diode is to prevent the current flowing to the drill motor from the battery, in other words the diode allows power to flow only one way).

Although highly unconventional, it is also possible to generate electrical power using another item found in great quantity in Afghanistan and that is 7.62x39 ammunition used in the AK47. What I'd do is set up a firing range and have the targets be small steel target paddles connected to a generator (it would be akin to using the wind or water power to turn a turbine). This method of generating power although very very very inefficient could be an exciting way to quickly dispose of lots and lots of ammo, and if done right could be rather artistic.

California Wireless Summit
October 28 - 30, 2005

Just spent some time at the California Wireless Summit...

The Summit will bring together representatives from community based organizations, activists, community technology and wireless experts, community Internet policy advocates, and local policymakers. Our objective is to build a statewide network of advocates working to bring Internet access and resources to all Californians.

On the first day "Gunner" the MC of sorts, placed tape down on the floor of the room (which was suppose to represent a spectrum) in order to gauge various participants point of view on two questions. The first question posed was something to the effect, do people strongly agree or disagree that WiFi is important. The second question asked people to place themselves on a spectrum and decide if broadband is a luxury or a fundamental right.

The first question I thought was pretty open ended, so on a scale of 1 to 5, where 1 was "WiFi was no factor" and 5 was "WiFi was the most greatest invention ever" I only thought this question deserved a 2. FYI no one in the group thought WiFi was no factor, several people thought like me WiFi deserved a 2 (in other words not vary important), some people thought WiFi deserved a 3 (in other words some what important), many people thought WiFi deserved a 4 (in other words vary important) and no one thought that WiFi was the greatest thing ever invented.

WiFi is neat because it allows people to access the net many places like in a yard, in a cafe, etc. BUT IHMO WiFi is just a tool that makes life easier and is not as important as say a power grid, running water, a stable government, a free press, etc.

On the other hand, many people in the group considered WiFi just as important as a power grid, running water, a stable government, a free press, etc. because it was a way to communicate ideas and gather information. The one question that came to my mind is if people did not know about WiFi technology could people still live a health life. Since WiFi is something that is not shelter which is necessary to survive, or a consumable like water or food, I don't consider WiFi worthy of being elevated to a status of being necessary! But with people's growing expectations, it seems many people think of yesterdays luxury as todays necessity.

As to the question if broadband is a luxury or a fundamental right, most of the people at the wireless summit thought it was fundamental right, no one thought it was a just a luxury and only one other person other than my self were in the middle of the spectrum. Since many of the people there were employed or made their living working for a non profit group advocating that minority groups among others have broadband access, it was not too surprizing a result. Since the subject of the digital divide is only one of my many interests and I do not benefit financially one way or another from the WiFi industry or work for an NGO that advocates for universal broadband, I was the odd man out so to speak. Broadband is nice, but in my mind it is a separate issue from basic information retrieval (i.e. I think it a fundamental right that people have access to information, but I don't think its necessary that all pages load in under a second). From what I've seen text content can be down loaded pretty quickly, the problem is all the graphical crap that takes up bandwidth, hence web pages take longer to down load.

After that little exercise, people had the opportunity to discuss various topics. The basic format of the summit was broken down into three tracks: deployment (which was kind of chaired by Michael), education and policy. Since I kind of know the BS involved in deployment of a WiFi network, I opted to sit in on the educational and policy discussions.

For the most part, the rest of the summit gave people an opportunity to social network and hash out ideas on how to get people interested in the topic of the digital divide. I'm more of a geek than a policy wonk, but since the opportunity provided itself I thought I'd see what other people were up to. There was nothing really new that I discovered, during the two days of listening to the educational or policy groups. Basically they pointed out it sucks to be poor and under-educated, duh! But none of them had magic solutions on how to solve the little problem of how to cure being poor and under-educated.

Some of the people at the summit I did find interesting, like Sr. Petra Chavez who runs CAMINOS, a program where immigrant women learn office skills, including classes in computer repair, to help them move out of their low-wage housecleaning jobs into more skilled employment. Then there were the techies like Ralf Muehlen from SFLAN one of the geeks who builds WiFi nodes up in the bay area and brings WiFi to Burning Man and Matt Rantanen who runs Tribal Digital Village a pretty big WiFi network here in San Diego.

Before the summit ended there was an opportunity for people to break up into little groups to discuss topics not covered in the various tracks. Since some people mentioned sustainability, I opted into that brief discussion.

IMHO I think the triple bottom line which takes into account financial, social and environmental considerations is the best way to look at a problem. In this discussion I decided to play devils advocate and pose the question why should minority and economically disadvantaged groups have broadband.

It is possible to get information off the web using POTS aka dial up service using a 56k modem. As a matter of fact I use dial up my self as a back up method to access information when the networks goes down which happens every so often. Personally I don't surf just for fun much when the WiFi network is down and have to resort to dialup. The reason being like most people I want instant gratification and hate waiting (but IMHO a users expectation is a separate issue).

There were several attempts by other people to illustrate to me why broadband was needed: like distance learning, the example that phone and cable companies do not wire up poor black and mexican neighborhoods, etc. Since this discussion was about sustainability, I took the hard ass approach and ignored the bleeding heart liberal arguments which too often ignore financial considerations (i.e people must eventually live within their means) when the topic of broadband was mentioned. Long story short, the arguments presented I thought were not able to over come the hard reality of economics, and people confused the ideas of wants vs needs.

To show the difference of want vs need, as I looked around the room at the summit I noticed that my old clam shell iBook G3 running OS 9, which I bought used off of CraigsList.ORG a few months ago, was the biggest POS there (OK I'll admit its not as bad as Michael's old PC laptop that requires a separate key board). In both cases these old laptops can access the web, and configure wireless nodes when working on the roofs of buildings. In general there is always going to be a new computer I would want because it has the latest bells and whistles. For example I "want" to trade in my 15" powerbook for a 17" powerbook, but when all is said and done, my old iBook takes care of 99% of my daily computing "needs" of checking the network and costs 1/10 the price of my powerbook.

Another personal example I can cite of want vs need is my daily driver. I have an diesel Benz which I bought when I was at university, it has 400,000 miles one it and it burns biodiesel so it is pretty environmentally friendly. In a country where the maximum speed limit is 70 MPH, there is no "need" for me to buy another car because my old Benz is pretty reliable and economical. Having said that I'd "want" a Ferrari to drive a every day, but I realize like my Land Cruiser it is not a practical vehicle on a daily basis.

Sustainability is a difficult subject to pin down, because in the end looking at the technical merits and doing a quick mental cost benefit analysis, I find it hard justify spending tens of millions for a muni broadband network or billions upon billions of dollars to build a statewide broadband network with speeds that are on par with say South Korea (which has one of the fastest on average in the world). Using just a financial model it is easy to understand why corporations ignore some groups, the bottom line being poor people suck! On the other hand using triple bottom line analysis (which is much harder to quantify because it is difficult to place an exact value on people's expectations or potential) one might see justification for internet access at higher speeds with a muni network because it allows a wide range of people traditionally disadvantaged (blacks and mexicans for example) greater economic and social mobility. In the end sustainability requires a balance must be struck between the wants of a society (which has unintended consequences) and the basic needs of a society. is a site that has lots of stats that show being poor and under-educated really sucks!

Basically from my point of view if the objective of community based organizations and activists is to do something to lessen the digital divide in poor neighborhoods by building a network, then solid answers to the following three basic questions must be found...

1) Why should a muni wireless system be built when there does not seem to be enough funds for other public services like police, fire, public heath, schools, upkeep of basic infrastructure like the power grid, sewer systems, etc. And in order to make the network sustainable what kind of funding mechanisms are there that will garner wide spread public support.

2) What speed is acceptable for a wireless network given the nurds knowledge of what is technically possible (i.e. understanding that WiFi is unlicensed spectrum and was never designed for large scale deployments) and the desires of social justice activists (i.e. people who just want to help out poor folks). In other words is there a simple to understand costs vs benefits design document that all interested parties can agree on.

3) Will WiFi actually help an undeserved community. In other words just because a person has a computer and access to WiFi, will that person have the wisdom and training to be more productive and allow them improve their lot in life, or will that person be a mass consumer of the many vices to be found on the net.

Now if I had all the answers that would be something, but alas I don't!

Solar Powered Meraki Mini
September/October 2006

The biggest problem with solar powered WiFi nodes has been WiFi itself, specifically getting M0n0wall or pfSense to play nice with the various Metrix kits, Soekris 45xx boards, etc, so we are trying out new software/hardware from the Meraki Summer Beta program. Basically Meraki was founded to help disseminate work from the MIT Roofnet project, with the hopes of bringing free or low-cost Internet access to people around the world.

Meraki Solar

The plan is to use the existing mountings and 802.11b omni antenna, place the Meraki in the existing case and change out the pigtail. Power will be supplied via a POE in place of direct 12 v. FYI the radio in the Maraki is a b/g 60mW which is a lot less powerful than the 200mW Senao cards used before.

The pro's of using Meraki are: less time, money, and technical knowledge are needed for a true mesh network. Because the Meraki is an all in one small form factor piece of hardware, power requirements should be much lower hence using the present solar setups might be over kill in terms of reserve power needed. Admin for the Meraki is web based with lots of pretty pictures courtesy of google maps.

Building an outdoor case
for the Meraki Mini

The Mini-Meraki is designed for indoor use only, so I have come up with a very cheap out door solution, I'll mount everything thing inside 3" diameter Sch 40 water pipe and use a bigger antenna. I have opted to use cheaper indoor only components totally enclosed in tough water pipe, because this setup eliminates potential points of failure and signal losses connecting a radio with a pig tail, an outdoor cable then antenna (which was a problem I encountered with previous setups).

Meraki Solar

1st iteration walkabout

Build a home brew outdoor case out of inexpensive and strong water pipe, which will house the Meraki and a much more powerful indoor only 9 dBi omni antenna. For site surveys, this unit can "WalkAbout" or be undocked from a solar charger, for more info check out my basic design layout for my 3rd generation solar wireless network node.

$180 in parts = WalkAbout

About a year ago after looking at the Afghan Solar Village concept I came to the conclusion that WiFi networks with interconnected nodes for bandwidth sharing and failover for the third world was not going to happen until the WiFi hardware/software side was easy to setup and use, and had a lower price point. For the Golden Hill, Mercado low income housing project, etc. the network(s) were getting pretty darn complex, with various feeds, many back-hauls to various nodes, etc., but we had the luxury of a few geeks willing to keep the network going and had a few extra bucks to throw at the problem.

mesh illustration

A mesh network has interconnected nodes for bandwidth sharing and failover. Wireless network products like the Apple Airport are NOT designed to scale up a wireless network to cover a neighborhood, or something like a metro area.

After installing the initial batch of 20 Meraki Mini's over a wide geographic area, I have a feeling that many more community networks will now be built both here and in the 3rd world, because the Meraki is so darn easy to use (it's plug and play compared to our old network hardware/software setup), and is less expensive than older gear. Since I'm not Code Monkey that has the ability to write meshing code, I figure my little contribution to break down the digital divide will be homebrew packaging this neat little device, so meshed WiFi can be used at weddings, bar-mitzvahs, burning man, natural disasters, etc.

walkabout assembly

Now you too can look like a geek and be part of a mesh network if ya dare carry this thing around the neighborhood. If stopped by someone for looking suspicious tell them its just a $180 worth of misc parts (without tax or S/H).

- Meraki Mini beta $50
- indoor 9 dBi omni $17
- misc sch 40 parts $32
- waterproof switch $9
- 12v 2.9 Ah battery $24
- solar panel charge controller $34
- 18 volt transformer $8
- pass thru $0.80
- misc wires w/ fuse holder, misc connectors, etc... $4
- "SoCalFreeNet" bumper sticker $1

geek break - pizza and beer

First day out with the WalkAbout I encountered lots of interesting things like homebrew electric cars, a few roofs (a new node was installed on B St.), and even managed to do some testing while enjoying a geeks favorite meal - pizza and beer!!!!!

The WalkAbout is cool because it is a rugged stand alone node which you can use to do a site survey. The WalkAbout has many different power options, you can plug it into a wall like an ordinary Meraki Mini but it has a small built in battery that can power everything for 12 hours plus, or it can be docked and charged on a solar stand.

solar ap with docked WiFi about

The WalkAbout concept works!!!! With the WalkAbout docked in my first generation solar stand, this thing will run for a month plus without a solar panel. I've made a homebrew quick disconnect base, so that the WalkAbout can be used as an ultra portable mesh node. FYI I made a screen capture shot showing the indoor 9 dBi omni antenna, seems better suited that an outdoor 8 dBi downtilt omni antenna and is cheaper to boot...

Its a long standing American tradition to make things more powerful, so I'm adding a 500 mW amplifier to boost the signal in a pair of WalkAbouts by almost an order of magnitude over the 60 mW in Meraki Mini. Thought I'd try this because we have an issue where a youth center has a Meraki Mini on the roof but cannot connect with a nearby building because there is a freeway over pass between the roof tops. So I'll test a pair of WalkAbouts, each with a 500 mW amp and 9 dBi omni antenna, this hardware set up should improve the odds making a useable mesh connection, and will be just under the Effective Isotropic Radiated Power (EIRP) legal limit of 4 watts.

Fresnel zone diagram

I looked thru my old physics text books to read up on specifics of the Fresnel zone, basically this stuff is important in WiFi because ya can check for obstructions that might penetrate the zone. While line of sight is a good primary site survey check, it may not always be adequate. This is because if obstructions (such as terrain, vegetation, buildings, etc.) penetrate the Fresnel zone, there will be signal attenuation, doh!

When dealing with an amp in a meshed network, there are two key concepts to be aware of: 1) an amp boosts a signal, it also boosts the background noise, what ya want to find is a sweet spot where a receiver can still work but the signal travels farther, 2) you need to amp the signal on both ends because of the alligator (big mouth, little ears) analogy. Long story short, what good is a boosted signal if the receiver is crap. If ya boost the signal only on one end, the signal of one WalkAbout AP will travel farther, and it might be possible that a POS laptop would be able to pickup a strong signal from far away but a laptop would not be powerful enuf to send a detectable signal back to the WalkAbout AP. If ya only amplified one side that would be akin to pissing into the wind, in other words not thinking the problem thru, this is why I'll test two WalkAbouts with the amp option to see if I can make a point to point connection under harsh conditions.

I made basic power calculations of various configurations:

   power 19.78 dB (basic Meraki Mini - 60 mW radio, 2 dBi antenna), approx cost $50
   power 26.78 dB (WalkAbout config - 60 mW radio, 9 dBi antenna), approx cost $180
   power 35.99 dB (WalkAbout w/ amp - 500 mW amp, 9 dBi antenna), approx cost $320

The decibel (dB) is a logarithmic unit used to describe a ratio, the ratio may be power, sound pressure, voltage or in this case WiFi wave intensity. The difference in decibels between the two is defined to be

   10 log (P2/P1) dB        where the log is to base 10.

cost to pimp the meraki

An Increase of 3 dB double transmit power, so by adding a 9 dBi antenna to the basic Meraki I am effectively making the signal over 4 times more powerful. An Increase of 10 dB produces 10 times the transmit power, so by adding a 500 mW amp to the 9 dBi antenna I have a pimped out WiFi setup that is over 40 times as powerful as the basic Meraki. Keep in mind that a setup 40 times as powerful will not send a signal 40 times as far, since the signal propagates spherically, the WiFi energy is distributed over the ever-increasing surface diameter of the irregular torus wave front surface (BTW this is AKA the Inverse Square Law).

So if wireless networks are going to be used in the real world (in places like Afghanistan for example) then amping both ends might make financial and technical sense in some cases, and gosh darn I'm going to see how far I can push the WalkAbout concept using off the shelf parts. I'll also paint the whole setup in some cool scheme to protect it from UV, which not only is a practical feature but also will give the thing a bit more attitude (another long standing American tradition).

Murphy's Law
strikes a 3rd time

After initial testing I discovered that the WalkAbout when left on the roof will collect water in the bottom of tube, which I don't think is a good thing for electronics. Basically the water was an accumulation of condensation which comes from moisture in the air, and is caused by too much darn moisture in the air for a certain temperature. Condensation forms when warm, moist air touches a surface (and that tube has lots of surface area) that is colder than the dew point of the warm air. When air becomes colder and its temperature drops below its dew point, it must release excess moisture to reach its new, lower dew point (it releases moisture in the form of water, which appears on the colder surface).

My solution to the condensation problem is to minimize the difference in temperature, by venting the warm air trapped in the sealed tube and caused by sun light hitting the surface during the day. Had I placed the WalkAbout on the roof during the summer, I would not have discovered the condensation issue, because the difference in the temperature is not as extreme as during the winter months and would not have reached the dew point.

basic layout of new solar stand

This is the basic layout of my 3rd generation solar node, to test stability in high winds I'll bring a complete setup down to the airport and place it behind my old Cessna and see how it stands up to prop wash.

Indoor vs Outdoor Antenna

Because condensation blew the voltage controller package in my WalkAbout setup, I am using the opportunity to check the performance of two different antennas (one indoor and one outdoor omni) which state they both have a 9 dBi rating.

vertical rf patterns

This graph was generated by using photoshop to scale and superimpose the two vertical RF patterns on top of one another.

Real world usage indicates that this graph seem accurate. With the outdoor 9 dBi omni which has a flatter vertical RF pattern a weak meshed link can be made with the first roof tube antenna setup (approx 600 meters distance) on a regular basis, the indoor 9 dBi omni (with a fatter vertical RF pattern) in the WalkAbout configuration could not link up on a regular basis. Signal strength within the building is much better with the indoor antenna (with a fatter vertical RF pattern) than the outdoor 9 dBi omni when placed on the roof in the same location.

GH network - geographic layout
GH network - mesh signal strength(s)

So if the objective is to provide the best RF coverage to users right under a node placed on a roof at the lowest cost, then this little test looking at the data of an outdoor 9 dBi omni connected to a 4 foot long LMR 400 cable with a RP SMA to N pig tail, with a Meraki vs an indoor 9 dBi antenna with a direct connection to a Meraki put into a housing built out of Sch 40, shows the latter better choice in terms of lower cost, lower signal loss (due to fewer connections), and lower complexity.

Solar WiFi, Digital Divide, VoIP
and Meshing Links

Homebrew Non Penetrating Roof Mount Ideas

Digital Divide in Africa
$200 Net PC to Close Brazil's Digital Divide
Using WiFi to Bridge the Digital Divide
Wireless Community Summit Tackles Digital Divide
VoIP Booming in Africa
India Rejects One Laptop per Child Program
Solar Wi-Fi To Bring Net to Developing Countries

Meraki Hardware Hack
Meraki Mini vs Fonera
Dema FON Blog, indipendent Fonera review
CUWiN Completes Port to Meraki Mini
Mesh Protocols and Implementations


home page button