(Summer ’12) Esther Jang ’14

PSC Fellow. Esther will travel to the Orkeeswa Secondary School in rural Tanzania this summer. She and her project partner Jessica have developed a way for the school to receive wireless Internet access by extending the cellular data network 7 km from the nearby town of Monduli. She hopes to increase the capacity of the school by setting up Internet access for its new computer lab, and also bringing over electronic teaching resources donated by MIT professors. Be sure to keep up with her updates!

Internet Manual for Orkeeswa

8/24/12

tl;dr: TO RECHARGE THE INTERNET CREDIT, take the sim card out of the shiny white VodaFone USB modem and put it into the regular Airtel USB modem. Then plug the regular Airtel modem into the computer, use an Airtel credit voucher to add money, and choose a (volume-based) bundle as usual.

How the system works:

The basic idea is to improve the reception of the USB internet sticks (modems) we normally use to access the Internet. The large antenna mounted on the roof is connected to the USB modem via a large data-carrying cable. The USB modem plugs into a 3G router, which takes the Internet access provided by the sim card inside the modem and produces wireless signal which can be picked up by multiple computers in the area. Since wireless signal does not penetrate solid walls very well, the wireless hotspot created will usually be limited to a single room.

Setup:

The antenna and cable are mounted permanently on the building. One end of the cable is connected to the antenna, and the other end should be connected to a small converter (a short piece of cable with one large end connector and one tiny one) that should be plugged into a port on the side of the shiny white Huawei USB modem. The USB modem should be plugged into the USB port of the black D-Link wireless router. The router needs to be plugged into power in order to work.

How do I know it’s operating properly?

INDICATOR LIGHTS:

USB modem:

When the modem is plugged into and receiving power from the router, a light on the top of the modem should be either blinking or steady. A blinking light means it is detecting or attempting to connect to a network, and a steady light means it has successfully registered on a network and is connected to the Internet.

If the light is green, the modem is connected to a 2G (GPRS or EDGE) network, and the connection may be very slow. If it is light blue or dark blue, the modem is connected to a 3G (HSPA/HSDPA or W-CDMA, respectively) network, and the Internet should be fast. The light blue light indicates the fastest type of network.

Router:

When the router is plugged into power, it automatically turns on and begins operating. There are several useful lights on top of the router:

1. The leftmost light, a circle with a check mark inside, is the system check light- it should be green and steady or blinking very slowly at all times, meaning the device is receiving power.

Troubleshooting: If the system check light is not on, check that the power cord is firmly and stably plugged into the socket. If so, check that the extension cord is working, and then check that the power in the room is working.

2. Directly to the right of the system check light is the Internet connection light, which looks like a picture of the earth. When data is being transferred over the Internet connection, the light flickers. Otherwise, as long as the modem is connected to the Internet, it is steady. When there is a 3G connection, the light is green; on a 2G connection, the light is orange.

Troubleshooting: If the Internet connection light is not on, check that the USB modem is plugged firmly into the USB port in the back of the router. If so, check to see if the sim card is still inside the modem. If it is, the sim card may be out of credit. See the credit recharge instructions below, under Maintenance.

To the right of the Internet connection light is the wireless light, which looks like 3 curves next to each other. When data is being transferred over the wireless connection (e.g. when the router is providing Internet to a laptop), the light flickers. Otherwise, as long as the wireless is turned on in the modem configuration, the light is steady.

Troubleshooting: If the wireless light is not on, the wireless capability of the router may have been turned off (allowing Internet access only via Ethernet cable). You can turn the wireless back on by changing the wireless setting in the router configuration.

Specifications:

Antenna: 2.4 GHz 24dBi antenna with an N-type female connector on the end of the cable. Purchased in the US from ZDA Communications for $65 including shipping and brought over in a suitcase.

Data-carrying Cable: LMR-400 coaxial cable with an N-type male connector on one end (mated to the N-type female antenna), and with an N-type female connector on the other end (inside the building).

At the Computer Lab, the cable is a single 12-meter (40-ft) cable.

At the Laguna Beach House, one 12-meter cable was not long enough to reach from the antenna into the office. There is a 12-meter cable attached directly to the antenna, and another 6-meter (20-ft) cable attached to the end of the first cable that extends into the office.

Cables were purchased online from Amazon in the US for $20-$50 each depending on the length, and brought over in a suitcase.

Router: D-Link Wireless Router 113 (DWR-113), purchased for $70 (110 Tzsh) at Sound & Vision in Arusha.

USB modem: Huawei E160E, originally from Vodafone. Purchased from Ebay UK for $37 and ordered to the US, shipped over using DHL or brought over by a volunteer. If you need to use the modem with your computer for some reason, and your computer tells you to install the pre-loaded Vodafone software, please DO NOT do so! The modem is unlocked already, and should be recognized by the older (red) Airtel/Zain interface.

Maintenance:

Data Recharge:

You will not be able to recharge the credit on the sim card using the unlocked VodaFone USB modem. Take the sim card out of the unlocked modem and put it into a regular Airtel USB modem. (There should be one in the router box.) Then plug the Airtel modem into a computer, use an Airtel credit voucher to add money, and choose a (volume-based) bundle as usual.

Router Configuration:

To configure the router, turn the router on and connect an Ethernet cable from one of the Internet ports in the back of the router (numbered 1 through 4) to the Ethernet port on your computer. Open a web browser such as Mozilla Firefox or Google Chrome and type in the URL window: http://192.168.0.1; this is the IP address of the router. If it doesn’t work, check the Ethernet connection on both ends, and try refreshing the browser a few times. The URL should take you to an admin login page that requires a username and password.

User Manual:

Documentation for the router can be found on Computer X (to be filled in later), or with Andy the volunteer if he is around.

One Step Forward, Two Steps Back

8/24/2012

Goodbyes are so hard.

8/18/12

The signal reception at the current height of the antenna turned out to be fine (consistently 3 bars HSPA). Perhaps I was being paranoid… but then again, maybe I could have gotten 5 bars.

8/17/12

I suppose the setup of the antenna at the school computer lab/library building is now finished, but I’m not quite happy with it. The reason is that the antenna is mounted not quite at the height of the peak of the roof, which may well be too low to get a very fast 3G connection (HSPA/HSDPA as opposed to W-CDMA). During testing, I found that the height at which I started receiving HSPA/HSDPA signal was around two feet above the peak of the roof. I had assumed when we started building that since our 20-ft pole had been more than high enough to reach the peak at the Laguna Beach House, it would be the same for the other building. The buildings look similar enough; they are both one-story concrete boxes with corrugated tin roofs. But as it turns out, the computer lab building’s roof is steeper (which I noticed briefly during testing but didn’t think it important), and the wall under it is slightly taller, so where we wanted the antenna is slightly higher than where we could safely mount it with proper lightning protection.

I should have measured beforehand how high the pole would come up above the roof, but instead I rushed ahead with construction without checking, which was foolish. If the Internet in the computer lab ends up being significantly slower than at the Laguna Beach House, I will take pains to redo the entire setup, probably to the great frustration of Elifuraha the fundi. But it should not take more than three school days- the first day to remove the grounding system and the new concrete base for the antenna pole, and the second two to either extend the antenna pole by splicing on another piece of pipe under it, or cutting down the pole and re-mounting it atop a small, high-up ledge on the side of the building that I hadn’t previously considered as a good base for the pole. After looking at it a second time (and considering that the Internet might be half the speed it could be because of the missing height), I think we could make it work.

I’m glad that we at least got 2 feet of extra height because I noticed my mis-assumptions when I went to mount the antenna. We had already attached the grounding wire to the pole (a 20-ft long, 2-inch diameter black pipe) before putting it through the hole we made in the tin roof. (The choice of location for the hole was at least obvious, because there was only one place on the edge of the roof that was not covered with plaster.) There was a limited length of ground wire at school- just a little over 20 feet (the piece that was left over from the 50-ft piece we had cut for the LBH ground). So at the time I thought it would be fine to raise the antenna just a few feet by placing cinderblocks underneath the base and then turning that makeshift base into a big concrete block at the end- then it would still be possible to use that piece of ground wire. In the end, we took a piece of pipe from the storage container that had been left over from the LBH installation, cut it to about 2 feet, cut one end into four pieces which we splayed apart, stuck the bottom of the antenna pole inside it, and put concrete in the bottom of the 2-ft extension.

 

Internet at Mungere

8/12/2012

Today I’m in Mtowambu- the name breaks down to Mto wa mbu (River of Mosquitoes). There is quite a lot of water here; it’s in the Rift Valley, so water flows down from the escarpments on either side, and the water table is high. They even have rice paddies! Thankfully, we have insect repellent and bed nets for the mosquitoes.

Jessica and I are visiting another volunteer-based school in Mtowambu called Mungere, where our host brother works. So far, it seems to not really be built yet; there’s only one teacher’s office and two classrooms, but no toilets or kitchen or anything else. The first year of students has been chosen- there are like 40 “preform” students, or students who are undergoing crash-course English training for a few months before staring Secondary school in total-immersion English. (In government schools, the preform period is only about 2 weeks; no wonder students generally fail to learn the material. Even the Orkeeswa Form Fours this year can’t read their textbooks effectively, and they’re the best and brightest from the region.)

A team from Engineers without Borders has come out for two weeks to build a set of composting toilets and make plans for the solar system. Afterward, they will climb Mt. Kilimanjaro for a whopping $1450 per person. Jessica and I are invited, but the price is a bit steep; plus, we’ve recently been asked by Orkeeswa to install a second Internet setup for the student computer lab, so we’ll have to get on that. In any case, the founder of this school wants us to come back, and one of her lures was to throw us in with all these lovely engineers with different knowledge sets- I must admit it’s working pretty well. One of them is Jeni, who used to work in aerospace, has her own solar energy company, and is in charge of setting up the solar system. Ada is a civil engineer who just has a huge amount of energy and always itching to get as many things as possible done right now- she designed the composting latrines, and she and her fellow engineer Valerie are in charge of getting those built within the next 2 weeks. And there’s a really cool techy French guy named Anatol who also worked as an aerospace engineer but is now a programmer. I got to introduce him to the Fieldtest app, which allows you to find out details about the Internet connection on an unlocked, jailbroken iPhone. He also has experience designing solar systems, so he and Jeni have somewhat opaque discussions sometimes- but I ask lots of questions so it’s not that bad.

It’s really nice to have Junior around again! He’s such good company, and always introduces me to people as his littlelest sister (which is slightly annoying, but in a nice way- I like having a sibling for once). I think I might really enjoy coming out here again to work with this school. But I also really miss home.

8/13/2012

Internet at Mungere seems possible:

1. I know where the cell towers are that provide what weak 3G signal there is on the roof of the first classroom block. Although at its best, the 3G signal strength was only -98 dB and around -110 dB on average. The towers are atop the part of the escarpment of the Rift Valley that overlooks Lake Manyara. They are hidden from view at the school behind a portion of the escarpment that juts out between the towers and the school.

2. I observed that there is much better 3G signal (-70s to -80s dB) less than half a kilometer away from the school, where the line of sight to the towers is not blocked. Perhaps just a little more height off the roof would be enough to improve reception to that level; or maybe a building can be built within the campus grounds in a spot with particularly good reception.

3. There will be two decently powerful antennas left over from this project; if I am planning on coming back to implement an Internet system at Mungere, it would be convenient to leave at least one of the antennas with them (and not have to deal with TZ customs again).

Internet at Orkeeswa: August 5, 2012

We have wireless Internet access at the school!

Yesterday we finished mounting the antenna permanently in the roof of the Laguna Beach House. We drilled a hole in the corrugated metal roof on the side closest to our target tower, and chipped a large hole in the concrete wall against the building for the pole to sit in. Then we secured the cables to the underside of the roof along the side of the building using really cheap but effective flexible metal cable clamps called “Elegant Clips,” which were nailed into the rafters and bent around the cable. We attached the cable along the wall until we reached a convenient flower-shaped vent on the side of the building, which we widened a little bit to thread the cable inside.

Since our 40-ft LMR-400 coax cable wasn’t quite long enough to get from the antenna into the office, we were planning to use a 40-ft and a 20-ft cable in series. It turned out really nicely that the first 40-ft cable reached just far enough to go into the vent, so the connection with the 2nd cable was inside and taped to the ceiling. Even so, I will probably seal up that connection with some gaffe tape and epoxy in case any rain drips in through the vent and decides to take a walk along the cable.

Then we drilled a hole in the office doorframe, threaded the cable through, and secured it with some more cable clips. The end of our cable will eventually be gaffe-taped discreetly to the bottom of the wall, but first we have to receive the crc9 to N-type pigtail (connector cable) so we can tape everything down as nicely as possible. People tend to appreciate nice-looking things more.

To recap: the free end of our 2nd cable was attached to a N-type male to stripped wire pigtail, which was plugged into the crc9 antenna port on the circuit board of an Airtel 3G modem stick. The modem is plugged into a wireless router, which sits on the concrete sill of a glassless window between the main administrative office and the main conference room; this successfully gives both rooms wireless Internet access. Later, a cabinet might be installed on that sill that opens into the office; this might weaken the signal to the conference room, but probably not as much as a concrete wall would.

When the whole thing was set up, I tried to show Peter Luis how well it worked. The first time J and I had tested the wireless in those two rooms on Tuesday, we had been able to browse quickly on separate computers at the same time; at one point, she was downloading videos while I was on gchat. Of course, it didn’t work the first few times he tried loading gmail on his computer. But it was working fine on mine, and we determined that Peter’s computer was in general really slow. After about 10 minutes and restarting the browser a couple of times, he was able to access non-html gmail and his chat loaded- this doesn’t happen at most Internet cafes, even in Arusha. He was very impressed, until he clicked on a link one of his co-workers in the US had sent him about Snoop Dog claiming to be the reincarnation of Bob Marley. He seemed to get the impression that teachers weren’t going to use the Internet for useful things, and that IEFT wouldn’t be able to afford the data for teacher use. But we’re going to have a meeting about Internet use policy in about two weeks (since Peter will be away next week). Maybe we can set a time limit per teacher per day, to give them an incentive to use their time wisely while not detracting too much from their productivity if they choose not to.

Next Monday we will be driving a grounding rod into the ground, and testing the reception with our big antenna on the roofs above the lower teacher offices.

0o0o0o0o0o0o0o0o0o0o0o0o0o

J was sick yesterday and most of today; she had a stuffy nose, nausea, and slight fever, and was really tired, so we rescheduled our second enrichment physics lesson for the science-oriented 9th graders, which was supposed to be on Friday. Today, after eating a ton of delicious potatoes (with rosemary) cooked by Mama for lunch, J feels much better but still not entirely well. If she still feels odd on Monday, we’re going to ask Peter to take her to get checked for malaria at Dr. Danny’s. But apparently people here generally get malaria like once every 5 years; it’s considered equivalent to a really bad cold/flu that requires special medicine. For people who have access to treatment, it’s not that bad, but for Maasai living out in the villages, it often goes untreated.

This morning I went to school on my own to help the Chemistry teacher Bendera with a Form 4 practical on reaction rate. The experiment was to combine a constant amount of HCl with Sodium thiosulfate (Na2S2O3) of various concentrations, all clear solutions, in a conical flask atop a piece of paper with an X drawn on the center, and use a stopwatch to time the progress of the reaction until the solution had become too cloudy for the X to be seen. I had prepared 0.1 M HCl and 0.05 M Na2S2O3 the day before, but on the first demonstration trial that Bendera performed for the kids this morning in class, we found that those concentrations were simply too low. It took 9 minutes for the reaction to occur with the highest concentration of Sodium thiosulfate. So we sent the kids out on break for a few minutes (at which point one of the kids actually went home) and whipped up some new solutions- 1 M HCl, and 0.5 M Na2S2O3. These worked fine; the reaction times ranged from about 10 to 60 seconds each. Nonetheless, it would have been really nice to have tested out this lab beforehand, or at least to have looked up online what appropriate concentrations would be for the reactants. The former might be really hard for a teacher who lives far from school and doesn’t have access to any of the classrooms outside of regular school hours, but the latter could be easy with Internet access, if it is allowed.

Finding bulbs for what are currently the sucking pipettes needs to be a priority; Mbayana, one of the Form 4 students, told me today after the lab that he expected people to be quite likely to accidentally suck liquid into their mouths during the official exam practicals due to anxiety. I am concerned.

I really like working with the Form 4 students; maybe it’s because they are around my age, generally seem to work very hard, and because I sort of like how cheeky they are- they have a lot of character. It’s hard to think that most of them will probably not pass the Form 4 exams with the marks to move on to Forms 5 and 6. But at least they will have the opportunity to repeat the year at Orkeeswa.

There is also a new program at the school starting next week called “Study Camp,” in which Form 4 girls will be staying for a few days each week until the exams in newly converted dormitory rooms at the Laguna Beach House. Girls have much fewer opportunities than boys to study at home because they have so many chores to do, so this could be a huge help to them. J and I are excited to stay a few nights at Study Camp to help tutor and watch over the girls.

0o0o0o0o0o0o0o0o0o0o0o0o0o

Pole sana! The charger for our huge battery pack burnt out three weeks ago when I unwittingly plugged it into the 220-volt outlet without knowing that it was only rated for 120-volt input. The sad thing is that I didn’t find out until the battery pack ran out a few days ago; we’re going to look for a charger with the same specs in Arusha, but we’re not too optimistic about finding one.

0o0o0o0o0o0o0o0o0o0o0o0o0o

Alex (our awesome office manager) left for England forever to go propose to his girlfriend. We are so happy for him, but will miss him greatly!

0o0o0o0o0o0o0o0o0o0o0o0o0o

Last week we went to see a newborn baby! Her name is Swabra, which is a Muslim version of Subira, which means Patience. She is the newest daughter of Jessica’s host family from last year, and was only 14 days old when we went to visit:

I also drew a picture of my host brother. (I also did one of J and one of Malkia, but those are unavailable at the moment.) Oddly enough, for the first time I’m finding it easier to draw people from real life than from pictures. I think my drawing style is changing.

Kichaa (madness): July 29, 2012

The corruption in this country is like a great parasitic growth- one of the kind that you see on tropical trees that starts out as a small, vinelike mass feeding off one of the branches and then extends its roots downward, curling around the trunk of the tree and sucking it dry of nutrients until all that is left is the parasite, forming an empty shell in the shape of the original tree which has long since rotted to nothing.

Today on the news was a story about five top government officials who embezzled over 3 billion TZ shillings from the government using a scheme involving electricity. The recent power shortages experienced all over the country in the last month have apparently been because the officials were re-allocating the funds for electricity provided by the government to their own Swiss bank accounts, and cutting electricity off from the citizens under the guise of technical problems. On the news it was presented as a huge scandal, and rightly so. But the discussion it opened up at the dinner table tonight revealed such insanity that I find it both difficult and necessary to put it into words before it disappears from my mind.

This is not an uncommon occurrence. Not long ago, there was another incident where the government ordered and supposedly paid for several very expensive pieces of equipment from the UK, including a huge radar receiver that was supposed to be able to pick up all of the air traffic in Africa south of the equator. The shipments crossed the border; they arrived, and were unpacked. Only then was it revealed that most of the boxes were packed full of nothing but nails. The radar did arrive; it was unpacked, and set up. It never worked; it was a fake. The whole order had cost hundreds of millions of shillings.

Even in the small town where we are staying, there is a company that turns the taps to let water flow for a few minutes each day into their local customers’ water tanks; sometimes, the people in charge of doing this will stop the water flow to certain houses and extract a bribe before turning it back on. Those who do not want to pay the bribe must deal with frequent water shortages.
——-

Today J and I went to school to help with the entrance exams for Orkeeswa Secondary. The top primary school students (Standard 7) from all around the Monduli region came to take the test; this year, 228 students roughly between the ages of 12 and 25 registered for the exam, and all but one showed up. The test was composed of a spatial reasoning section, a basic math section (some sums/multiplication and a few very simple word problems), a “pick the one that doesn’t belong” section, a reading comprehension section in Kiswahili, and a short passage in English with 5 blanks to be filled in. I expected most of the students to do fairly well on all of the sections excluding English, since the questions looked like they should be easy for sixth and seventh graders. Instead, the average score of these handpicked top students was around 30/50. The quality of their education is very low and dropping every year, as teacher salaries and funding for schools continue to be cut.

There are a little over 300 seats in the Tanzanian Parliament. Around 40 of them are currently held by Chedema, and a little less than 100 of them are held by CCM members who have only Standard 7 primary school educations or less, and were elected to office by bribery and/or support from their wealthy and powerful parents.

So how do they get into office with so little education? In order to get the votes of the approximately 70% of the population who live in small towns or in semi-nomadic tribes in the bush, members of parliament have used their wealth to send people around to the rural houses and bomas with new clothing (kangas, which are only around 8,000 shillings apiece) and other small gifts for families in order to gain their votes. Traditional men living in bomas are often polygamistic and may have around 20 to 100 children each; also, if the man of the household makes an agreement to accept a gift and vote for a certain politician, it is culturally unacceptable for him to go back on the agreement, and he is obliged to order all of the members of his household to vote the same way. But since none of them know the importance of the government, and may not even realize that they are paying taxes and bribes for their goods, they are happy to accept the small items in exchange for their votes. It is like when Native Americans sold New York for pelts and beads all over again, but is happening now, every five years before an election.
——-

Starting Monday, there will supposedly be a nationwide strike of the Teachers’ Union to try and get a 100% increase in their wages. It may not happen at all, because government officials like school inspectors, as well as some teachers, will inevitably go to work out of fear for their jobs, their pensions, and their children. But apparently sometimes strikes have worked in the past to pressure the government to open up wage negotiations, so I wish them all godspeed.

Today J and I sat down on the edge of the soccer field in the middle of town to relax and read. Many small children, maybe around 8 or 9 years old (Standard 3 in primary school) who had been playing soccer in the field flocked up to us and began talking to us, asking us about what was in the books and eager to read with us. After a while, all of the books we had in our bags- two Kiswahili study books, and Form Four Physics, Chemistry, and Biology books- were out and distributed among the children, who were flipping through the pages excitedly; it was so incredibly adorable that I didn’t know what to do with myself. One of the Standard Threes, Baraka, was attempting to read the English words in the physics book and asking questions about the diagrams, so we explained them as best we could in our slightly broken Kiswahili at a level that an 8-year-old might understand. After about two hours, we trudged home up the mountainside, all the heavier for the knowledge that these children had probably never had such ready access to books before, and certainly didn’t have anyone to sit down with them and read. These are the people the government is stealing from; it is edging its people further and further into poverty; stealing its water and electricity, raising food prices, worsening the quality of education, making it difficult for impoverished parents to nurture their young children at home or even send them to school. J and I talked for a long time about the logistics of maybe opening up a day care program for very young children in town, to give them the boost in their early development that might make it easier for them to learn in school when they’re older. But this, like Orkeeswa, would only be another small patch on the huge, gaping tear that is the corrupt, backwards government in this country.

I often wonder whether I am working on the right thing here. Every time I turn on the lights, flush the toilet, or buy bread from the store, I feel a surge of bitterness that can only be a fraction of what the young, educated people in this country feel; these things should be easier to get, and people should be paying less for them. Their daily lives should be easier, so that development can actually happen. Someday, they say. But when is that?

Interesting: July 25, 2012

In Kiswahili, the word “interesting” is apparently not used as in English- the diplomatic, malleable, all-purpose phrase that is so useful in everyday conversation. I really miss it during my feeble attempts to fumble through conversations in Kiswahili. Instead, I’m limited to words like “nzuri” (great/beautiful/lovely) and “sawa” (okay), which just doesn’t say the same thing!

Interesting is definitely what I would call the past two days of testing. Yesterday, J and I started taking measurements of the Laguna Beach House, the administrative building we want to get Internet access to first. We were considering several different places on the building where we might want to clamp the pole for mounting the antenna. Working under the assumption that providing Internet to any of the rooms in the building would be equally good, our first consideration was where reception would be the best. We suspected that the corner of the building closest to where we thought the 3G tower was would be where we got the best reception (right next to the wall on which we’d tested before), so we were planning to run a short coax cable from the antenna mounted there into the nearest room, which is currently an empty storage room but perfectly usable as an office or small classroom. We wanted to avoid having too long of a coax cable to avoid cable losses, and also wanted to avoid allowing the signal to travel over the corrugated metal roof, which could cause multipath (destructive interference caused by reflections).

With the help of Alex, Andy, Ming, and at one point Simon (a physics teacher), we moved a very tall and somewhat sketchy handcrafted ladder from its storage place by the kitchen up the hill to the cemented outdoor courtyard of the Laguna Beach House, and I used it to get onto the roof with my Macbook, a 3G modem, and the 3G ¼ wavelength monopole antenna. I got speeds of 0.9 Mbps download and 0.3 Mbps upload with 3 bars on the HSPA network on the roof next to the wall; further back on the roof right above the main administrative office, I got 0.5 Mbps download and 0.16 Mbps upload with 1-2 bars on HSPA. There was definitely better reception in the first location, although both seemed acceptable.

Unfortunately, Alex told us afterward that it would be really inconvenient to use the Internet in that closest storage room as opposed to in the regular office/meeting room, because that room might be converted at some point into a temporary dormitory for students studying for the Form 4 Exams (scary tests that decide one’s future- more on that later). So we’d have to make sure our coax cables made it into the main administrative office; it’d be a trade-off between the accessibility and the quality of the Internet connection. We measured how many feet of cable we would need to connect the antenna mounted where we had originally planned to a USB modem in the office: around 65 feet. Our two longest cables were about 39 feet long, so we could connect two of them together and it would reach all the way back. But would cable/connector losses significantly slow down the Internet speed? I was inclined to think not, because of the high quality of the cable: LMR400, which has a loss rate of some .06-0.7 dB/ft. The total loss should only be about 5 dB- less than half the attenuation caused by a single average-sized tree in the line of sight of the antenna.

To verify the effect of the cable losses, today we tested the Internet connection again on the roof, with the monopole antenna on the end of the N male pin of two 39-ft coax cables connected end to end. On the other side of the cables, we had our modified N-male to stripped wire pigtail plugged into a USB modem in the office. Oddly enough, we had quite a bit of trouble getting on the HSPA network in the morning, and also seemed to be getting very variable WCDMA signal strength (at times fluctuating between 0 and 5 bars). I started out sitting on the apex of the roof (at 15 feet above the ground), but moved to mid-height on the roof because I wanted to see if mounting the antenna slightly below the apex level would still be okay. At Meena’s Hardware in Monduli where we were planning to get our mounting pole, we’d found out the previous day that we wouldn’t be able to get two sizes of pole that fit exactly into each other and assemble a two-piece 15+ foot mounting pole as we’d planned to do. So the shorter our pole could be, the easier (read: more possible) it would be to transport it to the school.

It turned out that as soon as I started sitting in the middle of the roof, the Internet in the office (where J was monitoring the results) slowed down drastically, becoming almost unusable. But even when I’d been sitting at the apex of the roof, the network had been flaky, switching between WCDMA and HSPA, and speeds had been comparable to those in Monduli: around 0.1 Mbps download and upload. After J and I regrouped with some granola bars, I went back onto the roof with the 24 dBi 2.4 GHz antenna, planning to put the monopole antenna in front of the dish as before. We tested the speed right above the office with only a single 39-ft cable, with the same results as before. At the point by the wall where we had originally wanted to mount the antenna, reception was slightly better (occasional HSPA of around 2 bars with mostly WCDMA at 0-3 bars) but speeds were not significantly faster.

As a sort of throwaway last check, I decided to try connecting the 2.4 GHz antenna directly to the N-male connector of the coax cable, even though I didn’t expect it to be able to pick up 2.1 GHz signal as well as the monopole antenna. Surprisingly, the WCDMA signal strength jumped to a consistent 4 bars, and within a few minutes the network switched to HSPA with 3 bars. I propped the antenna up on my head, and the HSPA signal strength jumped to 5 bars! Height really matters. But the Internet connection was somehow still as slow as before. What a mystery! Maybe traveling through the cable introduced errors in the signal that lowered the data rate drastically without decreasing the perceived signal strength? Maybe there were just too many users on the network at the time when we tested it; but we’d never observed that before at this time of day. The Internet with 5 bars on HSPA today was somehow worse than with 2 bars on the wall on the other days we’d tested it. I now realize that what I should have done is test the speed at the location I was receiving full bars on HSPA with just my laptop, 3G modem, and monopole antenna to eliminate the effect of the cable. We will be sure to do this on Thursday when we next go to school.

The trouble is that in order to keep the pole length acceptably short, it would be best for us to mount the pole on top of the courtyard wall, which adds about 5½ feet to the height of the antenna. If the cables are the problem, we’ll need to mount the antenna right above the office in order to shorten the length of cable we are using. Trade-offs are hard.

Other exciting business from this past week:
On Saturday, we went to the wedding of our host mama’s younger brother, held in the village of their boma (family home/traditional hut) around 2 hours outside of Monduli. A few notable pictures here:

Today our host brother took us to see the Rift Valley, a very cool location on the outskirts of Monduli where continental plates are moving apart to reveal two sheer cliff faces and a huge, flat valley in between. According to J, this process is how new seafloor is created; in a few million years, the whole area will be in the ocean.

We saw a number of cool plants on the way:

This post is dedicated to Brandt: July 17, 2012

July 16, 2012

Three or four days ago, everything changed. We now have a reasonably reliable way of getting fast Internet at the school, sometimes at blistering speeds of around 0.3-1 Mbps (essentially the fastest we’ve seen outside of Arusha, and even faster than most Internet cafes there).

With the help of Brandt, a super awesome 3-week volunteer at the school who used to be a technical manager for Cirque du Soleil, we took apart two commonly used USB modems to see if we could attach an external antenna directly to the circuit board. On each board, we found a tiny female crc9 connector for an external antenna. Since we didn’t have a cable with that type of connector, we cut open an extra coax pigtail we had with one N male connector end (our antenna has an N type female end) and whittled down the wire to fit in the antenna port. Then we realized that the antenna we have was optimized to catch 2.4 GHz Internet signal, and not the 3G signal at 2.1 GHz. So we made a ¼ wavelength monopole antenna for the 3G band- basically a small piece of extra wire we had, cut to about 3.5 cm long (1/4 of the wavelength of the 2.1 GHz signal), and stuck it in the port. Like magic (except totally not), it boosted our signal at the school literally from no service to full bars on 2G, and sometimes with the Airtel SIM card, even 1 to 2 bars on 3G (which is still over 10 times faster than full bars on 2G). But the 3G service at first was flaky, and Brandt used an Airtel app on his iPhone to find out where the tower broadcasting the 3G network (HSPA/HSDPA) is. It’s a lot further away than the Monduli tower only 6 km away, which uses only 2G protocols (EDGE, GPRS, WCDMA).

We also tried using our largest antenna with the USB modem by cutting off the end of an extra RP-SMA to N-type coax pigtail that we had, whittling down the central wire and sticking it into the antenna port, and attaching the ¼ wave 3G antenna to the pin of the N-type connector with a piece of plastic insulation. Then we threaded the pigtail through the parabolic grid and wrapped it around the feed so that it was positioned at the focus of the antenna. Oddly enough, pointing this at the 3G tower only boosted the signal maybe one or two bars at most while we were on the 3G network- not as much as we expected for such a high-gain directional antenna. We are still not sure what is happening there.

But the main problem was that we had a very hard time connecting reliably to the 3G network in the first place. The monopole antenna, which was interestingly able to pick up both 2G (at 1.8 GHz) and 3G signal (possibly because 1.8 and 2.1 GHz are not that far apart), was already sensitive enough without the parabolic dish to receive very strong signal (-60s dBm) from the 2G tower. The parabolic dish only improved the 2G signal around 1 bar (bringing it up to -53 dBm). So we had to find a way to selectively join the 3G network; having the correct length of antenna is apparently not enough.

Luckily, the current Airtel USB modem interface for Macs has an option for choosing “3G-only,” wheras neither the Mobinil software interface on my unlocked Huawei E173 modem nor the Vodacom interface seems to have a “3G-only” or “HSPA/HSDPA-preferred” option. I might try to find new interface software that has the option and load it onto the modem if it’s compatible, but for now I won’t need to since the current Airtel interface seems to have it already.

All the first testing was done on the soccer pitch behind the school buildings, which we discovered was much higher up than we had originally thought- it actually has a line of sight to both the Vodacom and Airtel cell towers in Monduli. And it is a much more convenient place to have Internet access than the primary school. But while we were up there, we realized that the roof of the Laguna Beach House (the main administrative building at Orkeeswa) is at almost the same level as that soccer pitch.

Today’s testing was done sitting on the wall around the Laguna Beach House inner courtyard, facing out towards the 3G Airtel tower. The first major goal was to try to select the 3G network, which we did successfully using the Airtel USB modem interface. The second goal was to determine if soldering would work as an attachment mechanism for keeping the coax cable mated to the connector. Soldering into such a tiny connector is really hard! Ultimately we ended up soldering a thinner 2.1 GHz monopole antenna into the connector just because it was easier to deal with than the coax cable; the weight of the coax cable pulls against the solder in varying directions as it is moved. We were able to get a 3G connection with that antenna sitting on the wall outside, but not indoors. So we tried using the monopole antenna as a wire by twisting it around the coax cable inner wire, soldering the connection, and wrapping electrical tape around it. That connection was secure, so the other one broke off. In summary, we have to find a better way.

Tomorrow in Arusha, I’m hoping to order a cable with the correct mate (crc9 male) for that tiny antenna port on the USB modem. As a kind of experiment/test of the Tanzanian postal service, we’ll probably order two, having one sent to the States and mailed to us via FedEx, and the other shipped to TZ straight from the retailer.

A note past all the technical details that I’m trying to record scrupulously for future reference:

While J and I were sitting on the wall trying to get a 3G connection, a pair of curious students stopped by to ask us about what we were doing. When we told them that we were trying to get Internet access to the school, they first asked whether they (the students) would really be able to use the Internet, and we said yes. Then one of them asked, “When we have Internet access, what does that mean?” J explained that the Internet connects everyone in the world who has Internet access, and makes available huge stores of information that people have put there. The student asked, “Does that mean that if there’s a person in Australia, we can mail them there?” J replied, “Yes, through email!” Both boys seemed fairly impressed, and I think they would have continued to question us if they hadn’t been called away by what they’d been doing before.

Sometimes I get so focused on trying to solve technical problems that I miss opportunities to notice other areas of need and expand the scope of the project; I might not have thought twice about the incident. Fortunately, J was struck by the fact that they seemed to have no conception of what the Internet was and how to use it. We came to the conclusion that we should try to hold some seminars on basic Internet research skills like how to search for information online, and how to evaluate the validity of your sources. Maybe this would solve a common problem with teaching computer skills in developing countries that Andy and Ming once told us about: as interested and earnest the students are when they enter the classroom, the minute you give them free reign with Internet access, they will usually go straight to Facebook or download pictures of their favorite celebrities instead of being able to use it effectively for gathering information. But maybe that’s just the nature of high schoolers. We’ll see when we get there.

Crazy Times: July 10, 2012

As with most stories, there is a short and a long version of the events of the past week. The short version is that we got USB modems from Airtel and Vodacom, tested the speed of the Airtel network at the upper and lower staff houses, and found that the network at the lower staff house was faster. We then got a 3G router in Arusha that you can plug a SIM card directly into, since our TP-Link 3G router from the US would not work with either local ISP’s network. (I’m still not sure why, because the 3G modems themselves seemed to be recognized by the router correctly. But I decided it was worth buying a compatible router to save myself n days of possibly fruitless troubleshooting out of my ~7 weeks left here.) We put together all of the equipment and tested a one-hop link of a few meters down the road in front of the lower staff house. The broadcast connection was around 0.07 Mbps upload and 0.05 Mbps download on average, and the received connection was only a bit slower than that (forgot to record numbers). We also tested 2 hops at a short distance; the received speed was 0.03 Mbps upload and 0.01 Mbps download- still usable! Then we tested both USB modems at the Monduli office (thanks to J, or I would probably have forgotten about it’s existence), and with the Voda one found faster speeds than we had ever achieved in Monduli: 0.11 download and 0.9 upload. In summary, we should probably use Voda and broadcast from the Monduli office.

Meanwhile, new possibilities for overcoming the blocking hill have been emerging quickly over the past week. Andy and Ming, the super awesome older couple teaching computer classes at the Orkeeswa Secondary School, gave us the idea to get Internet to the Orkeeswa Primary School, which sits right on top of the blocking hill. Unfortunately, it’s not run by IEFT, and it’s very small and run-down; some of the classrooms have broken windows because apparently the young, misguided students play by throwing small rocks at them. :/ But it’s ok, because it turns out the school is at the perfect location (we can see Monduli town very clearly from there), and definitely more secure than the surrounding wilderness. We were even able to mark its location using our GPS units, which should make aiming the antennas really easy- we just aim in the direction our compass points.

Another thing we found: you can access the circuitry inside the USB modems really easily by taking off the back cover. I discovered documentation online of two people who had modified their USB modems by soldering coax cable straight onto the antenna leads on the circuit board, and today J and I opened up one of the extra USB modems we have (extra because the Vodacom modem is the same model, Huawei E173, as the unlocked one we brought from the US- no coincidences there!) and found what we think are the locations we’d need to solder our cable onto. Essentially, we’re thinking about attaching one of our huge antennas straight to the USB modem to make it way more intense than it was ever meant to be! We’d have to point it directly at the tower for whatever service provider we’re using, but it’s easy to find out where those are because they stick out above all the other structures in the landscape. So there’s another possible backup plan, although an unlikely one.

The long version is kind of crazy, so I’ll try my best to describe it in a coherent manner.

——

We got an Airtel and a Vodacom USB modem in Arusha sometime last week, and were only able to register the Airtel SIM card because the network in the Voda store was down (which certainly boded well for future performance!). Both Airtel and Voda require a photo ID to photocopy and keep with the SIM card registration, but our passports were being processed for volunteer permits, so we were told to come back to the store with them within the next few days. We came back to Arusha on Saturday, but everything was closed for either the weekend or a “holiday” that no one outside of Arusha seemed to know about. (What gives??)

On the dala-dala ride back that day, I sat a tiny little girl on my lap because the woman next to me had been trying to fit her two children and some groceries on her lap, and I had only a backpack on mine. I found out through some broken Swahili conversation as well as J’s much better Swahili conversation that the tiny little girl’s name was Fausta and she was 3; her sister Lea, sitting on her mother’s lap, was 7. At some point, Lea decided that she liked my hair, and therefore I was going to be her bestie and Fausta could have J as her bestie. Fausta was indignant and wanted to be my bestie too, but then Lea sold J pretty hard (“But look, this one has a bag!”) and all was well. When the dala-dala ride was over, the mother explained that the little girls wanted to invite us home as guests. Lea and Fausta took us by the hands and led us all the way to their home, which was a room just large enough for two twin-size beds, a couch, and a TV. They fed us oranges (machungwa), and afterward the mother charged me to speak more, to improve my Swahili. >.

Yesterday, J and I went back to Arusha to see if the 3G router we’d asked about at an electronics store had come in yet, as well as to go on a grand scavenger hunt for physics materials requested by Simon, the main physics teacher at Orkeeswa. The materials included arcane objects like a resistance box and a rheostat- two kinds of variable resistors for classroom demos. They also included “inextensible string” (which we learned today meant fishing line) and pendulum bobs (which we haven’t found yet).

After getting our photo IDs added to the SIM card registration records at Airtel and Vodacom, we stopped for lunch at an incredibly authentic Japanese restaurant a short way down the road toward Moshono. (All of their ingredients seemed to be imported from Asia, and there was even a Japanese toilet! There was a definite sense of homesickness about the whole setup.) There happened to be free wifi, so we tried it out and found that it was faster than any Internet connection we had ever encountered in Arusha! After eating, we asked one of the waitresses how the Internet was set up, and she showed us a Vodafone (not Vodacom!) 3G modem/router that took a Vodacom SIM card and created a wireless hotspot; apparently the owner had gotten it at a Vodacom store for 180,000 Tzsh, or around $120. Expensive, but possibly worthwhile for that fast an Internet connection. I dragged J around to three different Vodacom stores, and unfortunately none of them knew remotely what I was talking about.

Later, after pestering the owner of every store along the main road in Arusha that seemed remotely relevant, I found out that there were 3G routers at Sound & Vision, a much bigger electronics store that imports all of its wares from Dubai and sells in dollars. J realized that the 3G routers with a slot for a SIM card alone would probably have better reception than regular 3G routers that take a USB modem, so we bought one of the former. Via further pestering, we also found out about this tiny but very promising-looking lab supply store hidden away in a building with a lighting store and an arcade; both the resistance box and rheostat were acquired- totally unexpectedly! I also haggled on the street for a water heater for the science lab, and brought down the price from $10 to around $4.50. Ah, the little things in life.

Today’s testing results with the new router are at the top of this post. For me, the moral of today is that common knowledge (e.g. that Airtel works better than Voda) is often wrong, and also that I shouldn’t rule out possibilities in my head and forget about them until I’m sure that they should be ruled out right now as opposed to later.

After the work was done, we visited Lea and Fausta’s home, but no one was there! Instead, there were four very adorable little girls playing in the yard outside their home. I felt very awkward and intimidated for a while, as the little girls stared at me expectantly and said things in Kiswahili that I didn’t understand while J was calling Lea’s mother on the phone. But then one of them came up to me and took my hands; I took that as an invitation to play, and picked her up and spun her around! 🙂 Then everyone wanted me to spin them, which I did until I was completely tired out. Afterward we played a bit of confused two-ball goalless soccer, and J and I picked the girls up some more and talked to them as best we could. (Lucky for me that little kids can have fun without language.) Eventually we left just in time to get home by nightfall.

J and I have started to sing together on the dala-dala; we’re currently trying to learn lyrics and harmony parts for songs by Jason Mraz, Keane, and the Eagles. There’s also this incredibly fun song in Kiswahili about the benefits of education that we’re trying to learn, but the lyrics for that will be somewhat harder.

First Days: July 3, 2012

Saturday, June 30th

Another late post, because I didn’t have the adapters to charge my laptop in the Amsterdam airport.

Tanzania so far has been really surprising. It is fairly cold since we’re up in the mountains and surrounded by trees (oh no!). What strikes me most is how friendly the community is; many strangers welcome us to the town in passing, and everyone we have met has been eager to offer help and/or friendly conversation.

For breakfast this morning, the family (Mama Lukamai, Helen, Pendo) gave us bread (mkate), peanut butter (peanut = karanga, butter = siagi), tea (chai) with milk (maziwa) and sugar (sucari), and roasted/salted bananas which are delicious and remind one of salted seaweed. The food here is very simple but flavorful; they don’t use any spices except for a bit of salt, but the combinations of vegetables taste pretty amazing. P would like it. :p

For dinner the first night, I had beans, rice, and cabbage very finely shredded and pickled. Then for lunch today we had potatoes boiled with some peppers, along with avocado slices and leftovers from last night. Tonight we had beans, chapatti, avocado slices, and potatoes. The beans (maharague) are straight from the field; before being cooked, they are a very light purple color and grow in tan pods – we saw them drying outside in the yard.

Over dinner, I found out that Mama Lukamai is a government school teacher and grandmother (of 3, I think), and she’s taking her final two exams for her Master’s in Education next Monday. Her daughter Helen is an IT teacher at the Maasai Girls’ School, a local boarding school. Many of her other children went to university abroad, and several are teachers and engineers in various places throughout Tanzania.

Kiswahili I learned today: (much fewer than what I asked how to say, but these are what I could recall afterward)

Shoe: kiatu

Pair of shoes: viatu

Stove/kitchen: jiko

Wood: cuni

Potato(es): kiazi (viazi)

Metal: chuma

Thermos: chupa

How do you say: Unasemaje

I don’t understand: Sifahamu

I don’t speak: Sisemi

I don’t know: Sijui

Store: duka

Face(s): uso (nyuso?)

Tree(s): mti (miti)

Person (People): mtu (watu)

Meat: nyama

Animals: mnyama
—

Pole: Sorry (for your hard work/harm) or bless you

Ni: is

Kwa: for

Ya/za/la: of (this is confusing)

Au: or

—-

We were supposed to be picked up by Quinn from IEFT today and shown around Monduli, but she was busy meeting with trip leaders from the Groton school, a private high school in MA that brings a lot of students to visit Orkeeswa every year, and were scheduled to arrive today. The Groton families are an important source of funding and sponsorship for the students, since tuition is free.

Instead, at Staff House 1 (the “top house,” because it’s at the top of the hill) we wandered in upon Andy and Ming, volunteers (originally from England, but most recently from Cambodia) who have been setting up the computer labs. They live as technological nomads, having few possessions and wandering the earth as they see fit, volunteering their expertise to those in need. Very cool.

We checked out the top house for a place to put/mount the antenna. There is a tall wooden power line pole in the back of the house that we could use if there is enough cable. There seems to be a clear line of sight from there through the trees roughly in the direction of the school, but we can’t be sure until we have our GPS units working. There is also a water pipe to attach our grounding wire to, and ventilation holes in the walls (though we’ll have to seal up the break in the screen we will have to make). If we can’t use the tall pole, we’ll have to mount a pole of our own either on the side of the house or some distance from the house, with a sturdy base somehow secured to the ground.

We are also considering using Staff House 2 or the Monduli office as our starting point for the Internet signal, since rumor has it that reception is slightly better there. We will have to check it out.

Afterward, we walked into town with Andy and Ming. While at the grocery store, we happened to see Peter’s truck (with Quinn in it) drive up the road towards the staff houses, so we walked back up the hill to Peter’s house to see if we could catch him. He was still in a meeting at the staff house when we arrived, so we left a note on his door asking if we could exchange some USD for Tzsh with him. He eventually sent a Form 4 Orkeeswa student named Saing’orei to our house (which was next door) with 50,000 Tzsh to keep us going until we could go to Arusha to exchange money. Saing’orei took us into town, helped us bargain for a new cell phone (bringing down the price from 45,000 to 38,000 Tzsh), helped us figure out how to get a sim card and credit, helped us find a good Internet café, and walked us home as it was very dark when we finished. In a word, he was awesome.

The sparseness of the trees and absence of buildings makes the night sky look very big.

After dinner, J and I hung out with the women cleaning up in the kitchen; they very good-naturedly taught us more Kiswahili. Apparently I am too quiet; this is partly because I don’t know the language very well (not even basic grammatical structures), but I’m also not incredibly talkative in English either, especially with garrulous people around already filling the airspace. Which is funny because this blog is very verbose.

Tomorrow we are going to Arusha to exchange money and hopefully buy a 3G router and some USB-modem Internet access, so I can post these late blogs. Yay! Also going to buy electrical outlet converters, and toilet paper… Yeah… Kesho! (Tomorrow)

Monday, July 1:

Reality check

On Sunday, many shops and businesses in Arusha are closed; a very large portion of the population is Christian, so they go to church instead. We’d heard this from an IEFT volunteer, but had heard information to the contrary from an Orkeeswa student, so we decided to go anyway. We were disappointed to find that KrakiT Computers, the store we knew sold a 3G router, was closed. But we had a very lovely lunch and tea with Seth, who gave us a lot of useful context about the school, and showed us a chai place in a hotel with complimentary wireless Internet access!

But more on disappointments, just because I’m feeling that way right now.

We visited Orkeeswa for the first time today, and spent almost the whole day cleaning disgusting, moldy, foul-smelling rotten old chemicals out of beakers and soaking things in soap solution to try and make them acceptably usable for the science lab. There is now a bucket labeled HAZARDOUS WASTE at the school which when opened emits a strong odor of rotten dairy products. Apparently neither science teachers nor students consider it their job to clean up after doing experiments, because there’s no one assigned to the duty and no policy in place to hold someone responsible for equipment care.

There’s a larger property management problem at Orkeeswa, which we’ve been recruited to help devise solutions for. Since the school appears rich compared to the surrounding Maasai, there’s a small percentage of students and community members who walk off with school property; they borrow supplies like shovels, shoes, and books without asking, and those things disappear from the school forever. We’re trying to come up with a policy and mechanism for keeping supplies secure without having to hire an additional person to keep track of everything for Orkeeswa, which the school can’t afford. There are many ideas floating around, but we’re all hard pressed for the time we need to sit down, discuss, and agree on a policy.

Moreover, while looking around the school for a good spot to mount an antenna, I realized that there is a small hill next to the hill that Orkeeswa is on along the line of sight to Monduli that is just barely tall enough to block it. CURSES!!! I am extremely unhappy about this. It doesn’t look like a relay from Dr. Danny’s will help because the area around Orkeeswa is quite hilly, and the elevation profile shows that two hills almost or just barely block the line of sight between Dr. Danny’s and Orkeeswa (although we’ll definitely check in real life to make sure). A third hop would be too costly in terms of throughput. Even two hops is doubtful with the slow Internet speed in Monduli, and it would be hard to maintain and secure a repeater station on that blocking hill (even with its proximity to the school) where equipment would be out in the open for people to see and steal.

There is a tiny ray of hope that if we can get Internet access on that small hill (which is like a 5 minute walk from the school), IEFT will contribute a little funding towards making a small enclosing for it on the hill, or to moving this shipping container currently being used as a storage room to the hill. Then we’d either get it wired up with solar panels, or use this huge lead-acid battery pack that only works for 7 hrs at a time. Or we could even have the students set it up themselves every time they wanted to use it- it’s not that hard. Or cry! I’ll have to think of something.

Endgame: June 28, 2012

I shouldn’t have waited until now to write this post. I knew everything would feel different once I was on the plane.

The last few days have been very stressful and somewhat busy, but mostly because I have been sleep-deprived. The stress caused me to get a whole bunch of canker sores in my mouth, which cause a distracting amount of stabbing jaw pain (like a tiny person with a pickaxe is mining at my gums) as well as throbbing headaches. Sleep is incredibly important.

Monday:

The plan for the day was to get back at around noon, run a bunch of errands, test the system with two hops and with the MacBook as a router, and then go home to J. I did get back at around noon and tried to meet with Sindy Tan (from my high school, who is at RSI) to give her my bike for the summer. Unfortunately, I realized that my bike was still at the Tang from when I rode it to testing on Friday morning and then had to carry back the antenna back a kilometer and a half (whine whine whine). So I decided I’d be back on Wednesday to give Sindy the bike- on my honor, at all costs. I have issues about that sort of thing.

Then I got haircut, picked up some medicine for the trip (and my canker sores) from CVS, and set up the two hops using all four Bullets down the hallway outside the office in bldg 36. (Almost didn’t have enough Ethernet cables – had to scrounge from the other lab desks, again.) It worked fine, but was extremely slow: 3.29 mbps download and 0.73 upload, with original speed 58.96 download and 2.72 upload. I guess the halving of the throughput when you increase to 2 hops really makes a difference. One thing I didn’t test was the effect of the length of the Ethernet cable at the relay point on the connection speed. Now that I think about it, that could make a big difference; the data has to travel through every cable at the midpoint, so from BulletàinjectoràinjectoràBullet. I wonder how efficient POE is at transmitting data—the fact that we are also modulating the data over a DC power signal (I think) ideally shouldn’t make any difference. But any fluctuations in the DC power signal could also cause bit errors in the modulated signal.

Power over Ethernet also forces you to use Ethernet cables, which is easy but not very signal-efficient. I wish I knew more about radio equipment so I could engineer a solution.

More on what happened on Monday.

The last errand I’d scheduled for that day was to pick up the screw-on “butt” end of the first broadcasting Bullet from Tang, where I’d accidentally left it after testing on Friday, and to pick up my bike. It was an egregious error, by the way. Without the butt of the Bullet, it might not have been waterproof; we would have somehow created a solution, but it would probably have been makeshift at best, and would have wasted precious time and energy. Always keep track of every part of your equipment, even caps and screws and everything—you don’t want to be caught in the field without them.

It was drizzling when I stepped out of the office, and I was carrying nothing but my cell phone. I started running, although I found out later that I might have gotten wetter because of it. The rain started coming harder and harder; lightning cracked the sky and thunder rolled overhead. I counted the seconds – four miles away (to within an order of magnitude, anyway). I was pretty soaked and out of breath by the time I got to Tang, and the storm was getting closer by the minute. But after tucking phone and Bullet butt into a doubled plastic bag, I headed back out to bring the bike back for Sindy. Biking 1.5 km in a short skirt through stormy, flooding streets was quite the adrenaline rush. I squelched back to the Terrascope room dripping from head to toe. Luckily, I had a towel and change of clothes. I stayed with P that night.

Tuesday:

I was exhausted; I tested the system down the hallway (one hop only) using my MacBook as a router, and it worked fine! No idea what had happened that other time down Vassar St. Then I was supposed to go to Radio Shack to get supplies to repair Ethernet and coax cables with – crimp tools of the correct sizes, heat shrink, spare end connectors (N type, 8p8c). Instead, I crashed in the Terrascope room for 5 hours and that effectively ended the day. When I got up, I panickingly searched the web for the items I needed and ordered some of them with two-day shipping (because for some reason my confused brain thought it was still Monday). Instead of sleeping that night, I emailed out to the MIT radio society (w1mx) and MITERS mailing lists begging for N type end connectors, and watched as long as I could. It would have cost me $180 out of pocket to order the connectors and a crimp tool for them online, and they were only spare parts (for pretty important things nonetheless).

Wednesday:

J and I disassembled the antennas and packed everything away into 4 suitcases – 2 just barely less than oversize. We’d been worried about finding a suitcase for the huge 24 dBi antennas’ parabolic dishes, which were 2 by 3 ft and each came apart into 2 by 1.5 ft pieces, but our widest suitcase just barely fit them diagonally.

The MIT radio society also donated some spare N-type solder connectors (no crimp tool necessary)! Thank you, jhawk and w1mx!

All was well (except for maybe my canker sores). Both J and P brought me strawberry rhubarb pie to celebrate.

Thursday afternoon:

Traveling has been surprisingly light-hearted. We ran last-minute errands like binding the science lab manuals, and J made a trip to Radio Shack to get extra electrical tape and some 8p8c Ethernet cable heads that hadn’t arrived on time. I had pie and lunch with P right before leaving, which was great. J is also a very sunny person; I am lucky to be traveling with her.

Now coming into Amsterdam for a 4-hour layover (hopefully enough time to type this up, and maybe buy a real notebook?).

Last Minute Tests Before Departure: June 24, 2012

his blog post is two days late because things like shopping and catching up on sleep take a stupidly long time. (As you can see, I gave up on one of these tonight.)

The last long-range tests went well.

The Tang, where resides the amazingly wonderful Tetazoan cruft Chris Calabrese who let us use his apartment many times when it wasn’t actually convenient for him, is just under 1.5 kilometers away from the New Media Lab. Cbreezy’s 12th floor apartment is actually just a bit below the treeline in front of the E15 roof deck where we had our receiving antenna set up; we didn’t quite have line of sight there. But the signal strength was -76 dBm (to out of four power bars- not bad). The Internet speed test on the receiving end didn’t work because it was too slow, but google search was decently fast and the HTML version of gmail loaded fine. At the Tang, the speed was around 7-8 Mbps. I think at this point that one of the toughest things about implementing in Tanzania will be that the Internet speed at the source will already be quite slow- less than 4 Mbps for sure. Not sure whether we’ll be able to do anything about that.

The second test was almost as long a distance: 1.12 km from the 14th floor of MacGregor to the Media Lab roof deck. This time, we had perfectly clear line of sight, and a high elevation off the ground. The signal strength was -47 dBm (without the 100 ft Ethernet cable in the system), and the speed on the Media Lab end was around 8-10 Mbps (the wireless speed in the same room as the Netgear router). The speed at MacGregor was around 60 Mbps, so there was a loss of about a factor of six. Assuming the loss in speed is directly proportional to the loss in received power (which might not be true), I think that’s about 8 dB of loss (10*log(6)). Dunno what that tells me, though.

In the second test, we used our 100 ft Ethernet cable and discovered that the the signal strength seemed to fluctuate between -47 and -61 dBm; we aren’t sure why that would be happening, but the speed test didn’t reveal any speed difference with the cable so it did not seem to matter.
J (at the receiving end) also moved the computer around to different places to test the range of the Netgear router. There was no difference in speed 25 feet away from the router with only space or a glass wall in between. With a wall between, there was almost a tenfold decrease in speed.

One really troubling/puzzling thing happened: even with the Bullets configured correctly using static IP addresses, when J plugged the Bullet directly into her computer, there was a DNS lookup error. I’m not sure why that would happen, especially because it never happened before when we were testing the same setup at a shorter distance (across East Campus courtyard). This is another question for Ubiquiti tech support.

The promised update having to do with the GPS coordinates of the staff houses and school:

We got them. And then we used Google Earth’s Path tool to find the elevation profile along the line of sight paths between the school and the various staff houses and offices that could be our Internet source. The results were amazing.

Kudos again to Google Earth, and satellite imaging.
We are hoping that the little hill on the Orkeeswa side of the path (on the right side of the elevation profile, since the path started in Monduli) is right underneath the school rather than behind it. Apparently this isn’t just wishful thinking; the coordinates we were given for the staff house was exact, but the coordinates given for the school were actually the coordinates of a girls’ school close to Orkeeswa, a little bit further from Monduli. Our IEFT contact tells us that Orkeeswa is likely on that little hill, which would be really awesome, to say the least.

The profiles between the other possible Internet sources and the school were also decent, but not as good.

The last thing left to test is whether or not the relay connection will work. The two Bullets at the midpoint (if we use one, which may happily not be the case) are connected via an Ethernet cable between the two LAN ports in both of their POE injectors. We’ll do a short-range test on Monday.

I promised the schematic sketches I did. Here is a very drafty and messy second draft that should really be copied over at least once more:

Miscellany:

Things bought/obtained yesterday:
3 pairs of thin long pants
1 floppy-brimmed hat
1 mosquito net for sleeping in
1 huge, garishly pink hard suitcase

Things to buy today:
Lots of bug spray
Granola bars
Travel-size stove/water boiler?
Reading Flashlight (because I dumbly sent all of mine over already)
Blooming teas (gift for host family)

Things to buy at MIT:
Another MIT t-shirt (lost mine)
Screw-cap water bottle

Things to do today:
Go to Apple store, get computer fixed(?)
Email Ubiquiti

Things to do in the future:
Send a postcard to Bello Opticians (348 Shrewsbury Street, Worcester, MA 01604)
Email that random research scientist who asked us about our project in the elevator
Write a post about science lab testing at Hunter

To see entries previous to this one, check out her external blog here

« All Posts