October 27, 2007
I wasted no time in assembling one of the 30 dBi units and mounting it on the mast I had set up. It took a while, but everything went smoothly. It was bigger than I had imagined, but not as heavy. The whole thing can be put together with just one person, although two would have been easier. It can only be mounted about half-way up the mast because of the guy wires.
I was curious about polarization. On this antenna, the center feed horn can be rotated 90 degrees easily, but I don't see any way to rotate the dish itself without mounting it on a horizontal pipe (and then I would worry about its stability). The antenna's instructions say, "Feed horn must be orientated to grid as shown [feed horn parallel to grid bars] for optimal antenna performance." How much loss are we talking about here? For long-distance links, I suspect that horizontal polarization would reduce interfere from other WiFi signals (which are usually vertically polarized), but it appears that this antenna is only designed for vertical polarization.Node 14 site survey
October 19, 2007
We packed the usual site survey equipment and took a drive up to the area proposed for node 14, using Google Earth directions as a rough guide. It was cold, foggy, and a little difficult to find, but when we did arrive, we discovered two gates, both wide open, and were able to drive up a little beyond. The road did not permit driving all the way up to the towers; for the next trip, we need a truck, preferably four-wheel-drive in case of snow.
Building a prototype mast for village wireless VoIP
We walked around, took pictures of the towers, signs, and the area which looked good based on Radio Mobile and Google Earth. West of the road there is a large area which had been clear-cut for timber and seemed ideal. The area east of the road is thickly wooded and has no hope of line of sight without a professional tower. Rain made walking around unpleasant and I was concerned about getting the camera wet. We also could not actually see the site for node 13 because of fog, but we were able to see at least beyond the mountain in the right direction (according to the compass), so if Radio Mobile is any indication, the line of sight should be good. Cell phone coverage was also good.
October 18, 2007
I spent a better part of the day cutting, drilling, grinding, and assembling. The result was surprisingly stable–more stable than I could have imagined anything which just sat on the ground.
However, my concerns about the pipe being able to rotate were confirmed. The design needed to include something which would prevent the entire mast from rotating. Even half of a degree rotation is significant when dealing with high-gain antennas like we are using.
The cable tensioners (aka turnbuckles) can be confusing. I decided to always put the normal thread toward the ground and the reverse thread up. With this arrangement, because they are close to the ground, the instinctive clockwise turn will tighten the cable.
A shopping trip for village wireless VoIP
October 16, 2007
We took a trip down to Home Depot, having only half of a design in mind for an antenna tower. We considered a couple of tripod-like scenarios and an H-shaped base, but in the end decided on a 2.4-meter steel pipe and three steel cables running from the top of the pipe down to super-heavy-duty tent stakes (angle iron). Not being a plumber, I was surprised to learn that steel pipe is sized by the inner diameter, but of course the antenna descriptions specify the minimum and maximum outer diameter for mounting.At last, a test site for villiage wireless VoIP
October 12, 2007
The day after that we met the land owner at his office first thing in he morning. His primary concern was that our testing would not interfere with the Department of Transportation microwave relay station which was at the top. He talked with us about the project, made a copy of my driver's license, had an extra copy of the gate key made, and gave us a letter stating that we had permission to be there (in case we were stopped by the people who patrol the area).
Then we drove back toward node 13, through the locked gate, through several other unlocked gates, and up a steep climb. The fog became so dense at one point that we had to drive near walking speed, but at the top we actually came out above the fog and were greeted by a beautiful view for hundreds of kilometers. We surveyed the area as described earlier; everything seemed ideal.
When aligning high-gain parabolic antennas over a long distance, it is critical to have good communications between the two endpoints. None of the towers at the node 13 site were cell phone towers, so I was also very glad to discover that there was, in fact, cell phone coverage at the top, even though there wasn't at the base.
There were a couple of possibilities for mounting the antennas. We could attach them to a tree on the north slope, an old unused telephone pole at the top, or a free-standing system of our own. We decided for the free-standing system because we didn't know how the land owner would feel about the other two. In addition, the antennas required a specific size pole on which to be mounted, and devising a way to attach the pole to either the telephone pole or tree seemed difficult.Another disappointment, but hope
October 11, 2007
The next day we headed off towards node 13. We missed a turn or two, and then came upon a metal gate across the road labeled, “Private Road”. There was a laminated letter attached with the land owner's name and signature at the bottom. Armed with this information and a phone book, we drove back to the land of cell phone coverage and called him. He was surprisingly helpful; he could have just said 'no'. Instead he asked about why we wanted to be on his land and asked us to come by his office with our credentials.Disappointing change of plans in wireless VoIP research
October 10, 2007
It was time for our first site survey--a 'reality check' for a proposed site. Radio Mobile's altitude data is around 90 meters in resolution, so trees, buildings, and even small hills don't show up. Google Earth can help some, especially if people have posted geotagged photos, but a site survey is still a requirement.
My goals for the site survey were to
Below is a list I've compiled of what I would like to have on future survey trips. Needless to say, I was not able to bring all of these items on our first trip. I hadn't even purchased all of them yet.
We headed off towards node 12. Google Earth directions are not so accurate on tiny roads, as it turns out. I guess this shouldn't be a surprise. We ended up on many private driveways, and were even told at one point that we had probably "activated a neighborhood watch", whatever that means. We drove around in a few circles, and finally gave up on that route. As far as we could tell, the hill in question was private property. We had what we think was the owner's name and may have been able to get permission to do our testing there, but the bigger problem was that there was no usable road from where we were to the top of the hill, despite what Google Earth said. We could have walked there with a compass, GPS, and camera, but not with large antennas, batteries, etc.
I was surprised to learn that Google Maps directions are sometimes entirely different than Google Earth. Google Maps seemed to avoid the very little roads all together. Having given up on approaching the hill from the south, we followed the Google Maps directions. These led us all the way round the base of the hill and we ended up east of where we wanted to be, and significantly lower. We could look up and see our goal, but there was no way we could reasonably carry all of the equipment up that slope.
|The TR Podcast 175: the Zen of chipmaking and ARM's Cortex-A72 revealed||4|
|Elon Musk lays out vision for a battery-powered future||122|
|Inside ARM's Cortex-A72 microarchitecture||37|
|Asus' 144Hz MG279Q monitor may top out at 90Hz with FreeSync||59|
|Deal of the week: A Bay Trail netbook for $161, free case fans, and more||18|
|DirectX 12 Multiadapter shares work between discrete, integrated GPUs||98|
|Gigabyte's 9-series motherboards are Broadwell-ready||46|
|The TR Podcast will be live on Twitch shortly!||3|
|AMD delays FreeSync support for multi-GPU systems||41|