Most people mod their drone with powerful antennas to increase their range. I wanted to add a different type of antenna to mine for a completely different reason.
I recently ventured into providing Wireless Internet access to myself and all my neighbors. I live a good distance out in the country, and my only real options for internet were completely useless Satellite based, or a local WISP what was charging me $160 per month for 3 Mbps. I did the later for seven years before deciding that I needed to find something better. Since I work from home in a technology based job, I needed high speed internet with low latency. So I secured access to use a tower in town next to a building (thanks M6 Oilfield!!) where I could purchase a dedicated line. I now had 75 Mbps of internet that I could do with as I pleased, but I had to get it to my house first.
My first thoughts were to set up an easy Point-to-Point, but after talking with a few of my other neighbors who were in the same predicament as I was, I decided to go with a sector. This way I could provide access to all of them also. I chose a Mikrotik NetMetal 5, and a RF Elements 120 Sector, climbed 150 feet up the tower, and got it all rocking and rolling. On the house side I added a Mikrotik DynaDish 5. Before the DynaDish came out, I was rocking the Mikrotik QRT5, but the upgrade to an AC link was worth it (for other reasons). Now, 2 years later both my equipment and the number of neighbors connected has expanded. Most of them are farmers who just check email / surf the web a little, so the link is still way under utilized. There are always more that seem to want access though. I don’t actually charge my neighbors for access to the internet, instead I have them buy their own equipment and then we enact a barter system. When you live in the country, you are always needing to borrow tools / equipment for various things. So we instead help each other out in that respect. They get free internet, and I get to borrow a rototiller when I need it. 🙂
Now for the Drone. I recently purchased a Yuneec Typhoon Q500 4K. It has been an awesome drone, and I highly recommend it to anyone looking to get one. I have used it for various things, such as finding a break in our water main where it crosses the back creek, but I wondered if I could use it for the WISP I was slowly building out.
The furthest neighbor is 13 miles from the tower, while I myself am only 6 miles out. With different elevations, direction to the tower, and tree lines; finding a solid connection to the tower required lots of trial and error. You generally put up a temporary 20′ pole and pointed in the general direction and turned it until you found it. If that connection wasn’t good enough, you moved around until you found one that was. This may involve moving the temp pole, which was a real hassle. It may be that you needed to go higher than 20 feet also, which made testing even more problematic. All this time you also were dragging a laptop around to check the signal quality, and you needed a long Ethernet cable with power to use a POE injector to power the CPE. So with this in mind, I decided to mod my drone to act as a remote station.
One of the awesome new products that Mikrotik just p ut out was the Mikrotik SXT Lite5 AC. What makes this awesome is it technically has 2 wifi modules in it, both a 5 Ghz, and a 2.4 Ghz. The 5 Ghz makes the connection back to the tower, while the 2.4 Ghz acts as an Access Point, allowing a single connection back to manage the device. This means that I can connect to it wirelessly, without needing a long Ethernet cable / laptop (since I can use my phone). So I grabbed one of these from my local supplier, ISP Supplies. One problem solved. The next problem was power. I had thought about using a battery pack POE injector like the iDea4Tec Smart PowerBank but not only are they expensive for what they do, but at 1/2 a pound, it was going to be pretty heavy for my drone. Weight is at a premium, so the next idea was to use the drone’s own battery to power the device.
I was already going to be removing the camera from the drone. I didn’t really need to see where I was pointing, and since the camera resides on the front end of the drone, it would be directly blocking the CPE signal. It also saved me some weight. So removing the camera left me with the power connector freed up. Since I still planned on using the camera with the drone later, I didn’t want to butcher the existing power cable, so I snagged another one from Carolina Dronz (which also happens to be where I bought the drone to begin with). I cut off the end that would normally attach to the camera, and proceeded to figure out what voltage we were supplying. There are 3 wires, so I found the one that supplied consistent power and the ground. The other supplied power but it was at a lower voltage than I was wanting. The power line supplied ~11.7v (varies with the battery voltage) while the CPE will accept anywhere from 8v to 30v, so we were in business. All I needed now as to cut a Cat5 cable, and splice the proper Cat5 wires to the power and ground (Pins 4/5 to power, 7/8 to ground), and I now had a Drone to POE Cable.
Since the drone itself moves a little in the wind, you want the CPE to be able to keep its orientation too. Later this will be handled by mounting it to one of the GoPro 3-Axis Gimbals (I wasn’t going to mess with my current Gimbal). For now I have created a temporary mounting system. This involves making it free hanging beneath the drone by securing it with a rubber cables on each side, and some electrical tape to temporary hold it in place. I used the tape so I could determine the exact balance point of the CPE to make it always point horizontal while hanging.
So I did a quick pre-configuration of the CPE so it would connect to my tower, and away we went for our first flight. I powered on the drone, before the drone had finished its initialization, the Mikrotik was already booted up and I had connected to the wireless via my phone. I flew up ~25 feet and pointed in the general direction of the tower and was awarded with seeing the connection made. The drone holds its position extremely well (even in some wind). It also has a Turtle mode, so I was able to pan back and forth super slow until I was able to dial the connection in to find the perfect center. The Drone Controller itself then provides me with the exact elevation and azimuth of the drone, so now I also know exactly what size pole and the direction it needs to face. This means I could now find the best possible connection in under 5 minutes, where as normal it might take me 30 minutes or longer.
From here I plan on buying a real Gimbal for it, and doing some more real world tests from further out. I have a few clients whose connection isn’t as good as I would like, so it will be a good chance to see how I can improve their access.
I promise to post a lot more pictures with detailed instructions once its not so “hackish”. I would like to note that through out this entire endeavor, I had the support of my good friend Greg Sowell. He has been educating me all things wireless ever since.