The growth of CenturyLink was partially funded on government ineptitude. I applaud CenturyLink for figuring out the best way to play by the rules (if it really was legal), no matter how much the taxpayer got screwed and how unethical it appears to be. If it was legal, they won and heads should roll. Can you imagine being a fly on the wall when CenturyLink announced they were buying Qwest after growing like crazy partially on the $1.7 Billion dollars they weasled out of the FCC under questionable pretenses. Of course, using the local companies CenturyLink hid behind in rural areas must have made it hard for the RUS to figure it out since nobody there must be capable of looking at a corporate filing. Somebody at that division should either be fired for incompetence or outright stupidity but probably according to government union employee rules, neither is a valid reason for termination. Of course, if they couldn’t figure out Bernie Madoff or that drug cartels moved $350 billion dollars through CityBank over several years (for which nobody has been prosecuted nor will be), why should we expect the government to watch out for a measly $1.7 billion dollars of taxpayers’ money? As Comcast and Cox adroitly pointed out in their letter, private industry isn’t going to invest or compete in areas where the federal government is propping up their competitors. Of course, Hughes and WildBlue also come to mind against the WISP community but at least they are training the next generation of inexperienced users that have more than one option of what not to buy. But then again, Ford could use the same argument when Obama paid off his union cronies by giving them billions of dollars and the majority of the stock in GM. Jimmy Hoffa would be proud since he never owned a President as completely as the United Auto Workers do now with Obama.
But let’s not dwell on the continuing story of government incompetence and corruption and move back into the realm of real wireless projects. After all, that’s what these articles started out as. I’m in this industry because almost every project has something unique and challenging. My latest product is a simply Point-To-Point link (PTP). I’m currently working on a project between 2 buildings that needs a leased tower location as the relay point. The tower is located about 10 miles from the taller building and less than a mile from the shorter building. Bandwidth needs to be at least 40Mbps or more and it’s fairly rural so 5.8GHz unlicensed isn’t going to be an issue in terms of interference. The only challenge is that the longer range link is going to require that the antennas are at least 250’ in the air.
There is nothing special about this project and many of you have done similar installations. Normally you calculate the link path, which in this case would be work great with a pair of 2’ 30dBm parabolic antennas. However, the crimp in this equation is that we get charged on the tower based on antenna size. The base charge is roughly $300 for but each additional foot of diameter of antenna costs roughly an additional $100. If the base is a 1’ antenna, then a 2’ parabolic dish antenna costs $400 per month, a 3’ parabolic dish antenna $500 and so forth.
I’m sure you can see where we are going with this. It’s our building on the other side of the 10 mile link to the tower so it’s possible to put up the Seti array if we wanted to on the building as long as the link path antenna gain equals or exceeds 60dBi. I calculated that a pair of 2’ parabolic dish is sufficient for the link but if we use a 1’ flat panel on the tower and a 4’ dish on the building, we can save $100 per month. Since we are using redundant links, meaning 2 antennas, that actual savings is $200 per month. The calculated antenna gain with the 2’ dish was 60dBi for both antennas. It so happens that Ubiquiti manufacturers a 25dBi flat panel radio/antenna that is 14” across with a 25dBi gain. On the other side we can use a 34dBi gain antenna which comes close enough to our link budget for government work.
So, what is the drawback to this idea? Well, considering we are using a lower gain, less directional flat panel antenna 250’ in the air that means the noise figure on the tower is going to be higher than it would be for the 2’ parabolic. That also means that the S/N ratio is going become more of a factor and has to be considered. Assuming you don’t think the noise figure is going up significantly, then there are no issues. On the other hand, the tower climber will hug you since he only has to deal with 5lbs of equipment instead of 60lbs.
I mentioned the redundancy to the link so let’s discuss that. Since the radio costs are so cheap compared to downtime, we can afford to put 2 of the antennas on the tower. Yes, that adds another $100 per month but compared to the cost of downtime and tower climbing costs, it’s relatively cheap insurance. That also means that we have to mount two 4’ antennas on a building for the other side of the link. Running the numbers again shows that we could cut down the gain on the antennas and use a second flat panel on the building side and still get close to the same performance or a slightly. Since the secondary link is only there for backup purposes and the radio on the primary link can be replaced in 30 minutes or less, then if the secondary link has a slightly slow modulation rate such as MCS(12) instead of MCS(15) rates, then it’s not a big problem. That also means that the backup radio can be another flat panel instead of the 4’ dish since it only has to operate a short time while the radio is replaced on the back of the dish. That reduces the cost and the size of the roof mount on the building size also.
On the tower side, we have now limited the wind load to two 14” flat panels for redundancy. Redundancy can be handled many ways. The easiest is to power up both radios and use something like OSFP, STP, or RSTP to maintain redundant links. Another alternative it is to use something like an IP based power switch from Digital Loggers http://www.digital-loggers.com/lpc.html which can be programmed to power off one radio and power on another radio if there is a failure. I have used both methods in different projects. In this case, because towers are more susceptible to lightning hits, I’m more inclined to go the power switch route. The power switch basically turns off the power to the radio, thus disconnecting it from the power lines in case of an indirect hit. A direct hit to the tower probably fries both radios, regardless of whether it’s powered up or not. I can attest from personal experience that unpowered radios will get fried also when that happens. Although 8 unpowered radios were fried, 1 unpowered radio survived so you take your chances. A good strategy and this was the only thing I can attest this to, was that the unpowered radio was also several feet below all the other radios. I’m covering both bases with this install. This will also be the strategy on the tower with the two flat panel radios.
I mentioned early on that the tower was a relay point for the link which really means 4 radios. If we were using 2 more radios for the link, then we are adding to the wind load and thus the costs of the monthly rental fee. However, what if we can reduce the size of the radios so that their footprint doesn’t add to the cost? Since the second link is only ¾ of a mile, then we can use a very small footprint, Ubiquiti NS5M, which literally at 3” wide, is thinner than tower legs themselves. That means that it adds no cross section and even if there is a small charge, it probably will be very minimal. Again, we are going to use 2 radios for redundancy and these will get mounted towards the bottom where they pretty much become invisible against the legs.
PTP links don’t need to be symmetrical. Antenna gain takes into account both sides, not just one. If you are paying for tower footprint by the square footage, it makes sense to reduce the antenna size as much as possible. Obviously there may be some additional costs on the other side but usually Esmits minor compared to tower rental costs.