Very few stations have the sales expertise to sell off air time to a prospective advertiser. Although I once worked at a station that sold the 2 seconds of dead air during an antenna switch to an advertiser most stations prefer to be on the air, always available to it’s audience. Time off the air is lost revenue. Period.

At some point either the station engineer or station management will have to decide what is an acceptable amount of time for the station to be off the air during the year. Failure rates are generally referred to in terms of percent of yearly failure periods. For instance, if management feels that a station that is on the air 99.9% (three nines) of the year is acceptable that would be equivalent to 8.77 hours off the air. I wonder how an advertiser will handle it if some of that takes place during an important broadcast or time that they are paying for.

How about staying on the air for 99.999% (five nines) of the year? That is the same as 5.26 minutes off air during the year. My guess that would be more acceptable but is it possible to be on the air 100.00% or the year? The answer is directly related to how important being on the air is to the station management and ownership. The degree that a station maintains it’s on air presence is directly related to two factors. The first is luck and the second is the amount and type of redundancy that the station has built into it’s facility.

A complete signal path can be viewed as a chain with each part of the path being a "link" in this chain. For example, one part of the chain might be the studio processing or the STL, or the antenna. Each is important in keeping the station on the air. More than one "chain" may be necessary to keep the station within an acceptable number of "9’s". The more attention that is given to the possibility of path failure due to any unusual situation the better the station has of staying on the air. Its like trying to predict the unpredictable.

Signal path redundancy basically means that there are several different paths for the studio signal (audio, data, video) to reach the transmitter site. Redundancy in a radio or television station takes on different forms such as parallel, interlaced, isolated path, completely isolated, and mixes of the above. Examples of these forms are shown in the figures.

A parallel form of redundancy can be simply a microwave path with a phone line back up. Some sort of decision making process will determine when to have the primary microwave path (the STL) on the air and when to have the phone line on. The decision making process can be a manual command from the station personnel or can be a sophisticated automated process. In either case failure thresholds need to be determined to at what point is it necessary to change the primary signal path to an alternate path. These failure thresholds can be simple, as in the total loss of signal on the primary path, to a complex determination of signal degradation in the primary path relative to acceptable levels of signal degradation. Even more complex is the establishment of a pattern of a predetermined failure thresholds that have been exceeded over a period of time, and then taking action based on projected failure conditions. The decision making process is further complicated by increasing the levels of redundancy. With an interlaced form of signal path redundancy the signal that finally arrives at the transmitter may be a subset of other signals traveling in the same direction (i.e. a T-1 encoded signal path over a common telephone line).

An isolated redundancy signal path, as the name indicates, has isolated signal paths running from the studio or studios, to the transmitter site. An example would be two studio sites feeding a single transmitter site over two different paths. A completely isolated system has two sources feeding two different paths to one or two transmitters or even multiple transmitter sites. Obviously the latter situation would require a very sophisticated system of communications between sites to coordinate which was on the air and when. At this point the control communication path lends itself to a redundancy scheme as well!

In any talk of redundancy the engineer must look at each link in the chain and what happens when that link disappears. With STLs it can be a high wind that knocks the antenna around on the tower (even icing or damage to the transmission line). With a parallel redundancy scheme (interlaced or not) you are open to catastrophe if one item in the chain fails. Usually a mix of STL and T1 will solve the "phone company has a problem" syndrome.

The decision to minimize redundancy in the signal path is usually a purely financial one. How often have you heard the following: "Why spend the money on two STLs or a phone line back-up when that old STL is working just fine?" or "We haven’t been off the air in years!" One of many lessons that the World Trade Center disaster taught most of us is that no matter how secure we feel with our systems the unexpected CAN and DOES happen. Those stations that had completely isolated redundancy with their studio to transmitter links as well as different transmitter sites were the ones that were still on the air. The others were off the air for several weeks and were scrambling to find alternate facilities and equipment. Hopefully those stations had plenty of insurance! But, then again, isn’t that a form of isolated redundancy too?

How you figure out the degree of your vulnerability to a problem that causes a path or signal failure is actually pretty easy. On the left side of a piece of paper draw your studio and on the right draw your transmitting antenna. Between the two draw every piece of equipment that forms the "chain" between them. Now draw a line between the equipment to show the path that the signal takes on the way from the studio to the transmitter. Now simply cut one of the signal lines or remove one piece of equipment. How does your station keep it’s signal on the air? Now remove two signal lines or pieces of equipment. How is the station now? For a fun exercise draw in the power mains to each piece of equipment and then the breakers that the mains are on and then add the main power distribution panel or even the power company. It’s easy to see how complex designing a high reliability redundant system can be! The variables can become astronomical and eliminating each variable is now associated with a cost that must be weighed against the benefits. Again, how well can your sales staff sell off air time?

There are as many approaches to designing a reasonably priced redundant system as there are station management teams and stations. Some systems are very cleaver and well thought out. Others lend themselves to "Catch-22" scenarios. For example two STL transmitters being fed from a common AC power strip that the cleaning woman uses for her vacuum because there is no other outlet for her to use to clean the STL rack.

One of the more well thought out systems was recently installed by Kevin Plumb, Director of Engineering for ABC Radio, NYC for their flagship ESPN radio affiliate, WEVD. In his planning Kevin took into consideration just about every possibility of equipment or path failure and found ways to duplicate the path working around possible problems with various links in the chain. A simple block diagram of WEVD is shown in figure 5 (signal path). If WEVD had a redundant transmitter site then it would be one tuff station to take off the air! WABC’s sister station WPLJ-FM, for example, has 2 non-collocated transmitter sites; one atop the Empire State Building and one in Alpine, NJ.

As with most things in life calculating how much to spend on a well thought out redundancy plan is the same as calculating gambling odds. Calculating the odds that a phone line will be down during the next year or that the cleaning woman’s vacuum cleaner is defective is easy. Calculating the odds that your transmitter site will be affected by a 9/11 type of disaster is far harder.

So how much is redundancy at your station worth? How about adding a column to your next rate card. One for "Hours Off Air".