For the last few columns, I’ve focused on fiber optic cables, their types, installation techniques, etc. With that background, I want to cover some additional details you need to know to design and install a fiber optic cable plant.

You can’t have too many fibers

Network designers sometimes look at how many fibers they need and add a few spares. Don’t do that! Every major future system you hear discussed—5G and small cell cellular, fiber to the home, intelligent highway systems, smart grid, data centers—all are “fiber hogs.” Too much fiber is just right.

Fiber is really cheap; at about a penny a foot per fiber, it is cheaper than kite string and fishing line, and cables are only a few cents per fiber per foot depending on the cable type. Construction costs are way higher—about five to 50 times, depending on the installation type (underground in conduit, direct burial, aerial or premises). The only additional costs to installing a 144-fiber cable instead of a 48-fiber cable is the cost of splicing or termination, and if you don’t need the fibers now, you can leave them unterminated for future connections.

The cost of the additional fibers is cheap insurance against future construction costs. And if you are digging trenches to install fiber or conduit, put in a lot of conduits, and leave them empty for future needs. Remember the “Dig Once” policy.

We often get questions about what cable hardware should be used. That’s tough, because there are hundreds of types of cable and thousands of types of hardware to choose from, and they must be compatible.

Our first question is always “Which cable are you using?” The cable type, and often the manufacturer, determines the types of hardware needed. For aerial cables, particularly all-dielectric self-supporting (ADSS) cable, the hardware should be what the cable manufacturer recommends, because the two components are designed for each other. Long-term reliability depends on the installation being done to manufacturers’ standards.

This also applies to splice closures. Most closures and splice trays are designed for regular outside plant loose-tube cable, right down to each tray having space for 12 splices, the number of fibers in a buffer tube. But if you are working with traditional ribbon cables, the cables can be of several types of construction, such as central tube or slotted-core that require special splice closures and trays. Working with ribbons can be difficult since they only bend in one direction, making fitting the splices into the trays a challenge. Newer designs where the fibers are only loosely held in ribbons don’t have these problems.

My advice on choosing hardware is to always consult with the technical support personnel at the company supplying the cable. They know best.

Make the cable long enough

While visiting an installation of fiber-to-the-home in a suburban neighborhood, I watched a worker installing a large fiber count backbone cable on an aerial messenger. The supervisor showing me around stopped and ran over to the tech to have a heated discussion about the installation. He explained that the workers had been installing the cable and it to the ground at the location of an underground vault where it would be spliced to a number of distribution cables. The cables were barely long enough to reach the ground, but there needed to be adequate cables for splicing and to leave service loops in the underground vault for future expansion needs. The supervisor had told them several times that his splicers needed another 30–40 feet to bring the cables into the splicing truck. There they had to cut the cables back 6–8 feet for splicing.

It’s always easier to store excess cable in service loops than “stretch” a cable. In another project, the designer told me that he had tracked actual cable usage versus the direct length of every cable run. The total amount of cable used was 10–12 percent longer than the design runs to account for the fact that the cables don’t follow the design route exactly and excess cable is needed for splicing and service loops. You’ll find that out when you troubleshoot with an OTDR, of course, since the actual location of a problem is not going to be exactly where it would be shown on a design document.