Making the case for human hands rather than machines as the preferred fiber-polishing mechanisms.
By William Graham, Mississauga Training
Fiber-optic installers started polishing fiber-optic connectors in the late 1970s. At that time they injected epoxy in the connector and let it set for 24 hours. Then the two or three steps of polishing began. Data rates were low and basically, if there was light, the system worked. Through the 30- to 100-power microscope, the connectors looked acceptable.
In the mid-1980s ovens were invented, as were new types of epoxy that would cure in 10 minutes at 100 degrees Celsius. This practice continued, and in the meantime innovative companies searched for new opportunities.
The case for prepolished
Somewhere along the line, someone decided that installers didn't like polishing connectors, and also noted the high cost of polishing film, alcohol and other consumables. The pitch made to the end-users was that the manufacturer could save them money by selling them a connector that already had a polished fiber stub within it. Figures were compiled, based on labor and material. And to the end-user, the numbers looked good. Had the manufacturers involved the installer in this equation, the results might have been different.
With these connectors-which came loaded with a prepolished stub of fiber-all the installer had to do was strip and clean the fiber, then cleave it to close to 90 degrees, put it in the connector and mechanically lock it in place. Some index-matching gel was in the connector to reduce insertion loss and back-reflection loss. At the time these types of connectors were emerging on the scene, about 1993, back reflection was not much of an issue.
The tool used to install these connectors was awkward, and the cleaver not so accurate. However, these connectors were pushed relentlessly to customers. In 1996 a second make of this connector type came on the market, and other revisions followed to the point where most companies have their own brand of what they call "field-installable connector."
From those days in the 1990s, the connectors have evolved to the point at which toolkits can cost between $1,000 and $3,500 and the connectors cost between $12 and $18 apiece as companies search to save money. Recently I found one brand that includes the connector-installation tool for free with a connector purchase, and the connectors cost about $12 each.
As the years went on, cleavers improved as well as the geometry of the fiber and the connectors. Many companies accepted this cleave-and-crimp method as the way to install fiber-optic connectors, and did not seek additional opinions or options.
In the mid-1990s, as data rates increased, Fresnal (back) reflections were recognized as a problem. Back reflections are caused by differences between the refractive index of the mating connectors. This phenomenon is caused by dirt, temperature cycling, a gap between the mated pieces, or alcohol residue. The light shining back causes the laser to reduce power, possibly causing the system to fail.
Another, more-practical solution to Fresnal reflection was the angle physical contact (APC) connector. With an APC, the end of the connector ferrule, rather than being 90 degrees, was only 8 degrees. This had the effect of deflecting the reflected light into the cladding rather than back at the laser. This was, and still is, the most-effective solution for back reflection. We teach a class focused on fiber-to-the-home cabling, in which we use epoxy-based APC connectors with great success.
The pigtail tale
Despite the effectiveness of the APC connector, users began hearing about another fiber-connectivity option that was supposed to "come to the rescue"-the pigtail. The pigtail consisted of 6 feet of fiber attached to a prepolished factory connector. The connector had a physical contact (PC), ultra physical contact (UPC) or APC finish. The pigtail was spliced to the fiber, and the splice was protected in a splice tray.
Each of these connectivity approaches, the prepolished connector and the spliced pigtail, has its drawbacks. In the case of the prepolished, field-installable connector, we end up with two sources of loss rather than just the one that occurs with an epoxy/polish connector. Little is made of this fact, however, because the Telecommunications Industry Association (TIA) standard-development groups wrote specifications allowing for a loss of 0.75 dB on a mated pair of connectors. This level of allowable loss provides sufficient leeway for typical prepolished connectors to meet the standard.
With the pigtail approach, a splice tray is required, which occupies space and adds to cost. Also, the splice itself is both an additional source of loss and an additional potential point of failure.
Why are these fiber-optic connectivity methods so popular? We all have been told by providers of these products and technologies that we cannot expect a technician to hand-polish a connector with any success. Many end-user customers believe this and (literally) bought into the quicker, easier methods. Unfortunately, in this author's opinion, we are ending up with connections that: 1) have additional loss, 2) change their loss values as a result of temperature cycling at the splice point in the connector, 3) include additional points of failure, 4) come at a higher finished cost, and 5) have the potential to require early replacement.
The solution, as I see it, is to back up to the old way. Use an epoxy connector and polish it by hand. An epoxy connector can be polished in about 30 seconds by a trained technician. This applies equally well to a PC, UPC or APC connector.
Knowing how much to polish at each step, and when to stop, is critical for a low-loss, low-back-reflection connector. It is ridiculous to expect that a machine, on which we set a timer, can do this better than it can be done by hand. Multi-connector polishing machines have their place for high-production facilities, and a price is paid in loss and back reflection. It is not possible to polish 12 or 24, or more, connectors together without having some difference in reflection and loss. For small quantities of connectors, I question their need or cost-effectiveness.
William Graham is a master fiber-optic instructor and director of The Fiber Optic Association (FOA; www.thefoa.org). He has been certifying fiber-optic technicians through Mississauga Training (www.fiberoptictraining.com) for the past 17 years.
More perspective on epoxy/polish connectors and methods
Eric R. Pearson, the principal of Pearson Technologies Inc. (www.ptnowire.com), devotes a chapter of his book "Professional Fiber Optic Installation - The Essentials for Success" to connectors. In that chapter he writes, "Fiber-optic connectors can be installed by many mehtods. These methods were developed to address different concerns, such as reliability, convenience, time to install and cost to install. Most of the development of methods has been to reduce one or more of the cost factors. In most cases, the effort to reduce the total installed cost resulted in increased connector cost." -Ed.