Free Space Optics (FOS)

Free Space Optics, also known as FSO, or Optical Wireless technology is increasingly used by service providers as a rapid-rollout, instant infrastructure to meet massive growth in user numbers and the demands of new-generation high-bandwidth wireless applications. By using advanced wireless laser technology, modern FSO vendors offer point-to-point bandwidths exceeding 2.5Gbps over distances up to a few kilometers.

Free space optics is experiencing a renaissance due to the ability to provide gigabit bandwidths over modest distances without risk of interference using unlicensed infrared frequency spectrum. Unlike fiber, most of the power emitted by an FSO system is typically lost in the atmosphere - hence transmit power tends to be much higher than for fiber systems.

From the diagram it should be reasonably clear that a lower beam divergence (narrow)results in a lower loss, and a higher divergence (wide) more loss. So what prevents us using narrow beams? The answer is movement, either long-term due to tower sway (wind) or thermal (heating), or short-term (vibration, wind-loading).

Quite clearly, a large-area (>100mm) transmitter is much better than a small (<30mm) one in relation to beam-breakage due to birds, etc - if the area is larger than the size of the bird, beam-break will not occur.

For systems employing multiple transmitter devices, designers have favored two overall schemes:: multi laser - multi small aperture (ML-MSAk), where each device has a separate collimating lens, and multi laser - single large aperture (ML-SLA)where combining fiber optics are used to sum the power from several devices to a single emission aperture.

For ML-MSA, there is not enough overlap between the beams at moderate distances, and holes exist between. Thus there is no optimal alignment from transmitter to the receiver, and typically the power from only one laser enters the receiver, negating the perceived benefit of a multiple-laser system. At long range, things improve significantly, though the power density is still not even, and the performance sub-optimal.

By comparison, multiple laser – single large aperture ML-SLA solves all of these problems. At the faceplate, power from several lasers is already combined, giving even over a large aperture at lower (hence safer) power density. At medium range there is no overlap holes problem, and at long range, coverage is optimal.

Multi-beam systems also suffer losses from non-parallelism of the beams. Minute differences due to manufacturing tolerances cause each beam to point in slightly different directions: there being no such things as perfect mechanics, this problem is inevitable at least to some degree. For low-divergence (<3mRad) systems, the effect is severe; a mispointing of 0.3mRad (=0.017 degrees, or 10%) would move the beam centers by 30cm at 1km; seriously eroding theoretical fade margin by 3-10dB - a major problem.