"Introduction". In: Fiber-Optic Communication Systems

(Brent) #1
6.5. SYSTEM APPLICATIONS 261

6.5 System Applications


Fiber amplifiers have become an integral part of almost all fiber-optic communica-
tion systems installed after 1995 because of their excellent amplification characteristics
such as low insertion loss, high gain, large bandwidth, low noise, and low crosstalk. In
this section we first consider the use of EDFAs as preamplifiers at the receiver end and
then focus on the design issues for long-haul systems employing a cascaded chain of
optical amplifiers.


6.5.1 Optical Preamplification


Optical amplifiers are routinely used for improving the sensitivity of optical receivers
by preamplifying the optical signal before it falls on the photodetector. Preamplifica-
tion of the optical signal makes it strong enough that thermal noise becomes negligible
compared with the noise induced by the preamplifier. As a result, the receiver sen-
sitivity can be improved by 10–20 dB using an EDFA as a preamplifier [107]–[112].
In a 1990 experiment [107], only 152 photons/bit were needed for a lightwave system
operating at bit rates in the range 0.6–2.5 Gb/s. In another experiment [110], a receiver
sensitivity of− 37 .2 dBm (147 photons/bit) was achieved at the bit rate of 10 Gb/s. It
is even possible to use two preamplifiers in series; the receiver sensitivity improved
by 18.8 dB with this technique [109]. An experiment in 1992 demonstrated a record
sensitivity of− 38 .8 dBm (102 photons/bit) at 10 Gb/s by using two EDFAs [111].
Sensitivity degradation was limited to below 1.2 dB when the signal was transmitted
over 45 km of dispersion-shifted fiber.
To calculate the receiver sensitivity, we need to include all sources of current noise
at the receiver. The most important performance issue in designing optical preampli-
fiers is the contamination of the amplified signal by the ASE. Because of the incoherent
nature of spontaneous emission, the amplified signal is noisier than the input signal.
Following Sections 4.4.1 and 6.1.3, the photocurrent generated at the detector can be
written as
I=R|



GEs+Esp|^2 +is+iT, (6.5.1)

whereRis the photodetector responsivity,Gis the amplifier gain,Esis the signal field,
Espis the optical field associated with the ASE, andisandiTare current fluctuations
induced by the shot noise and thermal noise, respectively, within the receiver. The
average value of the current consists of


I ̄=R(GPs+Psp), (6.5.2)

wherePs=|Es|^2 is the optical signal before its preamplification, andPspis the ASE
noise power added to the signal with the magnitude


Psp=|Esp|^2 =Ssp∆νsp. (6.5.3)

The spectral densitySspis given by Eq. (6.1.15) and∆νspis the effective bandwidth
of spontaneous emission set by the amplifier bandwidth or the filter bandwidth if an
optical filter is placed after the amplifier. Notice thatEspin Eq. (6.5.1) includes only

Free download pdf