5 Optical receivers
1. (a) Explain what is meant by quantum noise in an optical receiver.
(b) Considering that all other sources of noise may be neglected, other than quan-
tum noise, assume that a pulse with energy E is detected over the interval
. Show that the error probability of the pulse being detected as a zero (zero
electron-hole pairs generated) is equal to 109 if
h
E = 21
where is the quantum efficiency of the photodetector.
(c) For a transmission rate of 10 Mbit/s and identical probabilities for ones and
zeros, calculate the receiver sensitivity corresponding to a bit error rate of 109
( = 1300 nm.
3. A digital transmission system sends 200 000 bits of information per second. During
the transmission noise is added to the signal such that at the decoder input the
pulses have an amplitude of 1 V in amplitude and the rms noise voltage is 0.2 V.
(a) Assuming that ones and zeros are equally likely to be transmitted what is the
average time of error occurrence?
(b) If the voltage amplitude is doubled while keeping the noise the same, how does
this time change?
5. An electrical amplifier with a bandwidth of 0.5 GHz and an equivalent spectral noise
referred to the input of 5 1024 A2 /Hz is used to amplify a 622 Mb/s signal. The
amplifier is used on a 120 km optical fibre link with 0.3 dB/km of attenuation. The
transmitter produces 1 mW pulses into the fibre on the 1.3 m wavelength.
(a) Estimate the BER on the link if a PIN photodetector with a responsivity
R0 = 1 is used on the optical receiver.
(b) What is the quantum limit for the signal-to-noise ratio in dB?
(c) What is the required optical power at the receiver to achieve BER = 109 .
Optical CommunicationsOptical receivers 17
(d) Assuming now that an APD is used with an excess noise factor F (M ) = M ,
determine the optimum gain for the APD and the corresponding BER.
(a) Calculate the quantum limit for the sensitivity of the receiver for a bit error-
rate of 1012 .
(b) What is the power penalty (in dB) for this receiver, if an electrical amplifier
is used with an equivalent input noise of 1022 A2 /Hz. Neglect all other noises
except shot noise and amplifier noise.
(c) Is it possible to achieve a sensitivity for this receiver of 30 dBm using an
APD with an excess noise factor of F (M ) = M ?
7. We would like to design a binary digital decision circuit for an optical receiver. A
1 corresponds to 2.5 V and a 0 corresponds to 0 V. Assume that the noise in the
receiver is dominated by thermal noise with equal rms-value for a 1 and 0 and
that the probability for a 1 and a 0 is the same.