Next: Results and Conclusions
Up: A novel fiber-optic subscriber
Previous: OA placement in proposed
In computation, one of the most important thing is loss model. If the
transmitter power is known (say PT) and SOAs are not
there between CO and receivers for which BER is computed, the loss (in
dB) between transmitter and receiver assuming that receiver is in ithgroup starting from CO is given by
|
|
|
|
|
|
|
(3) |
In the above equation, LW is insertion loss of wavelength
multiplexer at CO, Lsp splice loss,
fiber attenuation
per unit length, L the length of fiber between two consecutive
OADMs, Lm insertion loss of OADM, L12 insertion loss for drop
signal in OADM, Li insertion loss of
couplers used for
making star coupler, and n the number of users in a group.
The loss (in dB) between transmitter at CO and first OA is used to find the
input power to SOA. If the first SOA is placed before the ithgroup the loss between transmitter and first SOA is given by
|
(4) |
If the two consecutive amplifiers are placed before ith and
jth OADM respectively, then loss (in dB) between them is given by
|
(5) |
Further loss (in dB) between a receiver in lth branch and nearest
amplifier (which is before tth OADM) is given by
Here b is total number of branches in the network.
In order to compute the gain saturation, it is assumed that bits in
all the channels are synchronised. In order to find the impact of
other channels, two cases have been considered. In one case,
average value of saturated gain has been used. In the other case BER
for every bit pattern in interefering channels is computed. Thereafter,
average BER is computed.
Next: Results and Conclusions
Up: A novel fiber-optic subscriber
Previous: OA placement in proposed
Dr.Yatindra Nath Singh
2002-05-23