486 J
ILIN
Z
HENG
et al.
En/decoder within a group delay compensation function is a novel approach to
solving the dispersion problem. In the previous research, literature [8] firstly proposed
time-spreading/wavelength-hopping (TS/WH) group delay compensating a fiber Bragg
grating (FBG) en/decoder, which validated the feasibility of such concepts experi-
mentally. However, such en/decoders need to be improved if applied into practice.
First, the FBG fabrication equipment, such as phase masks, is required to vary with
the address-code or the amount of dispersion-compensation, which increases the com-
plexity and cost of experimental realization. Second, the in-band dispersion cannot be
compensated, which will deteriorate the system performance, especially in the case of
longer transmission distance or higher data rate.
In this work, we develop a novel dispersion compensating FBG-based en/decoder.
Compared with the previous dispersion compensating en/decoder, our new technology
is easier to be experimentally realized, and most importantly, the in-band dispersion
can also be compensated. Such advances are expected to make the dispersion compen-
sating en/decoder more practical.
2. Principle
Figure 1 shows the
τ
–
λ
(i.e., time-delay versus wavelength) distribution of TS/WH
signals evolving at different stages. The address code used here is the prime-hop
sequence (P0H2, p = 5) used in [8]. It can be seen after propagating in the G.652 fiber
that there is not only extra relative group delay among different wavelength-chips, but
also extra dispersion inside one chip. As reported in literature [8], the “decoding A”
in this figure represents using an out-band dispersion compensating decoder. Such
a decoder cannot compensate the in-band dispersion, and the residual dispersion will
result in pulse broadening and deteriorate the system performance. However, this paper
proposes the “decoding B” technique, which uses a total dispersion (both out-band and
in-band) compensating decoder, and can completely compensate all the dispersion.
For the encoded signals, the time delay between one tip pulse (central wavelength
λ
1
) and another tip pulse (central wavelength
λ
i
) is defined as (i =2,3,…,p,
p is the number of wavelengths). When the encoded signals propagate in the SMF with
the length of z, an extra time delay (i =2,3,…,p) among tip pulses is added:
(1)
where D is the wavelength dispersion coefficient of SMF. On the other hand, every
reflective channel of the encoding spectrum occupies a certain bandwidth, which is
defined as Δ
λ
in this paper (Δ
λ
<min(
λ
i
–
λ
j
), i ≠ j). So the in-band dispersion
(i =1,2,…,p) inside one chip is produced:
(2)
ΔT
i
λλ
i
–(),
in-disp
Dz
λλ
i
–(),=