close

Вход

Забыли?

вход по аккаунту

?

el%3A19970471

код для вставкиСкачать
lime-steered array with a chirped grating
beamformer
ing the R F throughput with a network analyser directly following
the photodetectors. The commercial gratings used here had peak
98% reflection at 1556nm, a length of 3.4cm, and a chirp of
1.2nm/cm.
J.E. R o m a n , M.Y. Frankel, P.J. Matthews and
R.D. Esman
L0o oI )
VI
a
-2 00
Indexing termx Gratings in fibres, Antenna phased arrays
W
U
The authors have developed a novel, simple chirped-grating based
beamformer for time-steered array antennas. The beamformer is
based on commercially available chrped fibre gratings and optical
circulators and has a measured bandwidth of over 18GHz. The
first-ever array demonstration achieves quasi-continuous 1D
steering over 45".
Optical techniques for squint-free ultra-wideband control of timesteered array antennas have been under intense study recently
[1-51. A variety of drawbacks including component complexity,
excess losses, andlor operational constraints have precluded a wide
acceptance of these techniques. A chirped-grating based architecture is an attractive altemative to overcome many of these problems. To date, all implemented grating-based beamforming
architectures have used discrete gratings, which do not permit continuous beamsteering and have technological difficulties in achieving repeatable grating spacing control [6]. Similarly, a previously
proposed chirped-grating based beamformer [7] requires a range of
gratings with different precise characteristics and has never been
reduced to practice.
w e have proposed a novel, simple beamformer that is implemented using only nominally identical chirped gratings, relaxing
most fabrication requirements [SI. Here, we demonstrate the
beamformer and use it to control an array antenna with the measurements showing squint-free steering over 45".
photodetectors
computer
network
analyser
W
O
._
9
a,
-2 00
-LO 0
15L9
1553
1557
1561
=
A , nm
Fig. 2 Measured delay for one and two cascaded gratings and delay difference
between
gratings
grating 1
.. - grating 1 + 2
........... delay error, ps
~
-
The gratings had a measured flat-amplitude reflection bandwidth of 13nm. The grating delay characteristics, shown in Fig. 2,
are matched to f i p s over the wavelength range of 1551 to
1561mn as measured at 12GHz. Similar data were obtained across
the complete 18GHz bandwidth. The maximum measured delays
were 3 2 0 ~ sfor a single grating and 6 4 0 ~ sfor two cascaded gratings. The deviations of the delay from linearity are not important
for time-steering purposes as long as all gratings are matched.
For array pattern measurements the photodetector outputs were
fed to the emitters of the microwave D-lens. The microwave Dlens was designed for -3.2GHz centre frequency operation but
provided adequate performance (kl dB amplitude and So"phase
ripple) over the 3.0 to 3.8GHz frequency range. It consisted of a
parallel plate waveguide with a series of 34 R F emitter probes
arranged on a half-circle with a 0.508m radius. A similar series of
RF receiver probes were arranged along the half-circle base. Three
photodetector outputs fed three emitter probes separated by d l 7
radian arcs. The receiver probes were separated by N 2 at 3.2GHz
(-0.047m). In effect, the D-lens provided a Fourier transform
function such that measurements at each R F receiver corresponded to a specific farfield radiation angle.
1432/11
>
.-
?
%
Fig. I Dispersive-grating based fibue-optic beamformer measurement
system
0
c
The fibre-optic beamformer implementation is based on cascaded chirped gratings, and is shown in Fig. 1 . The optical source
is a wavelength-tunable semiconductor laser. A wideband electrooptic Mach-Zehnder modulator (MZM) amplitude modulated the
optical carrier with an R F signal from the network analyser. A
fibre-optic coupler provided the first, wavelength-independent,
optical signal tap. The remaining optical signal entered a 6-port
circulator with nominally identical chirped gratings and optical
taps arrangGd such that thcj optical signal at each successive tap
propagated through an increasing number of grating elements.
The overall delays from each tap were equalised to within +Ips at
the grating centre wavelength of & = 1556nm using additional
non-dispersive fibre. Hence, the overall differential time-delay at
each optical tap was linearly related to the sequential tap number
and to the wavelength detuning from the centre wavelength. Fibreoptic attenuators were used to equalise the amplitudes of the
tapped signals to within 0.2dB. The overall beamformer bandwidth was limited by the available MZM and photodetectors to
I8 GHz.
Accurate simultaneous time-delay control of many elements
requires well-matched grating delay characteristics. The system in
Fig. 1 was used to evaluate the grating characteristics by measur-
652
D
W
-a
&
z
.2
-LO
0
-20
angle, deg
20
LO
1132131
Fig. 3 Meusured and caktiuka& array pattern with bcamfurmcr acering
angle and frequency as parameters
(i) main beam (h= 1556nm)
(ii) main beam (1
= 1551nm)
0
3.0GHz (measured)
-0-
-0A
-a-
3.0GHz (calculated)
3.3GHz (measured)
3.3 GHz (calculated)
3.6GHz (measured)
3.6GHz (calculated)
As mentioned, the signal delays among all three taps were
equalised at & = 1556nm, corresponding to the array being
steered for broadside radiation. Fig. 3 shows the signals measured
at three frequencies across the D-lens focal plane. A main lobe
ELECTRONICS LETTERS
10th April 1997
Vol. 33
No. 8
centred at broadside and two sidelobes on either side can be
clearly seen. The frequency responses have been offset for clarity
and we can observe the expected narrowing of the main lobe with
increasing frequency. There is very good agreement between the
measured data (filled symbols) and the expected response calculated assuming ideal conditions (open symbols).
Broadband steering is demonstrated simply by tuning the laser
wavelength, without any other adjustments. Tuning the wavelength to IL = 1551nm introduces a 137ps delay between consecutive taps, as determined from Fig. 2. This corresponds to the main
beam being steered to +25”, as can be directly observed from Fig.
3. There is good agreement between the measured data and the
patterns calculated assuming ideal true time-delay steering. There
is no observable squint across the 3.0 to 3.6GHz range within our
experimental resolution. Similarly, detuning the laser to longer
wavelengths steers the microwave beam to negative angles.
In conclusion, we have demonstrated the first chirped-grating
based time-steered antenna. Quasi-continuous beamsteering was
achieved over 45“ using a simple, novel architecture that utilises
commercially available broadband chirped gratings and optical
circulators. The results show that chirped grating-based architectures are a strong contender for optical control of time-steered
array antennas.
Acknowledgments: This work was supported by the Office of
Naval Research. The authors would like to thank M. Parent for
the loan of the D-lens.
0 IEE 1997
Introduction: Cross-phase modulation (XPM) is an example of
fibre nonlinear effects which can limit the distance and the capacity of WDM optical fibre transmissions [l - 41. However, the
transmission limit imposed by XPM is still unclear due to the
complexity of its mechanism: XPM on a signal channel is mainly
caused by the neighbouring WDM channels travelling at nearly
the same group velocity [4], then fibre dispersion converts the
XPM to intensity modulation (IM), or waveform deformation.
Previously, we experimentally investigated the impact of XPM in
dispersion-shifted fibre (DSF) transmissions, and demonstrated
that the variation of DSF dispersion can enhance the XPM effect
[3]. In this Letter, we analyse the XPM effect in detail experimentally and by numerical simulations in terms of two normalised
parameters, the XPM induced phase-shift and the fibre dispersion
interacting with the XPM. We provide the XPM limit in WDM
transmissions in a normalised formula and check its validity for
various parameters, such as fibre dispersion, fibre loss, channel bit
rate, bit walk-off and fibre input power. We also investigate the
XPM effect on dispersion compensated transmissions, and show
that the XPM can be enhanced by dispersion compensation and
that the optimum amount of compensation should be determined
considering the XPM.
2Ops/div
LO ps/div
63
m
9
c
b
4 March 1997
Electronics Letters Online No: 19970471
J.E. Roman, M.Y. Frankel, P.J. Matthews and R.D. Esman (Code
5672, Naval Research Laboratory, Washington, DC 20375, USA)
C
E-mail: jroman@ccf.nrl.nay.mil
-5
0
5 10 15
a fibre input power,dBm/ch
-10
References
d
m
U
and TOUGHLIAN, E.N.: ‘Photonic aspects of modern
radar’ (Artech House, Boston, 1994)
2 DOLFI, D., JOFFRE, P., ANTOINE, J., HUIGNARD, J.-P, PHILIPPET, D , and
GRANGER, P.: ‘Experimental demonstration of a phased-array
antenna optically controlled with phase and time delays’, Appl.
Opt.. 1996, 35, (26), pp. 5293-5300
3 GOUTZOULIS. A.P., DAVIES, D.K., and ZOMP, J.M : ‘Hybrid electronic
fibre optic wavelength-multiplexed system for true time-delay
steering of phased array antennas’, Opt. Eng., 1992, 31, (ll), pp.
1
ZMUDA, H.,
23 12-2322
4
SOREF, R.: ‘Optical dispersion technique for time-delay beam
steering’, Appl. Opt., 1992, 31, (35), pp. 7395-7397
5 FRANKEL, M.Y., MATTHEWS, P.J., and ESMAN, R.D.: ‘Two-dimensional
fibre-optic control of a true time-steered array transmitter’, IEEE
Trans. Microw. Theory Ttchnol., 1996, 44,(12), pp. 2696-2702
6 ZMUDA, H , SOREF, R., PAYSON, P., JOHNS, s., and TOUGIILIAN, E.N.:
‘Photonic beamformer for phased array antennas using fibre
grating prism’, ZEEE Photon. Techno1 Lett., 1997, 9, (2), pp. 241-
243
SOREF, R.A : ‘Fiber grating prism for true time delay beamsteering’,
Fiber Integr. Opt., 1996, 15, pp. 325-333
8 ROMAN, J.E., FRANKEL, M.Y., MATTHEWS, P.J, and ESMAN, R.D.: ‘Timesteered array with a chirped grating beamformer’. OFC ’97, 1997,
Dallas, USA, Post-deadline paper PD28-1
7
Analysis of cross-phase modulation (XPM)
effect on WDM transmission performance
N. Kikuchi, K. Sekine and S. Sasaki
Indexing terms: Wavelength division multiplexing, Cross-phase
modulation
The authors analyse the effect of cross-phase modulation (XPM)
on wavelength division multiplexing (WDM) transmissions by
experiments and numerical simulations, and provide the XPM
limitation in a normalised and closed form formula. They also
investigate the effect of XPM on dispersion compensated
transmissions. It is shown that the location of dispersion
compensators and the amount of Compensation should be
determined by considering the XPM effect.
ELECTRONICS LETTERS
10th April 1997
Vol. 33
Fig. I Expevimental and simulated examples of XPM pcnalty and eyepatterns in 10Gbit/s 2-channel W D M transmission
Wavelength separation: I nm, a-parameter: 0.6
Transmission fibre consists of 60km DSF (-0.4pslnmikm) followed
by 20km conventional fibre (17psinmlkni)
Eye patterns (experimental: left, simulated: right) correspond to fibre
input power of +lOdBm/channel
a XPM penalty (experimental:circles, simulated: lines)
0 best
0 worst
h Single channel
c WDM best case
d WDM worst case
Simulation model: We used the fibre transmission simulation based
on the split-step Fourier method and assume a two-channel
repeaterless WDM transmission with IMDD NRZ format modulated by mutually uncorrelated 128-bit long pseudorandom patterns. Fibre parameters were assumed such that the nonlinear
index n2 was 2.6 x 1W6cm2/W,the core area A , was 5 0 p * , and
the loss a is 0.25dBkm. The polarisations of WDM signals were
set to the same state to simulate the worst case. We suppressed the
four-wave mixing (FWM) effect by changing the fibre dispersion
with an amplitude of 5 p s i n ” by a 1km step, since the FWM
can be suppressed in practical system using the scheme with unequal channel spacing or dispersion management. The degradation
via XPM was evaluated in terms of the XPM penalty, and the
receiver sensitivity degradation at a bit error rate (BER) of IP9
from the single channel (non-WDM) sensitivity
Comparison with experiment: Fig. 1 shows examples of the experimental and simulated XPM penalty a and eye-patterns b-d in a 10
Gbit/s WDM transmission over a 60km DSF followed by a dispersion of +340ps/nm. The XPM penalty is dependent on the relative bit-phase o f two WDM sigiial pattenis at thc DSF input [3],
so we obtained the best and the worst XPM penalties changing the
relative bit-phase from 0 to 1OOps. In each case, both the experimental and simulated penalties showed good agreement. The bitphases corresponding to the best and the worst penalties are -50
and Ops, respectively. In the best case (Fig. IC), strong timing jitter
is observed. Conversely, in the worst case (Fig. 14, the pulse
No. 8
653
Документ
Категория
Без категории
Просмотров
0
Размер файла
308 Кб
Теги
3a19970471
1/--страниц
Пожаловаться на содержимое документа