Demonstration of Tunable Optical Generation of HigherOrder Modulation Formats using Nonlinearities and
Coherent Frequency Comb
M. R. Chitgarha1, M. Ziyadi1, S. Khaleghi1, Ahmed Almaiman1, A. Mohajerin-Ariaei1, Ori Gerstel 2, Loukas
Paraschis2, C. Langrock3, M. M. Fejer3, J. D. Touch4, Alan E. Willner1
1) Ming Hsieh Department of Elec. Eng., Univ. of Southern California, 3740 McClintock Ave, Los Angeles, CA 90089, U.S.A., [email protected]
2) Optical Networking, Advanced Technology and Planning, Cisco Systems, Inc U.S.A
3) Edward L. Ginzton Laboratory, 348 Via Pueblo Mall, Stanford University, Stanford, CA 94305,
4) Information Sciences Institute, University of Southern California, 4676 Admiralty Way, Marina del Rey, CA, 90292, USA U.S.A
Abstract: We demonstrate the generation of optical 16-QAM and 64-QAM at EVM 6.8% and
6.4% respectively using nonlinearities and coherent frequency comb. We also demonstrated a
successful transmission through 80-km SMF-28 after compensating with 20-km DCF with
negligible penalty.
© 2013 Optical Society of America
OCIS codes: (060.2360) Fiber optics links and subsystems; (190.4223) Nonlinear wave mixing
1. Introduction
There has been significant interest in higher-order modulation formats for optical communication systems due to the
higher spectral efficiency in terms of bit/s/Hz. Specifically, quadrature phase shift keying (QPSK) is an example of
4-ary phase encoding, whereas 16 quadrature amplitude modulation (QAM) is an example of 16-ary
amplitude/phase encoding.[1,2] Moreover, QAM can go to higher orders, such as 64 and beyond.
A common technique for generating higher-order formats is to use electronic circuits to drive an I/Q modulator.
However, key challenges for this approach are: (a) the limited linearity of the electronics at high baud rates, such
that the spacing of the data constellation points on the I/Q plot will no longer be uniform, and (b) electronic
approaches may become difficult at rates exceeding 100-Gbaud.
As an alternative, nonlinear optical processes hold the promise of high speed, format and phase transparency, low
noise, and high linearity.[3]. As an example, an NOLM loop can be used to multiplexed 4 OOK signals into a 16QAM[4]. These methods tend not to be transparent for phase modulation formats. We recently demonstrate a
method to multiplex QPSK signals into a 64-QAM using 2 cascaded nonlinear stages [5]. In this method, however,
two nonlinear stages degrade the quality of the generated QAM. A laudable goal would be to use fewer nonlinear
stages for better efficiency and the potential to go to higher-order formats.
In this paper, we demonstrate a scheme for tunable optical generation of higher-order modulation formats using
nonlinearities and coherent frequency comb. Due to the coherency of comb fingers, coherent multiplexing process
can be done completely in one nonlinear stage. Optical 16-QAM and 64-QAM are generated at EVM 6.8% and
6.4% respectively. We also demonstrated a successful transmission through 80-km SMF-28 after compensating with
20-km DCF with negligible penalty.
2. Concept
f
Amplitude Weights:
[ 1 , 0.5 , 0.25 ]
QPSK
Modulation
Frequency
Comb
e.g., Modulated
64-QAM fingers
f
Frequency
Comb
Pump
D3 D2 D1
P
PPLN
(Phase
Coherent
Addition)
QPSK
Modulated
Pump
fingers
QAM S1 S2 S3
f
f
+
64-QAM
1st QPSK
2nd QPSK
+
1st QPSK=45º
•n QPSK signals
generate 4n-QAM
3rd QPSK
+
=
2nd QPSK=135º 3rd QPSK=225º
Fig. 1. Conceptual block diagram higher-order QAM generation
• Vector addition of
QAM symbols
The conceptual block diagram of the tunable optical QAM transmitter is in Fig. 1. Multiple fingers from a coherent
comb source are selected and modulated using an I/Q modulator with QPSK modulation format. These signals along
with another set of coherent comb fingers with equal frequency spacing and a CW pump laser (ECW) are injected
into a periodically-poled-lithium-niobate (PPLN) waveguide to perform coherent addition of QSPK signals. This
process can be done through two cascaded second order nonlinear wave mixings (sum frequency generation (SFG),
difference frequency generation (DFG)). The multiplexed signal becomes proportional to (ECW(t))*Σ EPi(t).ESi(t).
The amplitude and phase of comb fingers determine the coefficients of this coherent addition and thus, by varying
these parameters, QAM with different constellation size and/or encoding can be generated.
3. Experimental Setup
Flat 20-GHz Comb Generation
20/25.1 Gbaud
BPF
400m
1.5
-20
-40
1
400m
0.5
0
LCoS
Filter
-0.5
MLL
-1
-1.5
-6
-4
-2
0
2
4
6
HNLF
DLI
5 nm
5 nm
-60
PPLN
-10
DCF
5 nm
PC
ATT
-80
1 nm
1542 1545 1548 1551 1554 1557
-20
A
λPUMP
-30
1 nm
EDFA
Coherent
Receiver
-40
-50
1546
1548
1550
1552
1 nm
1 nm
1554
Fig. 1. Experimental setup of higher-order QAM generation.
The experimental setup for the tunable QAM encoder is shown in Fig. 2. A mode-locked laser with 10GHz repletion
rate and 2-ps pulse width is used to generate a coherent comb with 10-GHz frequency spacing. This optical comb is
passed through a DLI with FSR 20-GHz to increase the frequency spacing. The 20-GHz comb is then passed
through an HNLF fiber to generate a flat and broad spectrum. A Liquid Crystal on Silicon (LCoS) filter can be
utilized to select and write complex weights on comb fingers and separate them into signal path and pump path. A
nested Mach-Zehnder modulator is used to generate the 20/25.1-Gbit/s BPSK/QPSK data (PRBS 231-1) on the signal
path. These signals along with the comb fingers selected for pump path and a CW laser pump are sent to a PPLN (5cm length) after enough amplification to perform coherent multiplexing of original signals. The multiplexed signal
is then filtered and sent to the coherent receiver.
4. Results and Discussion
a)
b)
c)
1
20-Gbaud QPSK-1 B2B
20-Gbaud QPSK-2 B2B
20-GbaudQPSK-3 B2B
20-Gbaud16-QAM B2B
20-Gbaud16-QAMTransmited
25.1-Gbaud QPSK-1 B2B
25.1-Gbaud QPSK-1 B2B
25.1-Gbaud16-QAM B2B
Generation of QAM from QPSK signals
1
Generation of PAM from BPSK signals
0.6
0.8
0.4
0.6
0.2
0.85
0.6
0.4
1
0.4
0.2
0.2
0
0
-0.2
-0.4
0
-0.2
0.8
4-PAM
-0.4
0.6
-0.2
0.4-0.6
-0.4
0.2-0.8
-0.6
0 -1
-0.6
-0.7
-1
-1
16-QAM (EVM 6.8 %)
-0.5
0
0.5
-0.5
0
0.5
1
2
4
-0.2
-0.8
1
-1
-5
-1
-Log10(Bit Error Ratio)
0.8
1
64-QAM (EVM 6.4 %)
-4
-3
-2
-1
0
1
2
3
8-PAM
-0.4
4
5
6
-0.6
5
10
15
20
25
30
35
Fig. 3. (a) 16-, 64-QAM generation using 20-Gbaud QPSK, (b)4-, 8-PAM generation using 20-Gbaud BPSK (c)OSNR(dB)
BER measurements
-0.8
-1
-1
-0.5
0
0.5
1
Fig.3(a) shows QAM generation using 20-Gbaud QPSK signals. By coherent multiplexing of two QPSK signals
with appropriate weights, a 16-QAM with EVM 6.8% can be generated. 64-QAM constellation with EVM 6.4% is
also shown in fig.3(a) which can be generated from three QPSK signals. If the original signals are two or three
BPSK signals, a pulse-amplitude-modulation (PAM) signals with 4 and 8 points can be generated respectively (Fig.
3(b) ). The performance of the higher-order QAM encoder is assessed using BER measurements. As it can be seen, a
16-QAM signal can be generated at both 20-Gbaud and 25.1-Gbaud. The 20-Gbaud 16-QAM is also transmitted
through 80-km SMF-28 and 20-km DCF fiber with negligible power penalty.
Acknowledgements
We acknowledge the support of DARPA, Cisco, CIAN and NSF.
5. References
[1] Y. Koizumi, Optics Express, Vol. 20, Issue 11, pp. 12508-12514 (2012)
[2]D. J. Geisler, Opt. Express 19, 8242-8253 (2011)
[3] Agrawal, G. Nonlinear Fiber Optics (Academic Press, 2001)
[4] G. Huang, JLT, Vol. 30, Issue 9, pp. 1342-1350 (2012)
[5] M. R. Chitgarha, ECOC (2012)