@fibresystemsmag |
www.fibre-systems.com
FEATURE DWDM AND ROADM NETWORKS
Figure1: A simplified schematic of C+L
Figure2: A simplified schematic of C+L hybrid EDFA/Distributed Raman Amplifier
can interact with and amplify a 1550nm signal. But, crucially, this also works at other wavelengths.
Non-linear thinking Distributed amplification exploits this effect that occurs continuously within fibre, as opposed to lumped amplification at the end of each span. Compensating for losses in the fibre span, the Raman noise factor is lower than it would be for the same span with an EDFA, Pelouch explained. ‘Distributed Raman improves the OSNR, allowing longer transmission distances and/or higher spectral efficiency,’ he says. Yet Raman amplification is less commonly used than EDFAs, because it needs higher power pump lasers to generate the desired gain at the signal wavelengths. Tis consumes more electrical power and increases the risk of damage to optical connectors and components outside a controlled environment. Sufficiently high powers can also change the fibre’s refractive index through the Kerr effect. ‘Both Raman gain and the Kerr effect are
nonlinear processes that scale proportional to 1/ Aeff of the fibre,’ Pelouch explained. ‘Te power-dependent Kerr effect limits the optical transmit power such that the benefits of increasing OSNR with increasing transmit power are balanced versus the nonlinear penalties, which also increase with increasing transmit power. Tis is true for all transmission systems.’ Consequently, in a hybrid Raman/ EDFA design, the main trade-off is between fibre area and spectral efficiency, he said. ‘Standard 80μm2 Aeff fibres tend to be a good balance,’ Pelouch said. ‘Our hybrid Raman/ EDFA amplifiers are designed for a 25-year lifetime, which would not be possible if there was any significant risk of damage in our submarine line system. Te reason is that it is a controlled environment. Also, the Raman
power does not cause a Kerr effect for the signals, except under special circumstances of four-wave mixing, which is easily avoided by design.’
Prime the pump Adding an extra Raman pump, compared to the hybrid Raman-EDFA design, increases bandwidth and OSNR. ‘We can add one pump wavelength to scale the bandwidth within the same basic optical design, without the need for a parallel amplifier,’ said Pelouch. ‘In fact, we only need to add one pump and one optical filter to implement this, relative to the previous design. Terefore, there are savings in both
But now, the next generation expects even more bandwidth, so we have to go beyond EDFAs
optical component count and space, which makes it an efficient design per bandwidth.’ Paul Harper and his Aston University team
has started looking to exploit Raman amplification alone in a lumped, rather than distributed, format. Typically, discrete lumped Raman amplifiers use separate highly non- linear fibres as the gain medium, rather than the transmission fibre, to ensure adequate Raman scattering. ‘You reduce the pump power required,’ Harper explained. ‘Te downside of using highly non-linear fibre is that they have all these other non-linear effects,’ he observed. Tat imposes an approximately 10-15dB upper limit on the gain attainable.
Te Aston team therefore reduces nonlinear
penalties through a ‘dual stage’ approach. Te first stage consists of 10km of non-linear inverse dispersion fibre (IDF) and provides 14.5dB gain. A 10km singlemode fibre (SMF) second stage adds 5dB gain with minimum additional nonlinear penalty. Such an approach is unlikely to replace EDFAs in the C-band, Harper admits. Instead, it will likely target the O-band, E-band and S-band low-loss windows, from 1260nm-1360nm, 1360nm-1460nm and 1460nm-1530nm respectively.
Banding together Similarly, Antonio Napoli, a project leader at transmission technology system vendor, Coriant in Munich, Germany, said that his company doesn’t plan to replace EDFAs in the C-band. Coriant also intends to limit Raman amplification to where it’s strictly required. ‘Under certain conditions Raman improves system performance, but might be costly and require higher power than lumped amplifiers,’ Napoli explained. ‘Terefore its usage should be pondered case-by-case.’ Instead, Coriant and its partners are planning to re-design the EDFA in the L-band from 1565nm-1625nm bandwidth, extending the current upper limit of 1615nm by 10nm. However, the company is also entering the O-, E- and S-bands. ‘Coriant is actively working with several
industrial and academic partners on multi- band systems as a whole,’ Napoli told Fibre Systems. ‘As a first step, we have prepared a testbed with (C + L)-band systems, to be used, for example, to validate the first theoretical results over the (C + L)-band. Te main part of our future work will be to define the most cost-effective solutions for O, E and S-bands in terms of amplifiers, but also for what concerns other devices like filters and the transmission techniques.’ Napoli envisages that solutions
Issue 20 • Summer 2018 FIBRE SYSTEMS 15
Xtera
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44
