Optical communications are becoming even more present in the present world. The constant search for higher bitrates and higher bandwidth is needed since the users’ number is constantly growing. This dissertation approaches the technology for Passive Optical Network NG-PON2. This technology allows the communication in different channels in both downlink and uplink. In my dissertation, it is tested and characterized an optical component that sends and receives data at an extended reach. The data rate used is 10 Gbps in downlink and uplink. So, a prototype of an OLT produced by PICadvanced S.A. was tested in different scenarios. Two OLT are tested in the downlink: a commercial and a non-commercial. Both have an Externally Modulated Laser (EML), allowing a comparison of the performance. Considering the test conditions and the setup don’t change, the differences registered are related to the components used: laser, modulator and SOA. Regarding the upstream, the transmission of information is realized by an ONU, which has a Directly Modulated Laser (DML). Considering that, the test conditions remain the same for both XFP, the differences registered are related to the receiver, which is an APD in both cases. The performance tests realized include the impact of fiber dispersion in downstream for distances up to 40 km and upstream for distances up to 20 km. The fiber dispersion impacts more in transmitters that have a directly modulated laser. The impact of the temperature in both scenarios was analyzed yet considering that the transmitter has a TEC to control the temperature and the receiver does not. Considering an ONU emits data in bursts, the uplink communication was tested with 2 ONU emitting information that an OLT receives and it was possible to conclude that the time slot variance does not impact in the performance. In addition, a simulation tool in VPIPhotonics was developed to allow the optimization of different scenarios tested in the downlink considering the results of the non-commercial OLT.

In recent years, due to the growing demand for faster and better communication traffic, the continued improvement of our network infrastructure has become crucial. With the development of technologies like 5G, high quality video streaming, virtual reality and the growing number of network users, our access networks will have to be able to adapt to the data rate imposed by these technologies. Due to their advantages regarding transmission speed, efficiency and cost effectiveness, most of these access networks are now implemented with passive optical networks (PONs) which, at present, are led by the NG-PON2 architecture. This PON technology makes use of optical transceivers, devices capable of transmitting and receiving information at 10 Gbps within these networks, that operate using four different channel wavelengths. Needless to say, these transceivers are subjected to intensive rounds of tests and measurements in order to be approved for commercial use. In this process, several metrics are utilised to ascertain their quality, one of which being the extinction ratio, a parameter pertaining to the effectiveness of their modulating capabilities. This body of work will be describing the steps taken during the development of a method with which to estimate this metric, expediting its measurement process. This research, conducted in collaboration with PICadvanced, ultimately seeks to increase the production yield of these transceivers in PICadvanced’s manufacturing facility, meaning, the speed with which these are approved. In this effort, this work will also touch on the optimised measurement setup devised for this feat, followed by the transmission algorithm developed that allows it. Throughout this research, more then 3260 extinction ratios measurements were performed, across three tests, in order to achieve two reliable extinction ratio estimation methods. During these three tests, three different photodetectors were utilised to rule out possible measurement discrepancies between them. Finally, a summarised economic projection of the impact of the developed method was made, showcasing its future prospects in a factory setting.

Optical communications are widely used in our daily lives for data transmission and offer high quality and efficiency. However, it is necessary to constantly evolve the technology and infrastructure to meet the needs and expectations of consumers, such as high definition videos, virtual reality, and 5G mobile networks that require higher transmission rates and wider bandwidths. This presents new challenges for PON (Passive Optical Network) based technologies, which are currently the most commonly used in access networks, with advantages of efficiency, transmission rate and implementation cost. The integration of optical components on a single chip (Integrated Optical Circuits or PIC) is essential for the evolution of PON technology as it allows for reduction of energy consumption and complexity. This document addresses the development of new data transmission standards, namely those that present rates of 50 Gbps. Additionally, the implementation of 50 Gbps systems was studied, including simulations of different modulation and transmission technologies, such as EML (Externally Modulated Laser) and DML (Directly Modulated Laser), using O-band wavelength. The objective is to analyze the advantages and limitations of each technology. Additionally, experimental tests and simulations with 25G-PON technology were performed to characterize and evaluate the performance of SFP28 transceivers. Finally, the design and implementation of a PIC architecture with 50Gbps support was also developed, using an EML.

Este trabalho foi financiado pelo FEDER - COMPETE 2020 no âmbito do projeto plugPON (ref. POCI-01-0247-FEDER-047221). Devido à grande procura de ritmos de transmissão cada vez mais elevados, as tecnologias baseadas em arquiteturas PON (Passive Optical Network) sofreram um enorme desenvolvimento nos últimos anos e são atualmente as mais utilizadas nas redes de acesso. Com o aparecimento da quinta geração de redes móveis (5G), surgiram novos desafios para as redes PON, tais como: aumento do ritmo de transmissão, minimização do custo e coexistência com tecnologias anteriores. Neste âmbito, foi desenvolvida uma nova norma, o 25G-PON. No seguimento, esta dissertação foca-se na caracterização e otimização de transceivers a 25 Gbit/s, que utilizam diferentes tipos de modulação, nomeadamente, EML (Externally Modulated Laser) e DML (Directly Modulated Laser). Assim, pretendeu-se entender os limites e as vantagens de cada uma das opções. Adicionalmente, foram realizados testes experimentais em NG-PON2 e em 25G-PON, com o objetivo de comparar a performance de uma tecnologia que já se encontra em funcionamento (como é o caso de NG-PON2) com uma tecnologia que ainda se encontra na sua fase de testes (como é o caso de 25G-PON). Este trabalho demonstra a viabilidade de transmissores que fazem uso de modulação direta, que com um custo de produção bastante inferior conseguem atingir os valores requeridos pela norma. Foi também possível demonstrar a viabilidade do uso de comprimentos de onda da banda O (banda onde ocorre a transmissão downstream e upstream em 25G-PON). Também se caracterizou experimentalmente, com sucesso, dois transceiveirs do tipo XFP (10 Gigabit Small Form Factor Pluggable) em NG-PON2 e dois do tipo SFP28 (Small Form Factor Pluggable 28) em 25G-PON. Adicionalmente foram caracterizados e optimizados vários ROSAs (Receiver Optical Sub Assembly), todos eles com um recetor do tipo APD (Avalanche Photo-Diode) em sua constituição.

The concept of connectivity in the dictionary is defined as the ability of a computer or an information system to connect to other devices. This concept is constantly evolving because the devices that require connectivity are becoming increasingly diverse, and information systems are more complex. This increase in connectivity leads to an increase on the amount of transferred information and, consequently, the need for the evolution of telecommunications infrastructure, particularly those that reach our homes, such as passive optical networks. These networks also have their own evolution plan through the standards defined by standardization bodies which have started in the early 2000s with GPON, implemented worldwide today, bringing downstream speeds of 2.5 Gbps and upstream speeds of 1.25 Gbps. Today, we are witnessing the proliferation of technologies that already support 10 Gbps symmetrically, with the first tests of networks capable of delivering 50 Gbps symmetric speeds underway. The evolution of passive optical networks is only possible with the evolution of the optoelectronic components that implement them, and it is in the physical layer that challenges increase with the increase in bandwidth per device or the need for tuneability, as is the case of NG-PON2 standard. In parallel, with the evolution of passive optical networks in recent years, research and industry surrounding photonic integrated circuits have also made significant strides, allowing many of the devices around us to already have simple photonic integrated circuits. In data centers, where the need for bandwidth is much higher, most transceivers are already based on photonic integrated circuits and they bring all the advantages that integration has brought to the electronics domain, such as scalability at production and technological level, price or power consumption. Hence, the hypothesis arises that photonic integrated circuits can be part of the solution for the physical layer of 10 Gbps and beyond passive optical networks. Starting from this point, this thesis begins by presenting the evolution and hardware requirements of passive optical networks, as well as the evolution of photonic integrated circuits and their maturity for this market. This is followed by a theoretical study of the blocks of a transceiver for passive optical networks based on photonic integrated circuits, moving on to its design and testing, presenting experimental results of transceiver prototypes based on photonic integrated circuits. Finally, a techno-economic study and price sensitivity analysis are presented to ensure the feasibility of the proposed solution, along with a discussion of future market possibilities for integrated optical circuits in passive optical networks.

A evolução das redes óticas tem conduzido ao desenvolvimento de novas tecnologias de maior débito, motivado pelo aumento do volume de dados ao longo dos anos. As redes óticas passivas (PON: Passive Optical Networks) são uma das soluções para responder às exigências das atuais e futuras infraestruturas de comunicação. Neste contexto, a tecnologia 25G-PON é uma das mais recentes normas, que visa aumentar a capacidade de transmissão e mantendo a compatibilidade com gerações anteriores. O presente trabalho tem como objetivo o estudo e caracterização de acopladores contra-direcionais (CDC) integrados em sistemas 25G-PON, recorrendo a medições experimentais e simulações realizadas com o simulador Ansys Lumerical INTERCONNECT. Foram analisados parâmetros fundamentais como a perda de inserção (IL), Razão de Rejeição (RR) e correspondente penalidade de potência (PP) quando se considera mais que uma tecnologia a coexistir. Foram desenvolvidos sistemas a 25 Gbit/s, obtendo-se as curvas de Bit Error Rate (BER) em função da Received Optical Power (ROP) em Back to Back (B2B), com uma fibra de 20 km, com um CDC com apenas uma tecnologia, e por fim para um setup com duas tecnologias, que nos permite analisar a diafonia inter-canal. Os resultados obtidos contribuem para o estudo de CDCs em sistemas de 25G-PON, em ambas as direções, Downstream (DS) e Upstream (US).

Neste documento é apresentado um estudo sobre uma tecnologia utilizada em sistemas de comunicação óticas, que se encontra actualmente em fase de desenvolvimento por parte da Google, nomeadamente, a tecnologia SUPERPON. São apresentados alguns conceitos teóricos essenciais à compreensão da tecnologia em si e das tecnologias que a precederam, tais como XG-PON, XGS-PON, NG-PON2 e Google Fiber. Posteriormente é feita uma análise dos principais componentes necessários à implementação de um sistema de comunicações por fibra ótica. Foram ainda feitas simulações, utilizando modelos baseados em componentes reais, com vista a perceber a diferença entre os vários tipos de transmissores e recetores. São ainda apresentados testes práticos que serviram para verificar a viabilidade de implementação da tecnologia SUPER-PON. Concluiu-se que para ser possível alcançar alguns dos objetivos conhecidos até ao momento, nomeadamente, alcance de 50 km na fibra com 10 Gbit/s por canal, é necessário o uso, em donwstream, de um transmissor modulado externamente por ter menos chirp, e um recetor do tipo avalanche por conseguir detetar potências mais baixas. Nesta situação não foi necessário qualquer tipo de compensação de dispersão, devido ao tipo de modulação usada. Já para a direção upstream, pode-se usar um LASER modulado diretamente, por ser mais barato para o utilizador final, e um recetor do tipo avalanche, contudo, nesta situação é necessário a utilização de algum tipo de compensação de dispersão. No contexto desta dissertação efetuou-se a compensação usando fibras de compensação de dispersão. Além disso, validou-se ainda que a filtragem de um canal com uma elevada taxa de rejeição dos canais adjacentes, permite a utilização de vários canais com interferência mínima. Esta dissertação foi desenvolvida em parceria com a empresa PICAdvanced.

The study of new technologies with the potential to support high transfer rates is important to respond to the constant evolution of Internet information consumption. The Next-Generation Passive Optical Networks (NG-PON) tries to overcome this obstacle. The most recent one, NG-PON2, uses several channels in the same fiber to transmit and receive data, considerably increasing the bandwidth of a single fiber. This new form of communication has introduced a new problem, the exchange of information between these channels with a minimum of disturbances on the network. The Inter-Channel-Termination Protocol (ICTP) was presented to certify that this communication is done correctly in a multi-operator environment. Based on a pre-built environment with the cooperative use of Gigabit PON (GPON) and NG-PON2 technologies, a system with similar characteristics to those of a real scenario was developed. In this system, two ICTP use cases were implemented and, for each one, two events were tested. The results obtained showed the possibility of being used in a real scenario, without failures and with small delays.

For the last two decades, the Internet has been target of a technological revolution. The amount of new services and applications that came up with this revolution demanded the search to obtain more bandwidth and bigger data rates. The only technologies capable of satisfying these conditions are the optical communications. The current technology that is capable to satisfying all these needs is the Passive Optical Network(PON). There are some variations of this technology currently being used such as the GPON and EPON. New evolutions of PON are being developed, such as the Next Generation Passive Optical Network 2(NG-PON2) This technology multiplexes different services into separated wavelengths, making its performance more effective. One tool of great matter in this business is the Photonics Integrated Circuit(PIC). The photonic integrated technology has been taking over the conventional optical equipment because it allows the integration of several functionalities into one single chip, saving space and energy. Having in mind the objective of developping next generation optical communication systems, the work presented in this document intends to test PIC in way that they can be used in optical equipment, as for example the Optical Line Terminal(OLT) which has functions of transmitter/receiver situated at the operator's station, and the Optical Network Unit(ONU) situated at the user's location. To make it possible for PIC to be integrated in optical networks these must be tested in order to assure that they perform their functions without any anomalies happening. For such reason, these PIC are at our labs, read to be tested. This work, consisted of the execution of tests to PIC as well as the optimization of the techniques and adequate procedures. In the first place, it is presented an overview of optical networks and PIC. Then the do's and dont's of testing PIC as well as the procedure and preparations of these tests. At last, the tests to the PIC will be performed which represents most of the work shown in this document.

Abstract In this work, we introduce a new concept of frequency interleaving applied to passive optical networks based in Dense Wavelength Division Multiplexing (DWDM). In order to explore colorless ONUs approach, a single and centralized optical source is used to create the interleaved wavelengths for the downstream and upstream path. In this case, frequency comb generation implemented by recirculating frequency shifting technique allowed the creation of all the wavelengths used in each path. The reception operation is accomplished by introducing, at the ONU side, a combination of a set of optical passive elements. The system with 4 × 12.5 Gbit/s NRZ-OOK (Not Return to Zero - On-Off Keying) modulation format were simulated through 20 km of standard single mode fiber link in both transmission paths. In order to guarantee a high quality optical frequency comb source, the optical carrier-to-noise-ratio (OCNR) was evaluated as well as BER (bit error rate), to analyze the system performance. All devices models and simulations were performed in Matlab environment.