Tech Briefs

Wavelength Translation Using SFP Transceivers Enables Extension of Optical-Communication Links

Optical-Electrical-Optical (OEO) wavelength translation (WT) based on smallform- factor pluggable (SFP) transceivers is meeting with increased interest in metro optical-network applications. It provides a flexible and low-cost solution when it interfaces with legacy optical equipment that does not support International Telecommunications Union (ITU) wavelengths. WT improves the performance of optical systems by changing the operating wavelength of the incoming optical signal to a wavelength that enables longer reach through lower attenuation or lower dispersion penalty.

alt Besides being protocol independent, WT presents an opportunity to deploy an optical source with a narrower linewidth, extending the optical link through better dispersion management. This is especially useful for higher-speed systems, such as OC-48 running on 1300- nm systems, requiring link extension. Data-com signals at 850 nm running on multimode fiber can profit also from WT that allows the system to run on single- mode fiber at 1550 nm, eliminating multimode dispersion.

The primary component that has made WT possible is the SFP transceiver consisting of a transmitter, receiver, and microcontroller. The transmitter module consists of a laser and laser-drive circuitry that may include a thermoelectric cooler (TEC). The laser could be: an uncooled, single-mode, distributed-feedback (DFB) laser with its center wavelength anywhere in the range of 1480 nm to 1580 nm (typically at 1550 nm); a coarse wavelengthdivision multiplexing (CWDM), ITUgrid wavelength, cooled DFB laser at a wavelength whose peak is centered at 100-GHz spacing on the ITU-grid; a 1310-nm multimode Fabry-Perot (FP) laser; or a multimode 850-nm vertical cavity surface emitting laser (VCSEL).

The receiver is a PIN- or APD-based module, depending on the link budget, with a trans-impedance amplifier (TIA) and limiting post amplifier. The attainable link reach after WT is a function of both transmitter and receiver and could be short (up to 10 km), intermediate (up to 60 km), long (up to 100 km), or extended-reach (up to 160 km).

Today, most SFP transceivers comply with a multi-source agreement (MSA) signed by many vendors to guarantee compliance and interchangeability of modules. These transceivers offer hotswappable modules that utilize single 3.3-V power sources to minimize power consumption.

alt The SFP transceiver receives a non-return- to-zero (NRZ)-encoded optical signal and converts it into a low-noise CML (current mode logic) or LVPECL (lowvoltage positive referenced emitter-coupled logic) compatible electrical signal. The transmitter is compatible with both CML or LVPECL input data levels. SFP transceivers are incorporated with a digital diagnostic feature to report transmitter and receiver status. They can be used for data rates from 50 MHz to 2.7 GHz and support SONET, Gigabit Ethernet, and Fiber Channel in addition to datacom products. A datacom SFP transceiver has to satisfy IEEE standard (Gigabit Ethernet 802.3) or American National Standards Institute’s (ANSI) Fiber Channel, FC-PI specifications. SONET protocol SFP transceivers must satisfy Telcordia qualification requirements.

Figure 1 shows a WT block diagram used to convert an 850-nm TDM signal to a 1550-nm signal for transmission to a remote site. The 850-nm signal could be a 1 Gb/s data signal on 200 MHz-km, multimode fiber that needs to be sent to another network site kilometers away. The conversion to 850 nm can be to a TDM signal at 1550 nm, a CWDM wavelength or a DWDM (dense wavelength-division multiplexing) wavelength on the ITUgrid. The choice of system is a matter of distance between sites, bit rate, and budgetary consideration.