Erbium developments drive multiport amplifiers
Optical-network designers can solve power budget problems in restricted spaces by using multiport optical amplifiers. These erbium-doped waveguide-amplifier (EDWA) arrays of individually addressable optical amplifiers enable a lower cost per port than other amplification solutions.
In a similar way to EDFAs, EDWAs compensate for optical losses between network elements or within any subsystem, exhibiting the same low noise figure, high-bit-rate transparency, and low power consumption. But unlike EDFAs, EDWAs can stand alone or be integrated with other components to offer the user systems on a chip.
An EDWA consists of an active waveguide embedded in an amorphous erbium-doped glass substrate combined with passive functions. Erbium atoms provide the active glass substrate with optical gain in C- or L-band wavelengths. The combination of active and passive functions on the chip provides high levels of component integration in EDWA design (see Photo).
A typical multiport amplifier can comprise four individually addressable ports via pump control amplifiers on one planar-waveguide platform providing four separate optical outputs of 10 dBm. Benefiting from uncooled laser pumps, the amplifier exhibits a maximum power consumption of <4 W, which means it consumes <1 W per port.
Typically, a multiport-amplifier array module measures 95×55×12 mm, giving system designers the latitude to add new or reconfigured components in network designs. With increasing channel counts in arrays, further size reductions are attainable through EDWA technology—for example, chips with arrays of 32 functions (equivalent to an array of eight amplifiers) in a footprint the size of a postage stamp.
To reach the compelling price per port that multiport amplifiers can offer requires a technology that enables high functionality and a manufacturing process that is low-cost. EDWA technology, based on an ion exchange process, manufactures amplifier modules economically due to low front-end investment, an efficient two-step process, and semi-automated production.
There are various applications for EDWA arrays in 10-Gbit/sec metro, metro core, and long-haul (LH) as well as 40-Gbit/sec metro and LH markets where single-channel and narrowband applications are key to achieving flexible bandwidth management. These arrays can be used in transmitter, receiver, and inline amplifications. At the transmitter end, the array can boost and equalise the signal of a set of sources such as tunable lasers, offering tunability and flexibility to system designers. Due to its low noise figure inherent in erbium physics, the EDWA array can readily be used as single-channel or narrowband preamplifier at the receiver end. Inline applications may equally be possible areas at nodes where more than two amplifiers are used.
The multiplexing and demultiplexing of WDM signals also present opportunities for EDWA arrays. For example, an array can be used along with an arrayed-waveguide grating to provide gain plus channel balancing. The pump control would enable the output power of each port to dynamically and individually change. The EDWA array would then be an alternative to arrays of variable optical attenuators.
EDWA arrays provide a compact solution to compensate for losses in optical-switch fabrics. Applying the EDWA array here would enable the channel balancing to improve the port-to-port uniformity at the output of the optical-switch fabrics.
Another interesting application is boosting the sensitivity of PIN receivers. Arrays of amplifiers, for instance, can be used in transparent LH optical add/drop multiplexers (OADMs) as single-channel preamplifiers to boost receiver sensitivity (see Figure 1).
The most important application of EDWA arrays, however, takes place in the metropolitan area where higher port count and signal-routine flexibility are key issues (see Figure 2). Here, the arrays are right on target and hit the need to substantially save space in future OADMs and reconfigurable OADMs. These arrays also enable system designers to take advantage of the single-channel or narrowband approach (see Figure 3). An array of amplifiers can be used before the multiplexer in an optical crossconnect (OXC).
Each port of the array supports a set of channels—typically four to eight channels over a 4-nm window. Consider a four-port array where each port supports eight channels. In this case, each port is a narrowband amplifier. If we assume that each port achieves 15-dBm output power, the power per channel will be around 6 dBm.
Another illustration of the competitiveness of EDWA arrays consists of a single-port amplifier placed after the multiplexer at the output of the OXC. The single-port amplifier has to support 32 channels over the entire C-band (4×8 channels). Here, the amplifier is a "wideband" amplifier. Assuming that the single-port amplifier achieves 17-dBm output power per port, the optical power per channel will be about 0–2 dBm.
The single-channel or narrowband approach can accomplish a number of advantages versus the wideband approach. It would ease the design of the amplifier, remove the need for a gain flattening filter (GFF), avoid transient issues encountered in an amplifier loaded by a wideband signal, simplify the control of amplifiers, and thus free designers from technical issues in the network design.
In addition, the location of the multiport amplifier would improve the signal-to-noise ratio when the amplifier is placed before the insertion losses in the optical path, such as before the multiplexer. Also, the optical budget per channel would be higher for the multiport amplifiers, because each port would support a lower number of channels.
The single-channel and narrowband approach offers system designers the ability to dynamically populate the system. This "pay as you grow" concept is what operators are looking for to reduce capital expenditures and improve cash flow.
Antoine Kevorkian is chief executive and sales/marketing vice president and Benoit Neyret is an applications engineer at Teem Photonics (Meylan, France).