The need to expand reach in metro networking equipment designs has never been more important. With service providers trying to deliver more services with existing equipment, designers are being forced to expand the reach capabilities of a networking box design.
At the same time, design engineers are being tasked with adding new levels of tunability to metro architectures. Carriers want to be able to tune their optical equipment so that a host of services can be handled on the same optical pipe.
Until now, OC-48 transmitters trying to meet the reach requirements of today's metro carriers were forced to employ expensive, large, external modulators with high-power CW lasers. A new transmitter architecture, which merges a vertical cavity surface emitting laser (VCSEL) with an electro absorptive (EA) modulator, has emerged that can provide tunability while achieving distances up to 600 km.
In this article, we'll take a detailed look at the VCSEL/EA modulator transmitter architecture. During the discussion, we'll compare this new architecture against traditional OC-48 transmitter designs. We'll also look at the packaging requirements for making the VCSEL/EA modulator transmitter come to life. Let's start the discussion with a look at the impact of loss and dispersion on an optical transmitter design.
There are several factors that limit the distance achieved by an optical transmitter. Two of the most troublesome common include:
- Optical Loss: Optical loss results in insufficient signal power available at the receiver.
- Dispersion: Dispersion results in spreading of the signal pulses at the receiver and therefore errors in resolving the transmitted signal.
Transmission distance refers to the overall distance data is expected to travel on an optical carrier. It represents the physical distance (typically denoted in meters or kilometers) between the electrically modulated optical transmitter (E/O) and the optical-to-electrical conversion of the signal at the optical receiver (O/E). Between the transmitter and the receiver, the loss and dispersion properties of the media may degrade the transmission performance.
The optical link budget represents the net allowable loss, from all sources including gain from amplification, through the link from transmitter to receiver. The amount of power required by the receiver for adequate performance is characterized by the receiver sensitivity, and the transmitter, naturally, determines the amount of transmitted optical power. Optical loss in the link (from fiber, passive components, etc.) is often offset with optical amplifiers that boost carrier signals to ensure that the power is sufficient to meet the minimum receiver sensitivity (Figure 1).
Figure 1: Loss element effects in a network link.
Dispersion refers to the fact that different wavelengths of light travel at slightly different velocities in an optical medium such as fiber. Because a pulse of light consists of a spread of optical wavelengths, this pulse will be broadened as it propagates, resulting in a spreading of the pulse at the receiver. This spreading increases the probability that the receiver will have errors in reading the received pulses correctly as 1s or 0s, commonly called bit errors.
A common way to characterize this probability of errors is the dispersion power penalty. This penalty is roughly an indication of the added power level required at the receiver to maintain a given bit error ratio (BER) in the presence of signal dispersion, compared to the back-to-back (0 km link) case (Figure 2).
Dispersion element effects in a network link.
Designers must meet the minimum requirements for both optical power and dispersion power penalty at the receiver in order for a networking design to be viable. The distance from the tunable optical transmitter to the optical receiver is limited by optical loss and dispersion.
Coping with Loss
It is fairly straightforward to deal with loss in the network. The three methods to dealing with link loss are increasing tunable transmitter power, increasing receiver sensitivity, and providing optical signal amplification. To determine what method to use, designers should turn to the link loss budget, which dictates the requirements of both transmitter power and amplification for a given receiver sensitivity.
Since amplifiers must be spaced sufficiently close together in order to achieve the desired link loss budget, amplifiers can have a negative impact on the cost of a network. As more and more passive elements increase link loss, particularly in tunable metro networks where flexible nodes include optical add/drop multiplexers and cross connect switches, the need for optical amplification is increased.
Fortunately, amplifier performance has increased while amplifier cost has continued to decrease. This makes it easier for network equipment manufacturers to develop flexible, cost-effective metro networks using optical amplification to manage the link loss budget.
Coping with Dispersion
Managing the impact of dispersion is more difficult than dealing with loss because fewer cost-effective alternatives are available. Though regeneration (the optical to electrical to optical conversion with signal conditioning) and dispersion compensation are technical approaches to coping with the ill effects of dispersion, the simplest way to deal with dispersion is to avoid it as much as possible by using transmitters which minimize dispersion.
Typically, the more pure the launch signal is in frequency, the less that the pulses will be broadened by dispersion in the network. A transmitter with a narrow continuous-wave linewidth and a small amount of chirp (frequency broadening due to modulation) will reduce the effects of dispersion and increases transmit distance. Until recently, however, the high cost of low dispersion modulation made it impractical to simply employ low-dispersion tunable transmitters in every network.
Alternatively, it is possible to reduce the impact of dispersion by introducing dispersion compensation in the transmission path. This can be done with specialty fibers or network elements (dispersion compensating modules) that generate negative dispersion, offsetting the dispersion created by the transmit signal.
The challenge with these dispersion compensating elements is that they tend to be very network-specific, meaning that the fiber being deployed or the dispersion compensating elements employed should ideally be matched to the tunable transmitters and specific network length in order to be effectively used. Additionally, dispersion compensating elements and fiber tend to be more costly, adding expense to the network. Finally, dispersion compensation elements add additional optical loss that must be compensated by amplification.
A New Alternative Emerges
Traditionally, finding an OC-48 transmitter that strikes an optimal balance between loss and dispersion has been next to impossible. But, things are starting to change. Recently, developers have begun exploring transmitter architectures that merge tunable VCSEL laser and EA modulator technologies. By doing this, designers have produced tunable OC-48 transmitters that can achieve the increasing metro and regional distance requirements, allowing non-regenerated transmission to over 600 km.
To achieve this feat, the proposed VCSEL/EA modulator transmitter supports a wide tuning range. Channel-to-channel output power flatness is achieved through internal gain control of the elements. Wavelength locking, to any ITU channel, is controlled and maintained with internal control circuitry and wavelength locker.
Alternative approaches to achieving similar performance exist, but none can meet all of the performance and cost requirements of metro and regional networking equipment designers. Let's examine these alternate approaches and then see how they stack up against the tunable VCSEL/EA transmitter architecture.
Due to their relatively low cost, thermally tuned directly modulated distributed feedback (DFB) lasers have been a widely used source for narrowly tuned OC-48 transmitters. Typical output powers of +3 dBm, or greater, provide plenty of fiber-launch power. However, directly modulated DFBs fail to deliver the reach needed in metro architectures because dispersion is induced in these transmitter designs from laser chirp. As such, thermally tuned directly modulated DFB lasers are typically limited to about 170 km in networks using standard fiber.
Even before the dispersion-induced limit of 170 km is achieved, a directly modulated DFB laser with +3 dBm of output power would still require at least one stage of amplification to offset the effects of power degradation in the link.
To solve the problems created by directly modulated DFB lasers, designers can turn to a transmitter architecture that blends a Mach-Zehnder lithium niobate (LiNbO3) modulator constructed with a high-power (+7 dBm or more) tunable continuous-wave (CW) laser. The high output power from the laser is necessary to compensate for optical loss through the modulator while still allowing adequate transmitter launch power.
The output power and spectral characteristics of the Mach-Zehnder/CW laser combo provides outstanding dispersion performance, significantly increasing reach performance (again with the addition of amplification stages to offset power degradation through the link) at OC-48, and higher data rates. However, the high-cost plus increased size and power consumption of this tunable transmitter does not make it practical in metro and regional networks, relegating this architecture to long-haul networks and higher data rate networking designs.
Combining EA modulators and VCSELs
An attractive cost-effective alternative to a tunable directly modulated DFB or CW laser plus LiNbO3 modulator is an electro-absorptive modulated laser (EML) transmitter including, in a single butterfly assembly, a laser and very low-chirp EA modulator. For OC-48, the most prevalent data rate used in metro and regional networks, the fixed wavelength EML approach has proven to be a cost effective alternative with adequate launch power and low dispersion characteristics to provide the necessary reach performance.
However, because of the EA modulator's limited wavelength range and suitable cost-effective packaging, until recently there were no widely tunable EML transmitter solutions available that could meet the longer reach requirements of regional networks while addressing the extreme price sensitivity of metro networks.
Today, this challenge has been met by the recent introduction of an integrated OC-48 tunable transmitter producing 1 mW (0 dBm) of average optical (modulated) output power with a dispersion power penalty significantly less than 2 dB at 600 km in systems using standard fiber, on any ITU channel.
The success of this tunable EML transmitter approach stems from the cost-effective integration of tunable VCSEL technology and advanced packaging capabilities. Let's look at both below.
Making Long-Wavelength VCSELs a Reality
The significant cost advantage of VCSEL lasers has been recognized at shorter wavelengths where VCSELs dominate. However, until recently, long-wavength (1550 nm and greater) VCSELs have not been readily available.
By adding a tuning structure to the VCSEL during the growth process, a low-cost, wide-tuning, monolithic tunable laser is produced. The VCSEL has the inherent advantage over edge-emitting devices that it can be tested at the wafer level, and therefore processed and packaged with extremely high yield and low cost.
Cost-effective wavelength locking is achieved with a small, low-cost universal wavelength locker that continuously monitors the transmit carrier signal and compares it's frequency to a known reference frequency at the specified ITU wavelength.
The EA modulator provides both optical gain and dispersion management. By integrating both functions in one device, the modulator enables the transmitter to achieve a launch power of greater than 0 dBm with significantly low dispersion to allow transmission, over a wide tuning range, in excess of 600 km (with amplifier stages).
One of the keys to making the VCSEL/EA modulator combo come to life lies in packaging. In this approach, yielded discrete components (low-cost tunable laser, locker and modulator with gain) are integrated within a single butterfly package assembly. This package enables batch processing as well as fast, reliable positioning and attachment. Thus, the complete assembly can be produced and tested at a single station in a very short amount of time with extremely high yield.
By including all supporting control electronics within the small transmitter package, the complete tunable transmitter conforms to the footprint of the fixed, directly modulated DFB transmitter MSA. This makes it an easy replacement for the non-tunable, shorter-reach predecessor or any of the other, larger and more-costly tunable alternatives.
About the Author
Wupen Yuen is the co-founder and CTO of Bandwidth9. He holds Ph.D. and M.S. degrees in electrical engineering from Stanford University, and a B.S. degree in electrical engineering from National Taiwan University. Wupen can be reached at firstname.lastname@example.org.