How DWDM systems work and why power matters


Every DWDM system has its own signal power characteristics at the multiplexer input. Over decades of designing carrier networks, a de facto standard has emerged: the multiplexer should receive a signal with a power level between -3 and 0 dBm. Systems are calibrated with this in mind, and passive components are selected under the assumption that the incoming signal has exactly this power.
Signal power at the multiplexer input
DWDM systemexpected range
−3 to 0
Typical 400G modulemost of the market
−10 dBm
GBC Photonicsnative, no EDFA
0 dBm
The difference between −10 dBm and 0 dBm is not a discrepancy within measurement tolerance. It is a tenfold difference in signal power.
10×
Most 400G coherent modules available on the market transmit a signal at approximately -10 dBm. It is worth understanding what this means in physical terms. The difference between -10 dBm and 0 dBm is not a matter of measurement tolerance, but a tenfold difference in signal power. A module transmitting at -10 dBm cannot be plugged directly into a system calibrated for 0 dBm, so an EDFA amplifier must be placed between the module and the multiplexer.

What happens when you add an EDFA as a patch


An EDFA amplifier between the module and the multiplexer will indeed equalize the power level. The problem is what you pay for it, and you pay in three areas at once.

An EDFA amplifier between the module and the multiplexer will equalise the power level. The question is what you pay for it.

01
A cost outside the original design

Buying and installing an amplifier that was never in the original quotation. Plus one more component in the path that can fail.

02
Higher power consumption

The amplifier needs power, so the energy consumption of the entire system rises, and across many links that item adds up.

03
Degraded OSNR along the path

The amplifier raises the signal level, but it also amplifies noise and adds its own spontaneous emission. The module enters the path with a worse OSNR than without it.

Some manufacturers build a miniature EDFA amplifier straight into the module. The module then transmits at 0 dBm, but at the cost of higher power consumption and a worse transmitter OSNR. An internal amplifier does not remove the OSNR problem, it simply moves it inside the housing.
The first cost is the purchase and installation of an amplifier that was not in the original design. The second is the higher power consumption of the entire system, as the EDFA requires power. The third and most serious is the degradation of the OSNR parameter across the entire link.

OSNR, or Optical Signal-to-Noise Ratio, is one of the most important parameters determining the range of a coherent system. Every active device in the path adds noise to the signal. An EDFA amplifier boosts the signal level, but it also amplifies noise and adds its own spontaneous emission. A module that would have entered the path with a good OSNR ends up with a worse one at the amplifier output than it would have had without it.

Some manufacturers try to bypass the power problem by building a miniature EDFA amplifier directly into the module. The module then indeed transmits at 0 dBm, but at the cost of higher power consumption and a worse transmitter OSNR than a module designed natively for 0 dBm. The internal amplifier does not remove the OSNR problem; it just moves it inside the housing.

How GBC Photonics solves this problem


The GBC Photonics 400G OpenZR+ 0 dBm module transmits a signal at 0 dBm without any EDFA amplifier, neither external nor internal. This is the result of a design decision made at the optoelectronic architecture level.
Optoelectronic architecture
InP technology, indium phosphide

Functions previously delivered by connecting discrete components are integrated into a single piece of semiconductor material. The amplifier set is designed to achieve maximum output signal quality without external amplification.

Tunable laser
Modulator
Receiver diode
Integrated amplifiers
0 dBmoutput power natively, with no external or internal amplifier
43 dBtransmitter OSNR, ample margin on longer routes
<22 Wpower draw of the whole module despite transmitting at 0 dBm
The transceiver circuit of this module consists of a tunable laser, a modulator, a receiver diode, and a set of amplifiers designed and built to achieve maximum output signal quality without external amplification. The entire unit is built using InP (indium phosphide) technology, which allows for the integration of functions previously performed by connecting discrete elements—namely the laser diode, optical amplifier, and modulator—into a single piece of semiconductor material.

The effects of this architecture are measurable. A transmitter OSNR of 43 dB provides a large margin for coherent transmission, especially on longer routes where every decibel of margin translates into link stability. The power consumption of the entire module below 22 W, despite transmitting at 0 dBm, is possible precisely because there is no internal amplifier straining the power budget.

What this means for practical deployment


The GBC Photonics 400G OpenZR+ 0 dBm module can be used in existing DWDM systems without any modifications, as most of them are calibrated for 0 dBm power at the multiplexer input, and the module meets this requirement natively. There are three practical consequences.
01
You do not need an EDFA amplifier between the module and the multiplexer

The deployment design is simpler, costs less and has fewer components that can fail. Upgrading an existing system requires no hardware beyond the original quotation.

Simpler design, lower cost, fewer points of failure
02
Transmit power is adjustable precisely across a 10 dB range

If a given route needs slightly lower input power at the multiplexer, you change the parameter over CMIS from the router or through the SRD environment. No hardware swap and no engineer on site.

Remote tuning, route by route
03
Better OSNR along the path from the outset

A signal entering the path with a good transmitter OSNR has more margin for degradation along the route. That translates into link stability in harder conditions, meaning more ROADM nodes, longer distances and older fibre.

More margin on difficult routes

Full tunability and Flex-Grid


Beyond output power, the GBC Photonics 400G OpenZR+ module has two features that are significant in the context of existing DWDM systems. It is fully tunable in the C-band, and setting the required transmission channel takes about 10 seconds. Field engineers do not need to order a module for a specific wavelength, as one type of module is kept in stock and configured for each node before installation.

The channel grid is also adjustable, meaning the module works in Flex-Grid systems. Flexible grid systems allow for allocating a channel width appropriate to the actual signal—that is, 75 GHz for 400G at 60 Gbaud, which is a requirement for OpenZR+ modules. If your DWDM system supports Flex-Grid, you have full flexibility in bandwidth management. If it has a fixed 50 GHz grid, the maximum throughput per channel will be 200G instead of 400G, and this must be taken into account when planning the project.

When output power is a deciding factor


For new buildouts, where the DWDM system is designed alongside coherent modules, output power is part of the design and can be accounted for from the start. A 0 dBm module is optimal in this case, but it is not the only option that will work.
New buildout
A DWDM system designed together with the modules

Output power is part of the design and can be accounted for from the start by selecting components around the chosen module. A 0 dBm module is optimal here, but it is not the only option that will work.

Important, but not decisive
Upgrading an existing network
A live DWDM system, IPoDWDM deployment without replacing infrastructure

Here you run straight into the power compatibility problem. The system is calibrated for 0 dBm and the module transmits at minus 10 dBm. Either you add an EDFA amplifier or you choose a module designed for 0 dBm from the start. For operators with networks built over the last fifteen years, this is often the argument that ends the comparison.

Decisive argument
The output power argument is strongest when upgrading existing systems. An operator with a functioning DWDM system who wants to deploy IPoDWDM without replacing infrastructure runs straight into the power compatibility problem. The system is tuned for 0 dBm, but the module transmits at -10 dBm, so you either add an EDFA amplifier or choose a module designed for 0 dBm from the start. For Polish and Central European operators who have extensive DWDM networks built over the last fifteen years and want to upgrade them without replacing the optical line system, this feature is often the deciding factor when comparing with the competition.

FAQ: coherent module output power

Because it decides whether the module drops straight into your DWDM system or needs additional hardware. Decades of carrier network design have produced a de facto standard under which the multiplexer should receive a signal between minus 3 and 0 dBm, and systems are calibrated accordingly. A module transmitting at minus 10 dBm does not meet that requirement, so an EDFA amplifier has to sit between it and the multiplexer. That single number holds the difference between a straightforward project and one you have to redesign.
Yes, and the scale is worth understanding. This is not a discrepancy within measurement tolerance but a tenfold difference in signal power, because the decibel scale is logarithmic. A DWDM system calibrated for a 0 dBm signal will not correctly accept a signal ten times weaker. The passive components in such a system were selected on the assumption of a specific input level, so this is not a matter of fine tuning but a fundamental mismatch between the module and the infrastructure.
The amplifier will equalise the power level, but you pay for it in three places. The first is the cost of buying and installing a device that was not in the original quotation, plus one more component in the path that can fail. The second is higher power consumption across the whole system, because the amplifier needs power, and across many links that item adds up. The third and most serious is degradation of the OSNR along the entire path, because the amplifier raises the signal level while also amplifying noise and adding its own spontaneous emission.
Partly, and that is the heart of the matter. A module with a miniature amplifier built into the housing does transmit at 0 dBm, so the power level problem disappears. The other two costs remain, however, namely higher power consumption and a worse transmitter OSNR than a module designed natively for 0 dBm. An internal amplifier does not remove the OSNR problem, it simply moves it inside the housing. When comparing offers it is therefore worth asking not only about the output power figure, but also about how the manufacturer achieves it.
Through a decision taken at the level of the optoelectronic architecture. The transceiver circuit consists of a tunable laser, a modulator, a receiver diode and a set of amplifiers designed to achieve maximum signal quality without external amplification. The whole unit is built in InP, indium phosphide, which allows functions previously delivered by connecting discrete components to be integrated into a single piece of semiconductor. The effects are measurable: a transmitter OSNR of 43 dB and a power draw below 22 W for the entire module, possible precisely because there is no internal amplifier loading the power budget.
Yes, transmit power is adjustable precisely across a 10 dB range. If a given route needs slightly lower input power at the multiplexer, you change the parameter over CMIS from the router or through the SRD environment, with no hardware swap and no engineer sent to site. The module is also fully tunable across the C-band, and setting the transmission channel takes about 10 seconds. In practice that means one module type in stock and a configuration matched to each node before installation.
Then the maximum throughput per channel will be 200G instead of 400G, and that has to be factored into project planning. OpenZR+ modules run at 60 Gbaud and require a channel at least 75 GHz wide, which a fixed 50 GHz grid does not provide. If your system supports Flex-Grid, you have full flexibility in bandwidth management, because channel width is allocated to match the actual signal. It is worth verifying the supported grid with your optical line system vendor before ordering modules, rather than after delivery.
When upgrading existing systems, and here the argument is strongest. An operator with a live DWDM system who wants to deploy IPoDWDM without replacing infrastructure runs straight into the power compatibility problem: a system calibrated for 0 dBm and a module transmitting at minus 10 dBm. There are then two ways forward, either adding an EDFA amplifier or choosing a module designed for 0 dBm from the start. On new buildouts, where the system is designed together with the modules, the question can be handled from the outset and a 0 dBm module is optimal, though not the only workable option.
Rozpocznij z nami nowy projekt!
Dziękujemy! Otrzymaliśmy Twoje zgłoszenie!
Ups! Wystąpił błąd podczas wysyłania formularza.

Alien wavelength - easily scale your network

Alien wavelength (alien lambda) is a technology with which you will increase the bandwidth of existing DWDM networks.
Czytaj więcej

Your fiber can carry 40 times more data. You probably haven't been told that.

CWDM, DWDM, and O-Band on a single fiber. Why now is the right time and how to get more out of the infrastructure you already have. Without laying new cables.
Czytaj więcej

Networking Trends 2024

Open networks, 800G or quantum encryption - what has the chance to change telecommunications and data center infrastructure in 2024? Check out the predictions of Salumanus experts.
Czytaj więcej