Pluggable coherent optics offer advantages in many areas. One of the biggest factors is electronic dispersion compensation. For DCI, traditional ‘direct detect’ systems needed a detailed pre-planned link with dispersion compensation ‘engineered in’. This is both expensive and challenging to deploy. Scaling and maintaining are complicated. Electronic dispersion compensation inside the DSP of coherent modules can correct for both Chromatic dispersion (CD) and Polarization Mode Dispersion (PMD) in the digital domain. Furthermore, this can be done dynamically and in real time.  Limited link planning is required.

An obvious complicating factor is the polarization of the transmitted optical signal, and the way it behaves and is modified by the optical fibre. As a method to double the data rate, coherent transmitters multiplex two signals at the same wavelength but orthogonal polarization states.  Before the receiver can apply the electronic compensation algorithm, it must demultiplex these two signals. However, the state of polarization will be constantly evolving due a small residual birefringence in all single mode fibre (birefringence is a difference in the index of refraction due to the state of polarization  and the orientation of the material through which the light propagates).  The birefringence of the fibre in the link can and will change.  As it changes, so does the state of polarization (SOP) at the receiver.  A coherent receiver must be able to deal with this – as mentioned above – in real time.

There are many ways the birefringence can change, and we can divide them into a few categories based on the rate of change. There are slow changes, like temperature, which cause expansion and contraction.  These changes happen over seconds and minutes. There can also be faster changes due to small vibrations.  These vibrations cause changes to the stress and strain along the fibre (for example: an aerial cable swaying in the wind or vibrations as a truck goes buy). These can happen with frequencies in the Hz or KHz.  Finally, there are rare, but very fast changes due to nonlinear electro-optic events, like a lightning strike, and these disturbances can occur extremely quickly in the MHz range.

 

 

 

 

 

 

While you often hear about changes in SOP in krad/s or Mrad/s, when you consider the different types of changes as they accumulate over 100’s and even 1000’s of km, it is more helpful to think about the evolution of the signal SOP in a statistical way.

There are different schools of thought here. Many consider this complexity is best described by a pure random distribution – the current SOP can change to any other SOP with equal probability. Others feel the SOP distribution is best modelled with a Rayleigh distribution where there is a much higher likelihood of smaller changes and then more occasional larger changes in state. In both cases, if the observer waits long enough, all states are eventually covered.  Test system that allow designers to validate performance in both cases is critical.

Whatever type of situation, the DSP module must then deal with a dynamic environment. First, to demultiplex these rapidly changing polarized signals and then apply compensation for multiple types of dispersions. Integrators and end users need to validate real live performance of this dispersion compensation and validate the OSNR penalty imposed by state of polarization changes. Such measurements can be extremely challenging without the right experimental set up and understanding of how to do ‘bad’ things to your module.

The VIAVI lab and production team has many years of experience in development, test and validation of coherent systems.  A combination of our ONT 800G DCO module and a photonic set up based around our MAP family–specifically the mPCX–provides a comprehensive solution to validate, test and stress dispersion compensation.

Figure 1: The MAP-300 mPCX-C1 module is used to emulate the change in the state of polarization change in an optical fiber.  The simple user interface allows engineers to modify the statistical nature of the polarization change quickly and simply, both in terms of peak velocities, but also statistical distribution and state coverage.

In addition to pure scrambling, specific stress tests can also be imagined.  One example is to manipulate the state of polarization into ‘aggressive’ states which could mean high optical powers going to one set of photodetectors with very low optical power on the other (polarization state) set. This tests the module dynamic range, including the photonics, the AGC and the DSP control loops. The MAP system can then be set to rapidly jump between bad states and validate coherent module stability under different statistical distributions of SOP change. The key point to remember is that instrument polarization controllers must be able switch between these modes simply and quickly.

Such testing is critical. It tests many aspects of the DCO under quasi static stressful states AND dynamic conditions. It also allows the DSP firmware to be stressed in different ways to ensure all the important control loops are stable while under real operating conditions. With high volumes of pluggable coherent optics expecting to be deployed in the near future end users will want to unsure that they really are ‘low touch’ and their high density 400G DCI networks can be deployed and scaled under all real operating conditions.

If you still have questions about coherent testing, we at VIAVI are happy to share our expertise from many years of coherent system testing with you. Feel free to connect with Paul Brooks or Matt Adams on LinkedIn if you have any questions or would like to learn more. Or, if you would like to read more from these two, check out the following blogs:

What’s All This OSNR Stuff?

What’s All This Throughput Production Stuff?

 

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

Do NOT follow this link or you will be banned from the site!
Close