Application Note 259

In today’s telecom industry, systems are becoming increasingly complex due to the use of reconfigurable optical add/drop multiplexers (ROADMs) and/or multiple-bit/symbol advanced modulation formats (as is the case with most 40 Gbit/s and 100 Gbit/s transmissions). Complex systems often mean complex tools for analysis, yet fi eld technicians—many of whom are used to working on copper networks—must now focus on fi ber networks, and they must be up to speed with minimal training. Combined, these two trends require that test devices be “smarter” by automating complex testing as much as possible.

In terms of testing requirements, standard measurements like optical signal-to-noise ratio (OSNR) are still as necessary as ever. However, it has become widely accepted that, when applied to next-generation networks, traditional IEC-recommended OSNR measurement techniques fail to deliver the required accuracy. As these next-gen networks employ ROADMs and/or 40G non-polarization multiplexed signals, use of this traditional OSNR measurement method leads to either over- or under-estimation of the OSNR, which is why these networks require the use of in-band OSNR measurements (for a detailed discussion of in-band OSNR, refer to Application Note 235 entitled Optical Spectrum Analyzers in Next-Gen Networks). Moreover, next-generation networks often combine different technologies (data rates, ROADMs, polarization multiplexing, etc.) that might affect some wavelengths, but not others, in which case OSNR measurements must be optimized on a per-channel basis. In this article, we first give examples of complex systems and the challenges they present when measuring OSNR. We then introduce the WDM-Aware technology that provides expert-level, automated OSNR measurements optimized on a wavelength basis. Finally, we conclude with a comparison of the WDM-Aware approach for OSNR measurements using the polarization nulling technique.

EXAMPLES OF COMPLEX WDM NETWORKSFigure 1 shows 10G signals that followed different light paths: channels C_002, C_004 and C_005 went through several ROADMs, while channel C_003 only passed through a single ROADM. One can recognize the typical effect of cascaded ROADMs on channels 2, 4 and 5, as it creates “shoulders” on both sides of the channel.

Accordingly, it becomes necessary to analyze the OSNR of channels 2, 4 and 5 with the in-band method, while the IEC and the in-band methods work just fi ne for channel 3, as shown in Table 1.

Another interesting case that is becoming more and more widespread is the coexistence of different data rates on the same fi ber, as illustrated in Figure 2.

When looking at OSNR values (Table 2), we can clearly see that the in-band OSNR must be used for 40G channel 39, while both the IEC and in-band methods work just fi ne for 10G channels without neighboring 40G channels, like channel 37.

WDM-Aware Technology: OSNR Measurements Optimized on a Per-Channel BasisFrancis Audet and Jean-Sébastien Tassé, Product Line Managers and Daniel Gariépy, Senior Optical Specialist

Table 1. IEC and in-band OSNR values for 10G signals with and without ROADM.

Channel IEC OSNR In-band OSNR

Difference ROADM? Requires in-band OSNR?

C_002 21.03 dB 18.42 dB 2.61 dB yes yes

C_003 27.57 dB 27.42 dB 0.15 dB no no

C_004 22.01 dB 20.95 dB 1.06 dB yes yes

C_005 19.28 dB 17.57 dB 1.71 dB yes yes

Figure 1. WDM trace of 10G signals with and without the fi ltering effect of ROADMs.

Table 2. IEC and in-band OSNR values for mixed 10G and 40G signals.

Channel IEC OSNR In-band OSNR

Difference Data rate

Requires in-band OSNR?

C_037 26.99 dB 27.22 dB -0.23 dB 10G no

C_038 25.94 dB 26.36 dB -0.42 dB 10G yes

C_039 10.93 dB 25.78 dB -14.85 dB 40G yes

C_040 26.34 dB 26.71 dB -0.37 dB 10G yes

Figure 2. WDM trace mixing 10G signals (channels 36, 37, 38 and 40) and 40G signals (channel 39)

© 2011 EXFO Inc. All rights reserved.

Application Note 259

A final example is that of complex systems where some channels have undergone various effects (PMD, carved noise, crosstalk, etc).Figure 3 displays two 10 Gbit/s signals (C_001 and C_003) on both sides of a 40 Gbit/s signal (C_002).

A detailed analysis of channel C_002 revealed that it experienced PMD, as well as crosstalk from the 10G neighbors. In this case, the user needed to set a few parameters in the OSA (number of scans, channel plan, etc.) to obtain a reliable OSNR value, which often required knowledge of the fi ber network confi guration and topology. Table 3 shows that the difference between the IEC and the in-band method for channel C_002 yields a -9.36 dB difference, a huge value. The in-band method is clearly needed here for proper OSNR analysis.

As we can see from the previous examples, the characteristics of each wavelength are becoming more and more unique as next-generation networks are deployed. This requires OSNR measurement optimization on an individual channel basis.

WDM-AWARE TECHNOLOGY: OSNR MEASUREMENT OPTIMIZED PER WAVELENGTHAs shown in the examples above, knowledge of the fi ber network confi guration (data rates, ROADMs or not, PMD or not, etc.) used to be necessary to properly set the OSA in order to get a very reliable in-band OSNR measurement. With the introduction of EXFO’s unique WDM-Aware technology, the expert knowledge is in the box, so that everywhere you go, you are bringing the expert along with you.

The WDM-Aware technology for OSNR analyzes the incoming signal’s properties to determine whether or not any external factors infl uence the noise measurement, such as the presence of:

› PMD

› Crosstalk from neighbors

› Polarization multiplexing

› Polarization scrambling

› ROADMs

It then automatically optimizes the OSA analysis so that factors affecting the measurement are taken into account and an optimal result is provided, as the expert would have done, thus providing an accurate OSNR measurement without user knowledge of the fi ber network. The automatic WDM-Aware settings therefore make WDM network testing easy by decreasing the need for OSA training and freeing up time for other jobs.

INSIDE THE WDM-AWARE APPROACH The WDM-Aware approach for OSNR measurements is based on the normally realistic assumption that the light signal under test is highly polarized and that the superposed noise is unpolarized. The measurement setup includes a polarization controller, a polarization splitter and a dual-channel scanning monochromator (i.e., an OSA), as illustrated in Figure 4 (note that detection, signal processing and display electronics have been omitted for the sake of simplicity).

The polarization splitter divides the light in two states of polarization, SOP-1 and SOP-2. The (internal) polarization controller is adjusted so that the SOP-1 and SOP-2 have a few decibels of difference, a condition that is easy to meet. Given that the signal is polarized and that the noise is unpolarized, the signal level in the two branches will be different, while the noise level in both branches will be the same. The detected spectra in each branch are then subtracted from each other. The two noise contributions are equal and hence canceled, resulting in a spectrum difference that is proportional to the polarized signal only (i.e., noise-free). Figure 5 illustrates this phenomenon.

Figure 4. WDM-Aware measurement setup.

PolarizationController

PolarizationSplitter

SOP-2

SOP-1

OSA

Figure 3. Example of 40G signal with PMD and crosstalk noise.

Table 3. IEC and OSNR values for signal with crosstalk and PMD.

Channel IEC OSNR In-band OSNR

Difference Data rate

Requires in-band OSNR?

C_001 16.73 dB 17.86 dB -1.13 dB 10G yes

C_002 9.33 dB 18.69 dB -9.36 dB 40G yes

C_003 13.57 dB 18.37 dB -4.80 dB 10G yes

Figure 5. The WDM-Aware approach splits the signal without requiring nulling on either branch.

Signal + NoiseP = S + N

Portion (k) of the signal plus half the noisePA = k * S + N/2

Portion (1-k) of the signal plus half the noise

PB = (1-k) * S + N/2

PB

PA

PBS Dual channel OSA

Polarization diversity OSA

P

© 2011 EXFO Inc. All rights reserved.

Application Note 259

It is then important to properly generate the reconstructed signal, proportional to the (data-carrying) signal but not the entire signal. To do so, a proportionality factor must be determined to find out the effective contribution of the signal. As a final step, the noise-free spectrum, multiplied by the proportionality factor, is subtracted from the initial spectrum (SOP-1 + SOP-2), so that the noise level is obtained.

As the name implies, a WDM-Aware OSA means that the unit is completely aware of the content and history of every wavelength (based on powerful data processing) and can fine-tune the analysis algorithm on a per-channel basis, meaning that within a DWDM signal analyzed, every channel will be analyzed with its optimum algorithm, based on the content of that channel.

The main benefit of the WDM-Aware approach is that the in-band OSNR measurement is optimized on a per-channel basis. This is made possible because the measurement time is much faster compared to other in-band OSNR methods, like polarization nulling. To further investigate this claim, we set out in the next paragraph to describe polarization nulling, and we then compare the repeatability and accuracy of both methods.

POLARIZATION NULLING METHOD FOR OSNR MEASUREMENTThe polarization nulling method uses the same setup as the WDM-Aware approach (Figure 5). It involves the adjustment of the polarization controller in order to extinguish, to the highest possible degree, the signal in one of the two branches. Such an adjustment occurs when one of the two outputs of the polarization splitter (acting as a polarizer) is aligned orthogonally with respect to the (polarized) signal. When this is achieved, the only light reaching the detector of that branch, at that particular wavelength, corresponds to half the noise power (since noise is assumed to be unpolarized, and hence split equally between the two branches). This provides a measured noise level, and since the OSA also measures the total power (noise + signal), the OSNR for the particular wavelength can be calculated.

In order for polarization nulling to provide acceptable measurements, the DWDM channels under test and the OSA itself must meet the following criteria:

› The polarization-mode dispersion (PMD) on the link should be very modest, preferably near zero. PMD partially depolarizes the signal within the resolution bandwidth (RBW) of the OSA, so complete nulling becomes impossible. In practice, there is some tolerance to PMD, especially if the OSNR that will be measured is not extremely high (e.g., <20 dB) and the RBW of the OSA is very narrow. Nonetheless, it ultimately limits the maximum attainable OSNR values.

› The polarization optics in the OSA, notably the polarization splitter, must also be able to provide a very high extinction ratio since the maximum measurable OSNR is directly dependent on the extinction ratio. Typically, this is limited to 30 dB.

› The polarization controller in the OSA must be able to rapidly find, for each of several wavelengths within the DWDM channel, the state of polarization (SOP) condition corresponding to the maximum extinction ratio. However, for OSNR values greater than 20 dB, it can become very time-consuming (i.e., long acquisition times) and, in fact, may not even be practical to measure accurately. This is exacerbated when multiple DWDM channels must be characterized simultaneously—for instance, 16 or more channels extending over much of the C band.

BENEFITS OF WDM-AWARE APPROACH FOR OSNR MEASUREMENT VS. POLARIZATION NULLINGSince full polarization extinction is not required, the WDM-Aware approach is much more robust than polarization nulling, since:

› It is not as susceptible to PMD compared to polarization nulling, which assumes a completely polarized signal.

› It is not limited by the polarization-extinction ratio, and hence much larger OSNR can be measured accurately, or equivalently—for a given OSNR sensitivity, the measurement can be performed much more rapidly.

› It is much faster to achieve a few decibels of power difference than complete nulling of a signal, especially for multiple-wavelength transmission (e.g., DWDM).

Figure 6 shows OSNR measurements of 40G signals using two different methods: polarization nulling and WDM-Aware. It gives the OSNR uncertainty as a function of the OSNR value. The plotted uncertainty comprises both a systematic ‘offset’ from the expected value (based on a carefully pre-calibrated test bed) and a random measurement uncertainty (standard deviation).

As predicted, WDM-Aware remains highly accurate for events with high OSNR—with a much faster testing time.

In-band OSNR measurements are also important whenever ROADMs are present in the light path. Figure 7 shows the OSNR results of a 40G network simulation in which signals went through filters (like ROADMs), comparing the values obtained with WDM-Aware and polarization nulling. The plot gives the OSNR repeatability around the average value as a function of the OSNR value. More specifically, we measured, under the same conditions, the OSNR standard deviation of five repetitions for the 10-minute measurements and of 10 repetitions for the 2-minute measurements.

Figure 6. OSNR uncertainty vs. OSNR value for different measurement techniques.

40G OSNR UncertaintyO

SN

R D

evia

tion

+ 2

o S

tDev

(db

)

0.00

20 21 22 23 24 25 26 27 28 29 30

1.00

1.50

2.00

2.50

3.00Polarization nulling, 10 minutes (dB)

WDM-Aware, 2 minutes (dB)

0.50

Figure 7. OSNR repeatability of filtered 40G signals for different measurement techniques.

40G OSNR Uncertainty—Filtered Noise

OS

NR

sta

ndar

d D

evia

tion

(db)

OSNR (dB)

0.00

20 21 22 23 24 25 26 27 28 29 30

0.50

1.00

1.50

2.00

2.50

3.00Polarization nulling, 2 minutes

WDM-Aware, 2 minutes

Polarization nulling, 10 minutes

EXFO Corporate Headquarters > 400 Godin Avenue, Quebec City (Quebec) G1M 2K2 CANADA | Tel.: +1 418 683-0211 | Fax: +1 418 683-2170 | info@EXFO.com

Toll-free: +1 800 663-3936 (USA and Canada) | www.EXFO.com

EXFO America 3400 Waterview Parkway, Suite 100 Richardson, TX 75080 USA Tel.: +1 972 761-9271 Fax: +1 972 761-9067 EXFO Asia 100 Beach Road, #22-01/03 Shaw Tower SINGAPORE 189702 Tel.: +65 6333 8241 Fax: +65 6333 8242EXFO China 36 North, 3rd Ring Road East, Dongcheng District Beijing 100013 P. R. CHINA Tel.: + 86 10 5825 7755 Fax: +86 10 5825 7722 Room 1207, Tower C, Global Trade CenterEXFO Europe Omega Enterprise Park, Electron Way Chandlers Ford, Hampshire S053 4SE ENGLAND Tel.: +44 23 8024 6810 Fax: +44 23 8024 6801EXFO Finland Elektroniikkatie 2 FI-90590 Oulu, FINLAND Tel.: +358 (0)403 010 300 Fax: +358 (0)8 564 5203EXFO Service Assurance 270 Billerica Road Chelmsford, MA 01824 USA Tel.: +1 978 367-5600 Fax: +1 978 367-5700

Application Note 259

APNOTE2591AN © 2012 EXFO Inc. All rights reserved. 2008

Printed in Canada 12/03

We clearly see that the higher the OSNR value, the greater the uncertainty, particularly for polarization nulling. Note that for significant OSNR, polarization nulling is further limited by the extinction ratio of the polarizer (or beamsplitter) used so that on top of the repeatability, a > 5 dB offset is generally added to the offset shown here. Again, the WDM-Aware approach delivers a much better accuracy for filtered signals, like those that passed through ROADMs, in a much shorter time.

CONCLUSIONNetworks are increasing in their complexity, both in terms of cascaded filtering (e.g., via ROADMs in mesh networks) and modulation formats using multiple bits/symbols. For instance, intra-channel noise will increasingly be spectrally carved by filters, the signal bandwidths will frequently be as large as the effective channels widths, and different data rates will be used more and more on a given network. All these factors will affect wavelengths differently, as some wavelengths might have passed through a different number of ROADMs, data rates in a single fiber might vary from one wavelength to the next, etc. Nevertheless, OSNR remains a critical network performance parameter, which requires OSNR measurements optimized on a per-channel basis, like the WDM-Aware technique. Purely polarization-based OSNR measurement techniques (e.g., polarization nulling) can perform well when networks and noise sources remain simple, but as demonstrated herein, the robustness and performance of WDM-Aware measurement renders it well suited for advanced network architectures and modulation formats.