415 ECE EXPERIMENT NO. 9 10Gb/s Optical Communication PURPOSE: To demonstrate and measure the performance of modern 10Gb/s optical communication in various lengths of fiber, and also to observe dispersive effects along with dispersion compensation methods. REFERENCES: 1. C. L. Chen, Elements of Optoelectronics & Fiber Optics, Irwin, (1996), Section 8.4. 2. Agrawal, G.P., FiberOptic Communication Systems, Wiley, (1997), Sections 3.3.7, 4.3.3, 4.5, 4.7, 5.4.4. 3. Agrawal, G.P., Semiconductor Lasers, AIP, (1995), Sections 9.2.1. 4. Proakis, J.G., Digital Communications, McGrawHill, (2001), pp 555556. PROCEDURES: IMPORTANT: Before making any type of fiber connections, you must clean the fiber ends with cotton tipped swabs using ethyl alcohol then dry with compressed air. WARNING: When making any type of adjustment to electrical connections, you MUST be grounded by wearing a grounding strap. The equipment is very sensitive and all antistatic precautions must be taken.
1. Arrange your setup as shown in Figure 1. Connect a variable optical attenuator (VOA) to the optical output of the bit error rate tester (BERT) Optical Transmit/Receive (Tx/Rx) module. Then connect the ouput of the VOA to a 90/10 power splitter. This will allow you to monitor the power at the receiver. The 10% are should be connected to a power meter for observation, and the 90% are should be connected to the optical input.
3/6/2011
111
Figure 1 Setup to record BER
2. Perform Bit Error Rate (BER) measurements when the output of the attenuator is connected directly (or a short patch cord can be used) to the input of the splitter for various powers at the receiver. The receiver has a limited range of optical power for operation, so determine the power range from no signal to error free performance before taking measurements. Record and plot (on semilog scale for the error) BER for varying optical power at the receiver.
3. Repeat the previous step, but with lengths of 20, 40, 60, and 80km of single mode fiber (SMF). Again, adjust your range of measurements as needed for the given lengths of fiber. Observe the effect of the different lengths of fiber on the BER for different optical powers. 4. Connect a spool of Dispersion Compensating Fiber (DCF) with 80km of SMF to see what happens to the BER. Again, record and plot the effect of DCF on the BER. 5. Modify the setup from the previous steps to the setup as shown in figure 2. Optical connections are indicated in red, and electrical are shown in black. This setup is to observe the eye diagrams created by the psedorandom data. Connect the 10GHz Trig Out on the BERT to the Prescale trigger on the 20G scope. Next, disconnect the cable on the Rx Data Out on the BERT and place a new cable leading from that port to any signal in connectors on the scope. Remember to wear the grounding strap when making any electrical connections.
3/6/2011
112
Figure 2 Setup to view and capture eye diagrams 6. Observe your data from the previous BER vs. optical power measurements for fiber various lengths. Pick an optical power that is present in most (or all) fiber lengths, including the DCF, and capture the observed waveform using a computer for all the varying fiber lengths. Note what happens to the eye openings. Estimate the signal to noise ratio in dB from the eye diagrams. 7. An optical communication system can suffer an increase in BER due to chromatic dispersion in the fiber. A certain BER may be observed with a give length of fiber at an optical power, but after the addition of more fiber (and dispersion associated with the extra fiber) the BER might increase at the same optical power. The BER performance, with the extra fiber, can be improved back to the error rate that was observed without the extra fiber by increasing the optical power. This is called a power penalty, achieving the same BER with two lengths of fiber by increasing the optical power. Observe your BER vs. optical power measurements for various fiber lengths to see if there is a pair of data points that have the above given scenario, different fiber lengths and optical powers but similar BER. Capture the eye diagrams for both cases. Why does the BER improve with increased optical power? 8. Pick a low dispersion (low BER) length of fiber, and capture many eye diagrams for varying optical receiver powers. Observe and explain what happens to the eye diagrams at low and high powers. Make sure you record the optical power associated with each eye diagram plot. 3/6/2011 113
QUESTIONS FOR DISCUSSION: 1. From your one of eye diagrams that is in a low dispersion case, estimate the SNR of the communication channel. From that same case, given that you recorded the optical power to the detector for the eye diagram, determine the thermal noise limit theoretically and see how closely they match. 2. Due to largesignal direct modulation of the semiconductor laser, the optical output signal is chirped. This chirp could either have the same or opposite sign as the chirp that is induced through a significant distance of fiber. How does this information affect your plots of BER vs. receiver power for various fiber lengths (without DCF)? Does this help explain some of the observed results? Is the sign of the laser chirp the same or opposite of the dispersion parameter of the fiber? 3. From all the BER vs. receiver power measurements for various combinations of fiber and fiber lengths, estimate the length of regular SMF needed to obtain maximum BER performance.
3/6/2011
114
ECE415 Prelab 9
1. What is dispersion? What different types of dispersion are there and what are their causes? What effects could dispersion have on data transmission? 2. What is a chirped waveform or chirped pulse? 3. What is an eye diagram or eye pattern? How is an eye diagram used to understand various characteristics of a communications channel? Given a relatively clear eye diagram, can you determine the sampling time for optimal performance, and where would it be? 4. What is Bit Error Rate (BER) and what is a Bit Error Rate Tester (BERT)?
3/6/2011
115