SEVENTH FRAMEWORK PROGRAMME THEME [ICT ] [Photonics]

SEVENTH FRAMEWORK PROGRAMME THEME [ICT-2013.3.2] [Photonics] Software-defined energy-efficient Photonic transceivers IntRoducing Intelligence and dynamicity in Terabit superchannels for flexible optical networks Grant Agreement no. 619603 D6.2 Report on the transmission performance of the flexible packaged devices under Lead beneficiary for this deliverable: ICCS/NTUA Contact Person: Hercules Avramopoulos Address: Iroon Polytechniou 9 15780 Zografou, Greece Phone: +30 210 772 2076 Fax: +30 210 772 2077 e-mail: hav@mail.ntua.gr Date due of deliverable: 30/04/2017 Actual submission date: 31/07/2017 Deliverable Authors: Nikos Iliadis, Giannis Poulopoulos, Nikos Argyris, Dimitris Apostolopoulos Participants: ICCS/NTUA Workpackage: WP 6 Security: PU (Public) Nature: R D6.2: Report on the transmission performance of the flexible packaged under Page 1

Version: 1.0 Total number of pages: 30 D6.2: Report on the transmission performance of the flexible packaged under Page 2

Abstract: The performance evaluation of the packaged devices is reported in the present deliverable. The programmable transmitter has been tested and characterized in the laboratory testbed of ICCS/NTUA examining the flexibility in terms of multilevel modulation formats. The flexible MUX/DEMUX prototype (characterized in WP4 deliverables) is introduced in a network topology at Ericsson Telecommunicazioni premises and a series of experiments revealed the ability of the device to handle various traffic profiles and volumes (ranging from 10 Gb/s NRZ up to 100G coherent signal streams). Keyword list: Flexible Programmable Transmitter, Flexible Multiplexer/Demultiplexer, Laboratory Evaluation D6.2: Report on the transmission performance of the flexible packaged under Page 3

Table of contents 1 Introduction... 5 2 s Single Transmitter performance evaluation under laboratory conditions... 5 2.1 Single Bit Operation... 5 2.2 Single Bit Operation... 7 2.3 Dual Bit Operation... 12 3 SOI MUX/DEMUX performance evaluation under real network conditions.. 22 3.1 10 Gb/s DQPSK data traffic... 22 3.2 40 Gb/s DQPSK data traffic... 24 3.3 100 Gb/s dual polarization (DP) QPSK data traffic... 25 4 Conclusions... 26 5 List of Figures... 27 6 List of tables... 29 7 References... 30 D6.2: Report on the transmission performance of the flexible packaged under Page 4

1 Introduction This deliverable reports on the performance evaluation of the flexible packaged devices (single transmitter and SOI MUX/DEMUX). System performance evaluation of the s single transmitter was carried out in a testbed developed by ICCS/NTUA in collaboration with researchers from HHI and IMEC. Furthermore, the experimental evaluation of the SOI MUX/DEMUX under real network conditions was also carried out at Ericsson Telecomunicazioni premises in collaboration with ICCS/NTUA. 2 s Single Transmitter performance evaluation under laboratory conditions 2.1 Single Bit Operation The experimental evaluation of the 's single transmitter based on InP segmented MZI modulator and 2 nd generation CMOS drivers is thoroughly described. Figure 1(a) illustrates the optical input/output ports and the RF electrical connections of the IQ SEMZM. The single transmitter has been assembled following a chip on board approach, in which a PCB is used to interface the drivers with external RF connections. Figure 1:(a) Optical input/output ports and the RF electrical connections, (b) Bit allocation on the PCB for both I and Q SEMZM. D6.2: Report on the transmission performance of the flexible packaged under Page 5

The module is consisted of two segmented modulators, in an IQ configuration. As it can be observed from Figure 2 both I and Q modulators consist of 10 segments driven by 5 independent differential binary RF signals. Figure 2(a) shows the connection topology between the segments electrodes and the binary drivers of the IQ SEMZM. Binary drivers (which were designed by IMEC) are responsible to control the contribution of each RF input to the modulators segments. Each RF input controls different number of segments, with Bit4 serving as the most significant bit (MSB)driving4 out of 10 segments in each IQ arm. It should be mentioned that two types of segments are distinguished with respect to their length (long and short segments). In Figure 2(b) the final assembly of the single transmitter based on 2nd generation drivers is depicted. Figure 2:(a) Connections mapping of the IQ Segmented Modulator with the 2nd generation drivers assembly (b) Final assembly of the single transmitter based on 2nd generation drivers. Table 1 summarizes the PCB pin-out arrangement for both the I and Q drivers. Regarding the optical IO, two lensed fibers with LC connector are used to couple the light in and out of the modulator. RF connections RF connections RF1 CLK+ RF8 Bit2+ RF2 CLK- RF9 Bit1- RF3 Bit3- RF10 Bit1+ RF4 Bit3+ RF11 Bit4- RF5 Bit0- RF12 Bit4+ RF6 Bit0+ RF13 CLK- RF7 Bit2- RF14 CLK+ Table 1:RF electrical PCB pin-out. D6.2: Report on the transmission performance of the flexible packaged under Page 6

The performance of the s single transmitter was evaluated both in single and dual bit operation. First, each bit of the module was driven independently to verify the proper operation of the device s inputs (single-bit operation). The evaluation was performed by optical eye diagrams acquired by an Equivalent Time Scope. For the MSB (Bit4) of the I and Q SEMZM, BER measurements were also obtained to assess its performance. As a next step, each SEMZM was driven with 2 RF signals (dual-bit operation) to further evaluate the performance of the device, where the MSB was combined each time with one of the other RF inputs. These measurements facilitated, on one hand, the performance characterization of the least significant bits (Bit1, Bit0) as the results in single-bit operation did not lead to straightforward verification of their operation. On the other hand, the capability of each modulator to generate multilevel signals (PAM-4) was also verified with BER measurements and respective eye diagrams. 2.2 Single Bit Operation Figure 3 presents the experimental setup that was employed for the evaluation of the single transmitter's performance with one bit driven in each test. Eye diagrams in each case were captured by an equivalent time Oscilloscope (70 GHz). Moreover, for the MSB of I and Q SEMZM BER measurements were obtained to assess the performance of a 14G NRZ signal. Figure 3:Experimental setup for the single bit evaluation of the 's single transmitter. The differential output (channel 1) of an 8-bit, 65 GSa/s Keysight Arbitrary Waveform Generator (AWG) with a repeating pattern length of 2 9-1 symbols generated an 14 Gb/s D6.2: Report on the transmission performance of the flexible packaged under Page 7

NRZ electrical signals (Figure 4(a)) of 500 mvpp, in order to feed the differential input ports of each input of the I or Q SEMZM. Figure 4:(a) 14Gb/s NRZ electrical data stream (b) Clock signal. A 20-GHz HP signal generator phase locked to the AWG provided the full rate clock reference at 14 GHz for the CMOS drivers which is shown in Figure 4(b). A broadband SHF amplifier was used to adjust the input power levels of the clock signals at 1.2 Vpp before entering the RF circuitry of the device. Two electrical phase shifters were employed in our experimental setup to ensure the precise synchronization between the clock signals and the binary data streams in the respective RF input ports of the I or Q segmented modulators. More particularly, the first electrical phase shifter ensures the synchronization of the two differential clock inputs (CLK+, CLK-) before entering to each segmented modulator, whereas the second achieves to fully synchronize the aforementioned differential clocks with the binary data tributaries. A Distributed- Feedback (DFB) laser emitting 14.5 dbm at 1550 nm provided the optical carrier for the I and Q SEMZMs. At the receiver side, the output NRZ optical signal from the I or Q SEMZM was amplified by an Erbium Doped Fiber Amplifier and then launched into an off-the-shelf 50 Gb/s PIN-photodiode with 0.65 A/W responsivity. A variable optical attenuator (VOA) was used to adjust the incident optical power at the photodiode s input in order to facilitate the BER measurements as a function of received signal power. The photocurrent was captured by a digital Real-Time oscilloscope with 33 GHz analog bandwidth and 80 GSa/s sampling rate for subsequent offline processing and BER assessment. Figure 5depicts the indicative eye diagrams at 14 Gb/s, acquired with an Equivalent- Time oscilloscope, corresponding to the output optical signal of Bit 4, Bit 3 and Bit 2 of the Q SEMZM. From the eye diagrams, we can deduct that the Bit4 and Bit2 are operational (clear eye openings) whilst Bit3 has a severely degraded performance. D6.2: Report on the transmission performance of the flexible packaged under Page 8

Figure 5:Eye diagrams at 14 Gb/s corresponding to the output optical signal of (a) Bit4, (b) Bit3 and (c) Bit2 of the Q SEMZM. Figure 6 illustrates the BER curve obtained for Bit 4 of the Q SEMZM at14 Gb/s plotted against the average received optical power which was calculated by the measured photocurrent. BER measurements were performed to the digitized signal captured at a Real-Time Scope after processing including symbol timing recovery, re-sampling and automatic thresholding for symbol detection. The BER performance was evaluated by comparing the received sampled signal to the original bit sequence and counting the actual erroneous bits of the received signal. As can be observed for an average received power higher than -7 dbm, the received signal has a BER below 3 10-7. It is clear that for received power levels equal to -10 dbm the achieved BER lies below the soft-fec limit (7% overhead, pre-fec-ber 10-3 ). D6.2: Report on the transmission performance of the flexible packaged under Page 9

Figure 6: BER curve at 14 GB/s for Bit4 of the Q SEMZM. Following the same rationale, Figure 7presents in turn the corresponding eye diagrams at 14 Gb/s, of the output optical signal of Bit4, Bit3 and Bit2 for the I SEMZM. It is obvious that Bit4 is fully operational as in case of Q SEMZM, whereas the eye diagram of Bit2exhibited worse performance compared to the respective of the Q SEMZM. Same as before Bit3 seems to have degraded performance. It should be mentioned that both in case of the I and Q SEMZM Bit1 and Bit0 had low output swing power in order to be evaluated independently as a result their operation was verified during the dual-bit operation. D6.2: Report on the transmission performance of the flexible packaged under Page 10

Figure 7: Eye diagrams at 14 Gb/s corresponding to the output optical signal of (a) Bit4, (b) Bit3 and (c) Bit2 of the I SEMZM. Figure 8 depicts the BER curve obtained for Bit4 of the I SEMZM at14 Gb/s plotted against the average received optical power which was derived from the measured photocurrent. The same procedure was followed for BER assessment as in case of Bit4 of the Q SEMZM. It is obvious that for an average received power higher than -5 dbm, the received signal the received signal has a BER below 3 10-7, whereas the achieved BER lies below FEC limit for received power levels higher than -9 dbm. D6.2: Report on the transmission performance of the flexible packaged under Page 11

Figure 8: BER curve at 14 GB/s for Bit4 of the I SEMZM. 2.3 Dual Bit Operation Figure 9 presents an upgraded experimental setup that was employed in order to further evaluate the performance of the 's single transmitter by driving 2 bits of the I or Q SEMZM simultaneously. Following the same rationale as in single bit evaluation, the two differential outputs of the AWG (channel 1 and channel 2) with the same pattern length (2 9-1) generated two differential 14 Gb/s NRZ electrical signals of 500 mv peak-to-peak, in order to feed 2bits of the I or Q SEMZM. The same 20 GHz HP signal generator was used to provide the full rate clock reference at 14 GHz for the segmented modulators. A DFB laser emitting 14.5 dbm at 1550 nm provided the optical carrier for the I and Q SEMZMs as in previous case of the module's single bit evaluation. D6.2: Report on the transmission performance of the flexible packaged under Page 12

Figure 9: Experimental setup for the dual bit evaluation of the 's single transmitter. As a first step the operation of the two least significant bits of both SEMZMs was verified (Bit1 and Bit0). Figure 10 (Q-SEMZM) and Figure 11 (I-SEMZM)depict the optical eye diagrams at the modulator for the case where both Bit4 and Bit1 (or Bit 0) is driven. We observe that a small increase of the output power when one of the two LSBs is driving segments of the modulator at the same time with the MSB leading to the remark that the two LSBs of both SEMZMs are operational. D6.2: Report on the transmission performance of the flexible packaged under Page 13

Figure 10: Eye diagrams at 14 Gb/s corresponding to the output optical signal of combined Bit4-Bit1 (upper diagram) andbit4-bit0 (lower diagram) of Q modulator. Figure 11: Eye diagrams at 14 Gb/s corresponding to the output optical signal of combined Bit4-Bit1 (upper diagram) andbit4-bit0 (lower diagram) of I modulator. D6.2: Report on the transmission performance of the flexible packaged under Page 14

The dual bit operation concept was also facilitated the performance evaluation of the device with multilevel intensity modulated signals. A PAM-4 signal was generated on both SEMZMs by driving Bit4 and Bit2 as the LSB of the 4 with 2 different binary streams. Since Bit3in both I and Q SEMZMs exhibited poor performance in the single bit evaluation, it was not used for the optical PAM-4 generation. At the receiver side, the generatedpam-4 optical signal from the segmented modulators was amplified by the same Erbium Doped Fiber Amplifier and then fed into the 50 Gb/s PIN-photodiode as previously. The variable optical attenuator (VOA) was used to facilitate the BER measurements as a function of received signal power. The photocurrent was captured by the same digital Real-Time oscilloscope for subsequent offline processing to assess the obtained BER measurements. Figure 12 illustrates the obtained eye diagrams at 14 Gb/s, acquired with the Equivalenttime oscilloscope, corresponding to thepam-4 optical signal generated by electrically drivingbit4 and Bit2 of the Q SEMZM. As is already mentioned Bit4 and Bit2 are comprised of five and one segment respectively. To fully evaluate the performance of the I and Q SEMZMs different combinations of the segments of Bit4 and Bit2 were used. More particularly, Figure 12(a) illustrates the obtained PAM-4 eye diagram when all segments of both Bit4 and Bit2 were switched on. An amelioration of the received PAM- 4 optical signal was achieved by switched off one long segment of Bit4 (Figure 12(b)). However, a degradation in the received signal was observed by switched off in turn a short segment of Bit4 (Figure 12(c)). In all the aforementioned cases, the one segment of Bit2 was switched on in order to contribute to the generation of the optical PAM-4 signal. D6.2: Report on the transmission performance of the flexible packaged under Page 15

Figure 12: Eye diagrams of 14 Gb/s PAM-4 optical signal with (a) all segments of Bit4 and Bit2 switched on, (b) one long segment of Bit4 switched off and (c) one short segment of Bit4 switched off. Following the same rationale, Figure 13 shows the obtained PAM-4 optical eye diagrams at 14 Gb/s generated by electrically driving Bit4 and Bit2 of the I SEMZM. The exact same procedure was followed as in case of the dual bit evaluation of the Q SEMZM. Figure 13(a) illustrates the obtained PAM-4 eye diagram when all segments of both Bit4 and Bit2 were switched on. Moreover, Figure 13(b) and (d) depicts the received PAM-4 optical signals by switched off one long segment and two segments with different length (long and short) respectively. Finally, a PAM-4 optical signal with balanced eye diagram openings was obtained by switched off two long segment of Bit4 (Figure 13(c)). In all the aforementioned cases, the one segment of Bit2 was switched on in order to contribute to the generation of the optical PAM-4 signal. D6.2: Report on the transmission performance of the flexible packaged under Page 16

Figure 13: Eye diagrams of 14 Gb/s PAM-4 optical signal with (a) all segments of Bit4 and Bit2 switched on, (b) one long segment of Bit4 switched off, (c) two long segments of Bit4 switched off and (d) one short and one long segment of Bit4 switched off. Figure 14 illustrates the BER bathtub curve of the received PAM-4 optical signal at 14 Gb/s generated by electrically driving Bit 4 and Bit 2 of the Q SEMZM. The same procedure was followed for BER assessment as in case of the single bit evaluation. It is obvious that for an average received powers ranging between -1 dbm and 5 dbm the obtained BER curve lies below the FEC limit. The respective BER curve of the received PAM-4 signal generated by the I SEMZM appeared to have an error floor in the achieved BER which was equal in the order of 10-4. The aforementioned error floor could be explained due to the degraded performance of Bit2 of the I SEMZM which affects significantly the quality of the received PAM-4 optical signal. D6.2: Report on the transmission performance of the flexible packaged under Page 17

Figure 14: BER bathtub curve at 14 GB/s by electrically driving Bit4 and Bit2 of the Q SEMZM. It is worth mentioning that during the experimental evaluation of the 's single transmitter by driving one (single bit operation) or two (dual bit operation) bits of each I and Q SEMZM, the module exhibited a gradually degraded performance. More particularly, the binary drivers that are responsible to control the contribution of the RF inputs to the respective segments of the I and Q SEMZM seemed to have a degraded performance and as a result were not properly operational. Consequently, the quality of the obtained eye diagrams for both SEMZMs was severely affected. The same experimental setup was employed (Figure 3) as in case of the evaluation of the I and Q SEMZM's performance with one bit electrically driven in each test. Figure 15 depicts the indicative eye diagrams at 14 Gb/s, acquired with the Equivalenttime oscilloscope, corresponding to the output optical signal of Bit 4 and Bit 2 of the I SEMZM. It is obvious that the eye diagram of Bit 4 and Bit 2 showed worse performance compared to the respective eye diagrams obtained during the preliminary evaluation of the I SEMZM. Moreover, it should be mentioned that Bit 3 had a severely degraded performance and it was not included in the following eye diagrams. D6.2: Report on the transmission performance of the flexible packaged under Page 18

Figure 15: Eye diagrams at 14 Gb/s corresponding to the output optical signal of (a) Bit4, (b) Bit2 of the I SEMZM. Figure 16 illustrates the obtained eye diagrams for different combinations of segments of Bit 4 for the I SEMZM attempting to ameliorate the received NRZ optical signal. As is already mentioned Bit 4 is comprised of three long segments and two short segments. All segments of Bit 4 were switched off one by one and the received NRZ optical signal is shown in Figure 16 (a) to Figure 16 (e). Moreover, Figure 16 (f) depicts the output optical signal of the I SEMZM by switching off one long and one short segments of Bit 4. In all the aforementioned cases, it is obvious that a severe degradation in the received optical signals was observed. D6.2: Report on the transmission performance of the flexible packaged under Page 19

Figure 16: Eye diagrams of 14 Gb/s NRZ optical signal with (a)-(c) one long segment of Bit 4 switched off (d), (e) one short segment of Bit 4 switched off and (f) one long and one short segment of Bit 4 switched off. Following the same rationale, in order to assess the performance of the Q SEMZM the exact same procedure was followed as in case of the I SEMZM. However, it was not D6.2: Report on the transmission performance of the flexible packaged under Page 20

feasible to receive any response from the Q SEMZM due to the degraded performance of the binary driver and the SEMZM itself. As a final step, the IQ SEMZM transfer function was measured as shown in Figure 17. More particularly, in Figure 17 (a) and (b) the normalized transmitted power of the IQ SEMZM was plotted against the driving voltage to each branch of the nested I SEMZM by biasing at the same time the Q SEMZM at its minimum transmission point. Subsequently, the exact same procedure was followed by biasing the I SEMZM at its minimum transmission point and driving each of the branches of the nested Q SEMZM in order to obtain the 's single transmitter's transfer function. Nevertheless, no fluctuation of the single transmitter's output power (close to -40 dbm) was observed with respect to the driving voltage of the Q SEMZM branches. Figure 17: IQ SEMZM transfer functions with respect to the driving voltage of the (a) first branch of the I SEMZM (b) second branch of the I SEMZM. D6.2: Report on the transmission performance of the flexible packaged under Page 21

3 SOI MUX/DEMUX performance evaluation under real network conditions The performance of the 's packaged SOI MUX/DEMUX device was also evaluated under real network conditions. A 10 Gb/s and a 40 Gb/s DQPSK signals were launched consecutively into the inputs of the MUX/DEMUX module to verify the proper single channel operation of the device under different traffic conditions. The experimental evaluation of the MUX/DEMUX was carried out at Ericsson Telecomunicazioni premises in collaboration with ICCS/NTUA. Figure 18(a) depicts in rack configuration the Ericcson's commercial equipment (transmitters, receivers and EDFAs) that were employed in the experiment, whereas Figure 18(b) shows the 's packaged MUX/DEMUX module. Figure 18: (a) Ericcson's commercial equipment and facilities, (b) 's packaged MUX/DEMUX module. 3.1 10 Gb/s DQPSK data traffic As a first step to assess the filtering capabilities of the MUX/DEMUX device, a 10 Gb/s DQPSK single polarization data stream was fed into one input channel of the device. The MUX/DEMUX device comprises bandwidth flexible and wavelength selective filtering elements in each input channel, which were properly tuned to handle to aforementioned data traffic. The experimental setup that was employed is shown in Figure 19: Experimental setup for the evaluation of the MUX/DEMUX module on data traffic conditions. D6.2: Report on the transmission performance of the flexible packaged under Page 22

. Figure 19: Experimental setup for the evaluation of the MUX/DEMUX module on data traffic conditions. The standalone filtering element of the aforementioned input channel was thermally tuned in order to be aligned with the emission wavelength (1556.179 nm) of the DQPSK data stream. A polarization controller (PC) was employed in the input channel of the MUX/DEMUX ensuring TE polarization for the transmitted data traffic (as required by the MUX/DEMUX design configuration). Moreover, the 3 db-bandwidth of the filtering element was adjusted at 12.5 GHz (Figure 20) in order to induce negligible distortion to the transmitted data stream. The filter's output was amplified by means of an Erbium Doped Fiber Amplifier (EDFA) and afterwards was directly detected confirming error free transmission of a 10 Gb/s DQPSK data traffic through the MUX/DEMUX module. Figure 20: Spectral response of the standalone filtering element (acquired with an Optical Spectrum Analyzer) handling a 10 Gb/s DQPSK single polarization data traffic. D6.2: Report on the transmission performance of the flexible packaged under Page 23

3.2 40 Gb/s DQPSK data traffic Following the same rationale, the exact same procedure was followed as in previous case, in order to evaluate the performance of the MUX/DEMUX by launching a 40 Gb/s DQPSK single polarization data stream into one input channel of the device. The experimental setup that was employed was the same as before illustrated in Figure 19. The standalone filtering element of the aforementioned input channel was thermally tuned in order to be adjusted with the emission wavelength (1556.199 nm) of the DQPSK data stream, whereas the 3 db-bandwidth of the filtering element was configured at 29 GHz. Figure 21 depicts the spectral response of the filtering element acquired with an Optical Spectrum analyzer revealing a small spectral ripple of less than 3 db. The 40 Gb/s DQPSK data traffic was again successfully directly detected (error free operation) after passing through the MUX/DEMUX module. Figure 21: Spectral response of the standalone filtering element (acquired with an Optical Spectrum Analyzer) handling a 10 Gb/s DQPSK single polarization data traffic. D6.2: Report on the transmission performance of the flexible packaged under Page 24

3.3 100 Gb/s dual polarization (DP) QPSK data traffic Figure 22 illustrates the experimental setup that was employed for the performance evaluation of the MUX/DEMUX on a 100 Gb/s QPSK dual polarization (DP) data traffic transmission scenario. Figure 22: Experimental setup for the evaluation of the MUX/DEMUX module on a 100 Gb/s DP-QPSK data traffic transmission scenario. A commercial 100G-DP transmitter (provided by Ericsson) generated the 100 Gb/s DP- QPSK data stream which was launched into an input channel of the MUX/DEMUX module. The DP-data traffic was coupled into the device through a 2D grating coupler and was subsequently de-multiplexed into two identical data streams via the respective standalone tunable filtering elements. The 3-dB bandwidth of the aforementioned filtering elements was set at 35 GHz, whereas their center wavelength was adjusted at 1556.07 nm. Since the two identical output data streams of the MUX/DEMUX were both TE polarized, a polarization controller was employed to rotate the polarization state of one output data stream by 90 degrees. On the other data stream an optical delay line (ODL) was used to ensure bit synchronization at the receiver side by compensating delays introduced by differences between the optical paths of the two data streams (additional polarization controller, differences in the polarization maintaining (PM) fiber lengths). Finally, a polarization beam combiner (PBC) was employed to combine the two data streams with orthogonal polarization states into a dual polarization data traffic which was received by Ericsson's coherent receiver. However, error free transmission was not feasible to be established, D6.2: Report on the transmission performance of the flexible packaged under Page 25

possibly due to the fact that the delays introduced from differences in the optical paths of the two separate data streams were not sufficiently compensated. Therefore, the successful reception of the transmitted dual polarization data traffic was very challenging for the coherent receiver. 4 Conclusions The performance evaluation of packaged devices has been reported in the present document. The Single Transmitter has been evaluated in a series of laboratory experiments at the ICCS/NTUA premises examining the capability of generating highorder modulation schemes by exploiting the flexible design of the Segmented IQ MZM and the CMOS drivers. The MUX/DEMUX prototype has been experimentally evaluated in the transmission testbed of Ericsson Telecomunicazioni. Various data traffic profiles ranging from 10G up to 100G has been handled by the flexible MUX/DEMUX revealing the potential of the device operating in a flexible network topology. D6.2: Report on the transmission performance of the flexible packaged under Page 26

5 List of Figures Figure 1:(a) Optical input/output ports and the RF electrical connections, (b) Bit allocation on the PCB for both I and Q SEMZM... 5 Figure 2:(a) Connections mapping of the IQ Segmented Modulator with the 2nd generation drivers assembly (b) Final assembly of the single transmitter based on 2nd generation drivers.... 6 Figure 3:Experimental setup for the single bit evaluation of the 's single transmitter... 7 Figure 4:(a) 14Gb/s NRZ electrical data stream (b) Clock signal... 8 Figure 5:Eye diagrams at 14 Gb/s corresponding to the output optical signal of (a) Bit4, (b) Bit3 and (c) Bit2 of the Q SEMZM... 9 Figure 6: BER curve at 14 GB/s for Bit4 of the Q SEMZM... 10 Figure 7: Eye diagrams at 14 Gb/s corresponding to the output optical signal of (a) Bit4, (b) Bit3 and (c) Bit2 of the I SEMZM... 11 Figure 8: BER curve at 14 GB/s for Bit4 of the I SEMZM... 12 Figure 9: Experimental setup for the dual bit evaluation of the 's single transmitter... 13 Figure 10: Eye diagrams at 14 Gb/s corresponding to the output optical signal of combined Bit4-Bit1 (upper diagram) andbit4-bit0 (lower diagram) of Q modulator... 14 Figure 11: Eye diagrams at 14 Gb/s corresponding to the output optical signal of combined Bit4-Bit1 (upper diagram) andbit4-bit0 (lower diagram) of I modulator... 14 Figure 12: Eye diagrams of 14 Gb/s PAM-4 optical signal with (a) all segments of Bit4 and Bit2 switched on, (b) one long segment of Bit4 switched off and (c) one short segment of Bit4 switched off... 16 Figure 13: Eye diagrams of 14 Gb/s PAM-4 optical signal with (a) all segments of Bit4 and Bit2 switched on, (b) one long segment of Bit4 switched off, (c) two long segments of Bit4 switched off and (d) one short and one long segment of Bit4 switched off... 17 Figure 14: BER bathtub curve at 14 GB/s by electrically driving Bit4 and Bit2 of the Q SEMZM... 18 Figure 15: Eye diagrams at 14 Gb/s corresponding to the output optical signal of (a) Bit4, (b) Bit2 of the I SEMZM... 19 Figure 16: Eye diagrams of 14 Gb/s NRZ optical signal with (a)-(c) one long segment of Bit 4 switched off (d), (e) one short segment of Bit 4 switched off and (f) one long and one short segment of Bit 4 switched off... 20 D6.2: Report on the transmission performance of the flexible packaged under Page 27

Figure 17: IQ SEMZM transfer functions with respect to the driving voltage of the (a) first branch of the I SEMZM (b) second branch of the I SEMZM.... 21 Figure 18: (a) Ericcson's commercial equipment and facilities, (b) 's packaged MUX/DEMUX module... 22 Figure 19: Experimental setup for the evaluation of the MUX/DEMUX module on data traffic conditions... 23 Figure 20: Spectral response of the standalone filtering element (acquired with an Optical Spectrum Analyzer) handling a 10 Gb/s DQPSK single polarization data traffic. 23 Figure 21: Spectral response of the standalone filtering element (acquired with an Optical Spectrum Analyzer) handling a 10 Gb/s DQPSK single polarization data traffic. 24 Figure 22: Experimental setup for the evaluation of the MUX/DEMUX module on a 100 Gb/s DP-QPSK data traffic transmission scenario... 25 D6.2: Report on the transmission performance of the flexible packaged under Page 28

6 List of tables Table 1:RF electrical PCB pin-out... 6 D6.2: Report on the transmission performance of the flexible packaged under Page 29

7 References [1] www.spirit-project.eu [2] D6.2: Report on the transmission performance of the flexible packaged under Page 30