Design, Simulation and Testing of the OOK NRZ Modulation Format for Free Space Optic Communication in a Simulation Box

This article deals with the construction of a modulator and demodulator for Free-Space Optical (FSO) communication. In FSO optics, the modulated optical signal is propagated in the constantly changing environment (atmosphere). The optical signal is strongly influenced by the actual composition of air, which is directly linked to the change in refractive index in the turbulent cells. This article examines primarily the appropriate modulation format for FSO. For this purpose, one type of an OOK-NRZ modulator and one type of a demodulator were designed. This article also describes the construction of two types of photo detectors (the high impedance and the transimpedance ones). All electronic constructions were tested in the MicroCap simulator and they were experimentally measured as well. For the OOK NRZ modulation the, maximum transmission speed achieved the value of 160 Mbps. Measuring the quality of the modulation formats was carried out under mechanical and thermal turbulences. The last part of this work gives the results of the measurements of fog influences.


Introduction
At FSO the communication runs by means of light in the atmosphere under the full-duplex operation. There is a carrier wave that is used for communication. This wave includes one or more wave divisible channels. The power of these channels is transmitted by one or more volumes. The FSO transmission path consists of two optical heads integrating both the optical transmitter and the receiver. The transmitting unit is composed of a modulator and radiation sources. The receiver section consists of a detector and a demodulator. Both the receiver and the transmitter are equipped with adaptive optics. Adaptive optics focuses the optical beam on the detector and this way it increases its profits. In the optical transmitter, the laser, laser diodes (LDs), LEDs, or infrared emitting diodes (IREDs) are used as the sources of radiation. The sources of radiation are modulated either indirectly or directly by changing the excitation current of the optical source. The communication takes place in the transmission windows of 850 nm, 1550 nm, and at longer wavelengths. At the receiver, the photodiodes of the PIN type as well as more sensitive avalanche photodiodes APD are used [1], [11].
The advantages of FSO are the installation speed, operating in an unlicensed band, security (complex interception), high data rate and ease of portability. FSO can be used for the last mile networks as a replacement for the RF [1], [2].
As compared with the RF system, FSO is less affected by rain and snow, but, on the other hand, it can be more affected by atmospheric turbulences and fog. The RF communication is very sensitive to rainfall (hydrometeorits). During the turbulences, the turbulent cells of different refractive indices are created. These refractive indices vary due to the fluctuations of their density and temperature. These changes in refractive index bend the light beam (from its original path), resulting in fluctuations of the received light output. The size and frequency of the fluctuations of the received power depend on the amplitude of the turbulent cells. The change of path occurs when the beam passes through a cell that is bigger than this volume. When the beam passes through the smaller cell, there appears the deformation of the wave front [18].
The aerosol particles (mist) strongly influence the light beam. The light beam is mostly absorbed by the fog particles of the size equal to the wavelength of the light source. This leads to a long-term loss of the power in a matter of hours [3], [4], [5], [6]. The attenuation of the atmosphere ranges from 0.22 dB·km −1 (visibility of 20 km) for clean environment to 271 dB·km −1 (visibility of 0 − 50 m) for dense mist and it is therefore necessary to deploy the hybrid FSO/RF [3], [10].
FSO uses LEDs and laser sources with the wavelengths corresponding to an atmospheric transmission window (with the minimum attenuation) [19].
An important parameter for the FSO links is the used modulation format, which can increase its reliability. At FSO, the On-Off Keying modulation (OOK) is used most often. It is used in the NRZ and RZ variants. The OOK is used because of its good spectral efficiency and easy circumferential solutions [6], [7], [8], [9], [10], [11].
The optimal limit level of FSO is affected by the turbulence and absorption of the atmosphere. To stop this negative effect, it is necessary to apply some adaptive decision limit levels. This necessity increased the interest in transition to other types of the modulation formats. These may be the SIM modulation formats (Subcarrier Intensity Modulation). In the SIM modulation format, the modulations BPSK, QPSK, and others modulations. The solution of the adaptive decision levels in the SIM modulation formats is not necessary any longer [2], [12], [13], [14], [15], [16], [17].
The atmospheric influence affects the modulated optical beam and it leads to incorrect detection of the received signal. This is the reason why at present the modulation formats that would be resistant to these interfering influences were searched for. Concerning the modulation formats, it is necessary to find a compromise between the photo-efficiency, bandwidth, and sensitivity to interference, cost and demands of the manufacture [2,13].
In the first chapter of this article, the authors focus on the observations of the influences of the atmospheric phenomena on the FSO modulation formats. In the second chapter the theory of the used modulations formats including their mathematical descriptions is given. In the third chapter the principles of modulation of laser diodes are described. This chapter also contains the principles of a detection process of the optical signal with the basic circuit diagram of photodiodes. The fourth chapter describes the design of the circuits and their testing kits for the modulation formats OOK-NRZ and BPSK. In the fifth chapter the description of the measuring station and the techniques for measuring the impact of the atmospheric phenomena on an eye diagram is given. In the sixth chapter the data taken from the eye diagram of the OOK-NRZ modulation are analyzed. Finally, the last chapter is devoted to the simulation in the OptiSystem environ-ment. The simulation gives an account of the influence of the structural parameter of the refraction index C 2 n on the OOK-NRZ modulation format.

Modulation of Optical Signal
A modulation technique influences parameters, such as bandwidth and energy efficiency, which affect the total system power. The most frequently used modulation for free space optics is OOK (On-Off Keying). The other modulation formats are PSK (Phase Shift Keying), DPSK (Differential Phase Shift Keying) and orthogonal modulation formats. The choice of the modulation technique depends on the type of use, where the compromise between the complexity, optical power, photo efficiency and bandwidth is important. An electrically modulated signal is converted into an optical signal in an E/O converter. The most frequently used modulation type is the intensity modulation when the immediate power of the transmitted energy with the required wavelength is modulated. A photo-detector converts an optical signal into an electrical signal on the site of a receiver. Finally, the information is obtained from the electrical signal by demodulation [14].
The modulation formats exhibit different sensitivity on the turbulence flow in the transmission environment (atmosphere), which causes decrease in received optical power. The level of the received optical power is significant for maintaining the correct bit-error-ratio (BER) with the same bit-rate. The modulation formats should be energetically efficient with low difficulty on the construction of both the receiver and the transmitter. They should also have the smallest bandwidth [2].
The main reason for using the OOK modulation is its simplicity of implementation, but its fixed limit (threshold) level is not optimal because of the fluctuations in the received power. In the PPM (Pulse Position Modulation) the threshold level is not used. There is increased photo-efficiency to the detriment of the complexity of implementation (the need for synchronization) and the bandwidth [13], [31].
The Subcarrier Index Modulation (SIM) does not also use the threshold level and does not require such a wide bandwidth as the PPM modulation, but it has smaller photo-efficiency [13], [31].

On-Off Keying Modulation
It is a type of the intensity modulation belonging to the ASK modulation, in which binary information is represented by the presence or absence of an optical signal. The decision level compensates the probability of the error log 0 and log 1 and is dependent on the received power and noise. This type of modulation is most widely used in free space optics. The production of the transmitter and the receiver is simple. It is resistant to nonlinearity on the laser and the external modulator. The OOK is more sensitive to atmospheric turbulences and other disturbing influences when the fluctuations of the received level occur. At the error rate of BER 10 −9 the photo-efficiency of this modulation is PPB = 39 dB [2], [16], [20], [25]. The rate of errors for this modulation is defined as follows [24], [26], [27], [31]:

Laser Diode Driver
In our tests the generator Stanford CG635 was used as the source of the pseudo random sequences. From the generator the signal entered the modulator, in which the LVDS signal was converted into a modulation signal for the laser diode (Tab. 1). The laser diode was placed in a temperature and current stabilized cube (TCLDM9) from the Thorlabs Company. The Thorlabs thermal stabilizer of the type designation TED200C as well as the LDC205C current stabilizer were also used.
The light signal emitted from a laser diode was collimated to the photo-detector (for 850 nm -A220TM-B and 1550 nm -C230TME-C) using a collimator. The laser beam went through the simulation box, in which it was influenced by the atmospheric effects (the wind turbulence, temperature turbulence and fog). The influenced modulated laser beam fell on the photodetector, in which it was converted back to an electrical signal and afterwards demodulated. On the receiving side there were eye diagrams, which were using an oscilloscope (LeCroy 204Xi). The parameters of the eye diagrams deteriorated due to the atmospheric environment.
The laser diode is modulated by using the input TCMLD9, which is driven by a temperature and current controller. The RF input is directly connected to the laser diode through a bias-tee ( Fig. 1). In the test the bias-tee was designed to bring the DC current and at the same time to bring the modulation signal to the laser diode. The RF input had an SMA connector with the termination of 50 Ω. To this connector an AC component with the frequency of 500 MHz can be brought. The modulation voltage necessary for achievement of the desired modulation current can be derived by the following formula, in which R vst = 50 Ω: For the modulation of the laser diodes the changes of the driving current are used (Fig. 2). This current consists of a DC and an AC component. The DC component keeps the laser diode in the operation mode, in which its characteristic is linear (exceeding I th ). The principle of the modulation when the laser diode is modulated by the OOK signal is shown in (Fig. 2).

OOK Modulator
The square signal with the voltage amplitude peak-topeak U pp = 0.5 V emits from the generator PRBS, since it is necessary to change the voltage of the modulation signal according to the type of the used laser diode. The OOK modulator is placed behind the gen- erator PRBS consisting of a voltage divider and the operational amplifier LM7171 (Fig. 3). The LM7171 allows the current of the output strain up to 100 mA because the signal from the modulator is connected directly to the laser diode through a Bias-tee. The frequency characteristic was simulated in program MicroCap 11. The decrease of the amplitude by 3 dB occurred at f −3dB = 160 MHz (Fig. 4) [26]. The measured eye diagram from the output of the assembled OOK-NRZ modulator is illustrated in (Fig. 5).

Electronic Design of Photo-detector and Demodulator OOK
For the electronic design of a photo-detector in the Mi-croCap, the transimpedance connection of a photodiode with the low-noise operational amplifier OPA847 of a large bandwidth was used (Fig. 6). The amplifier OPA847 is optimized for this circuit. It has the largest bandwidth with the lowest input noise and with the largest slew rate. As for the electronic design, the operational amplifier OPA846 was chosen for the subsequent amplification [28], [29].
The photodiode Thorlabs FDS010 was used. It is spectral sensitivity was from 200 nm to 1100 nm, the minimum bandwidth was 350 MHz and the capacity was from 2 pF (depending on the supplying voltage). At the maximal amplitude of the light signal, the total amount of the optical power falling on the photodetector is 0.166 mW. It results from the characteristic of the photodiode sensitivity (Fig. 7) that the current of 0.0662 mA (at λ = 650 nm) flew through this photodiode. In order to maintain the greatest bandwidth and to increase the signal sufficiently, the bandwidth was calculated (see Eq. (3)). The bandwidth of 95 MHz and the output voltage of 0.45 V were selected at the current of the photodiode of 0.0662 mA. It was also counted on the capacity of the photodiode of 10 pF, with the involvement of the reverse voltage of −5 V, to which the capacitance between the differential inputs as well as the input capacitance of the operational amplifier (2.5+1.2) pF was added. To stabilize the circuit and to avoid the peak amplification of certain frequencies (oscillation), the photodiodes capacity had to be compensated by the parallel capacity C 1 at the resistance R 1 (see Eq. (4)) .
The capacity of the photodiode is compensated by the capacity C 1 . The capacity C 1 is estimated by using Eq. (5). The SMD resistance already includes the parasitic capacitance of about 0.2 pF, therefore the capacity of 0.2 pF was only added to the resistance. It was verified in the simulation. To determine the optimum value of the compensating capacity the temporal analysis was done. It was a temporal analysis of the transimpedance amplifier circuit done by the input square signal from the photodiode with the frequency of 20 MHz and the current I ss = 66.2 µA. During the simulation the compensatory capacity C 1 was changed (Fig. 8), where particular colours of the waveforms correspond to these values (0.2 pF -red, 0.3 pF -green, 0.4 pF -blue, and 0.5 pF -purple). The optimal value of the compensating capacity was between 0.3 and 0.4 pF. When small capacity was selected, the circuit was unstable and it began to oscillate. On the contrary, when a larger capacity was responsible for, the optimal capacity led to the deformation of the output signal and to the decrease of the bandwidth of this connection. To minimize the output noise, the following electronic components were added between the noninverting input and ground: the resistance value of R 2 = 6K8 and the capacity of the values of C 2 = 100 pF and C 3 = 100 nF. Resistor R 4 was connected to the ground through capacitor C 6 , so the amplification of the non-inverting amplifier was A=1 for the prospective DC component and the required gain for the AC component. Thanks to this, the bottom cut-off frequency of the amplifier was f −3dB = 31 kHz. Subsequently, the frequency characteristic of the transimpedance connection was depicted in the simulation. The frequency of the input signal was changed in the range of 1 MHz to 100 MHz. The output voltage was measured in decibels when the value decreased by 3 dB at f −3dB = 93.2 MHz (Fig. 9).

Measuring Station for Testing Atmospheric Effects on the Modulations Formats at FSO
For the purposes of the measurements, a measuring station was established (Fig. 10). It consists of a measuring box, the length of which is 2.5 m. It is equipped with powerful fans to enable measuring of the influence of the mechanical turbulences on the optical beam. For the simulation of the thermal turbulences the direct heaters were used and the glycerin mist generator (Antari F-80Z) was used in order to create the fog. The box was placed on an iron construction, ahead it the holder of a laser diode was installed on an aluminium construction. The laser diode shone upright through the faces of the box, the light beam landed on the photo-detector. The photo-detector was also placed on the aluminium construction.
To provide the temperature and current stabilization of the laser diode, the Thorlabs equipment was used (TDC205C -setting of the DC component of the current, TED200C -setting of the operating temperature, TLCLDM9 -the holder of the laser diode with the input for the modulation voltage). To generate the signal PRBS (pseudo-random bit sequences), the Stanford CG635 generator was used, the output of which was connected to the modulator where it was possible to change the amplitude of the modulation voltage for a certain laser diode. This signal directly modulates the laser diode installed in the TLLDM9 holder. The laser beam from the laser diode was collimated and focused on the photo-detector where the optical signal was converted back to the electrical signal. The signal from the photo-detector was observed directly using an eye diagram on an oscilloscope (in case of the OOK modulation). In case of BPSK modulation the signal was observed at the output of the demodulator only. The measurement was performed with a constructed transimpedance photo-detector and also with Thorlabs detectors (PDA 10C and PDA 10A-EC), the bandwidth of which is larger. At measuring, the same attenuations were ensured by the effect of the box walls when during each measuring the optical beam passed through one Plexiglas wall. The second wall was either withdrawn, or the passage was created through it, so the further decay was impossible.
For the measurements, the photodetector by Thorlabs PDA10A for the wavelength of 850 nm and photodetector PDA 10CF-EC for the wavelength of 1550 nm were used. Because of using another photodetector and also due to difficult collimation of the optical beam, the parameters were not equal. It is therefore impossible to compare the results gained from measuring with the wavelength of 1550 nm.

Thermal Turbulences
During measurements the thermal turbulences the bottom wall of the box was removed as well as the wall ahead the detector. Hot air was blown into the box using direct heaters (CTH-5000 Empire), which were placed under the box. The output of the heaters was 2kW/piece, the temperature of the generated hot air was 48 • C and the flow rate reached 3.5 m·s −1 . The heated air was sucked by, a ventilator, which was placed on the top of the lid. The warm air could also freely issue fromthe open side. The measurement was carried out with several combinations of theswitched fans. Because of the reference, the eye diagram was always measured at the inactive hot-air fans. Then the hot-air fans were gradually switched on in the following combinations: the 1 st and 3 rd fan switched on, the 1 st , 2 nd and 3 rd fan switched on. It is shown in (Fig. 11). In (Fig. 14, Fig. 15, Fig. 16) there are eye diagrams for the OOK-NRZ modulation at the transmission speed of 20 Mb·s −1 .
The fan at the top of the box remained active even at the thermal turbulences due to the extraction of the excessive hot air to prevent the box from being filled up with the air of the same temperature. The temperature in the box with the three fans on was measured using  a thermal camera (FLIR E50).Alternating of the areas with warm and cold air can be seen in Fig. 12).

Measuring Eye Diagram
For the detailed evaluation of the measurement the double value of the reserve signal-to-noise was deducted from the eye diagram (Fig. 13) from the LeCroy oscilloscope). Before each measurement of the atmospheric effects, the reference value of the eye diagram was recorded. Afterwards, the eye diagram was measured with the atmospheric influences. Then the ratio between the reference of the eye diagram opening and the value of the opening of the eye diagram under the influences of the atmosphere was calculated. This ratio was brought to the graph.
Where the variable A ref represents the variable reference eye diagram and the A inf represents the variable eye diagram with atmospherical influence.
The optical beam was influenced significantly by the thermal turbulences; there was significant fluctuation in the received optical power. This caused the closure of the eye diagram. With each added thermal fan the size of the turbulences increased. These thermal tur- bulences caused increasing of the closure of the eye diagram. The worst values were achieved for significant difference intemperature when the 1 st and 3 rd thermal fans were switched on. Thus, two areas with the thermal turbulences were formed and in the middle there was airflow of the room temperature. Figure 14 shows the eye diagram without any atmospheric turbulences. Figure 15 shows the eye diagram when the 1 st and 3 rd thermal fans were switched on. In this case the atmospheric turbulences have significant influence on the fluctuation of the eye diagram. Figure 16 shows the worst state of the eye diagram when all thermal fans were switched on.  In Fig. 17, there is the dimension of the turbulences depending on the data rate. It can be seen that as the thermal turbulences grow, the discharge of the optical power grows too. At the wavelength of 1550 nm was significantly influenced by the thermal turbulences (Fig. 18), but it did not depend on the size of the turbulences so much as it did at the wavelength of 850 nm. With increasing size of the turbulences, there was not such a significant decrease of the received power. The wavelength of 1550 nm the beam was too collimated on the surface of the photo-detector. This surface was much smaller than at the 1stThorlabsphoto-detector and at the deflection of the beam due to the turbulences there was bigger power leakage.These thermal    turbulences manifest themselves mainly during clear weather when the surface of the ground is heated and the warm air convects. The problem may occur when the head is installed over the roof of the building heated by the sun, and the warm air interferes with the optical path. For these reasons, it is necessary to have a sufficient diameter of the optical beam on the receiver and also to plan for a sufficient power reserve.

Simulation of Thermal Turbulences in the OptiSystem Environment for OOK-NRZ Modulation
The simulation in the OptiSystem environment was set according to the parameters of the real measurements with a photodiode transimpedance wiring. For the modulation of the optical radiation the direct modulation was not used. The Mach-Zehnder modulator was used instead. The wiring of the OOK-NRZ modu- lation consisted of a laser diode with the optical power of 4 mW. It was followed by the Mach-Zehnder modulator. To this modulator the output from the OOK-NRZ modulator was connected. The OOK-NRZ modulator was driven by a signal from the generator of pseudorandom sequences. Then the modulated optical signal passed through the simulator of the atmospheric environment (FSO Channel), where the averages of the optical transmitter and receiver were set as well as the atmospheric attenuation. For the measurements of the turbulences the structure parameter of the refractive index C 2 n was set. The optical signal fell on the photodetector PIN. The following values were set: the gain of the operational amplifier (using resistance), sensitivity of the photodiode and the noise parameters. The signal was then amplified by an operational amplifier with the gain of 27 dB. The amplified signal was transferred to the analyzer of the eye diagram where the size of the eye opening was deducted. The scheme of wiring for the OOK-NRZ modulation in the OptiSystem is shown in Fig. 19.
In Tab. 2 the set parameters of the optical free space link are described. These parameters were set in the simulation according to the real measurements.
In Fig. 20, Fig. 23, Fig. 24 there are eye diagrams for the OOK-NRZ modulation for the bit rate of 20 Mb·s −1 . The eye diagrams without any structural parameters of the refractive index and with the structural parameters of the refractive index C 2 n = Fig. 20: Comparison of eye diagrams for OOK-NRZ modulation with structural parameters of refractive index C 2 n = 6.2 · 10 −10 m −2/3 .  the increase of the thermal turbulences causes the increase of the optical power leakage. This increase is also associated with the transmission speed when the reference value of the opening eye diagram decreases concurrently with the increasing transmission speed, which intensifies the influence of the turbulences. In the OptiSystem simulation, the effects of the turbulences on the eye diagram were similar for both wavelengths, but the wavelength of 1550 nm was less affected by the structural parameter of the refractive index C 2 n = 6.2 · 10 −10 [m −2/3 ] (Fig. 24). The effect on both wavelengths equalized together with the increase of the structural parameter of the refractive index (Fig. 25). At C 2 n = 8 · 10 −6 [m −2/3 ] (Fig. 26) the wavelength of 1550 nm was more affected.
By comparison of real measurements and simulations. The similar behavior is observed in properties of eye diagram dependent on intensity of turbulence. When the turbulence intensity increase the eye diagram reduces and it leads to degradation of BER parameter. This behavior results in degradation of transmission capability of Free Space Optical link. This behavior is more evident in the case of simulations (OptiSystem), than in the case of real measurements.

Conclusion
During the measurement of the impact of the turbulent flow on the optical beam, no significant effect of the mechanical turbulences on the received optical power was observed. However, regarding the thermal turbulences, the fluctuation of the received power at the receiver was considerable. This fact is proven by the measured values of the eye diagrams for different types of the modulation and bit rate. The fluctuations of the received optical power caused the decrease in opening the eye diagram. At increasing the turbulence size the closure of the eye diagram was more significant. The most significant closure of the eye diagram occurred at a big temperature difference when the area with the air flow of the room temperature was located in the box between two areas with temperature turbulences. The bit rate also affects the closure of the eye diagram and the increasing transmission rate leads to the closure of the eye diagram. The closure causes the increase of the BER. The thermal turbulences are mainly manifested in clear weather when the surface of the land is heated and the warm air rises. The problem may occur at installing the head over the roof heated by the sun and the warm air influencing the optical path. For these reasons, it is necessary to provide a sufficient diameter of the optical beam on the receiver as well as to count with a sufficient power reserve.