• 沒有找到結果。

Experimental setup of the measurement system

The detailed statement of group-velocity dispersion measurement principles has been adequately discussed in previous sections. We will explore the working mechanism and achievements of the measurement system further in this section. The schematic setup of our novel dispersion measurement system is shown in Fig. 3-1, which can be divided into two parts. Firstly, the wavelength swept light source has the slow periodic wavelength variation of intensity-modulated light and the light goes through the fiber under test which will induce timing variation of the signal through the dispersion effects. Secondly, the optical signal is detected by a fast photodiode followed by a RF spectrum analyzer so that the group-velocity dispersion coefficient of the test optical fiber can be obtained by directly analyzing the frequency spectrum.

Fig. 3-1 Experiment setup of the dispersion measurement system

27

The wavelength-swept light source utilizes slow sinusoidal wavelength filtering to an originally broadband signal with fast intensity modulation. A Mach-Zehnder modulator provides the white light beam with amplitude modulation and operates at a speed around 10 GHz in order to enhance the detection sensitivity. The external modulation method is relatively expensive but reduces chirp quite substantially. It is constructed as transverse integrated-optical devices with lower operating voltages. The light can be conveniently coupled into, and out of, the modulator by the use of optical fibers. The waveguide is fabricated in an electro-optical substrate (often LiNbO3) by diffusing materials such as titanium to increase the refractive index. When the electro-optic modulator is biased by a DC power supply at half-wave voltage, we ensure that this device works in the linear region. A bias tee is a three port network used for setting the DC bias point of some electronic components without disturbing other components. The low frequency port is used to set the bias by the a piezo controller; the high frequency port passes the radio frequency signals but blocks the biasing levels; the combined output port is connected to the Fabry-Perot filter, which sees both the bias and RF [1]. The sinusoidal signal of 1kHz is therefore applied through the bias-tee device.

An Erbium doped fiber amplifier (EDFA) is put before the fiber under test to function as a power amplifier to provide enough optical power. An amplified spontaneous emission light source typically has high output power, wide spectral range, large spectral bandwidth, and high stability against temperature change.

28

The fiber-type Fabry-Perot tunable filter (FFP-TF) generates wavelength-tunable narrowband signals and its lensless fiber construction allows high finesse and low loss transmission profile. The filter is applied to the light passing through the Mach-Zehnder modulator so as to induce a slow varying sinusoidal wavelength scanning. The piezo controller is driven by the LABVIEW software of the computer to change the etalon length of FFP-TF for periodic wavelength scanning.

The temperature test of the fiber-type Fabry-Perot filter is shown in figure 3-2. When the Fabry-Perot filter is heated by a hot plate, its center wavelength is shifted to the longer wavelength. From the fitting line, we can discover that center wavelength approximately increases 0.57 nm while temperature rises by 1℃.

26 28 30 32 34 36

1550 1552 1554 1556

C e n te r W a v e le n g th (n m )

Temperature (

C)

Data

C.W.=1536.094+0.57* temperature

Fig. 3-2 The sensitivity of the Fabry-Perot filter

29

Table 3-1 The optical devices and instruments

1. ASE light source : maximum output power : 17.17 mW 2. Mach-Zehnder modulator

3. Signal generator  2

4. Fiber Fabry-Perot tunable filter : central wavelength : 1520~1570 nm 5. Bias T

6. Piezo controller : 0~150V

7. Erbium doped fiber amplifier : signal gain : 25dB 8. Photo detector

9. Receiver 10. Hot plate

The broad band light ASE source is indirectly modulated and output a periodically sinusoidal wavelength-swept light beam. The gradual scanning of the selected wavelength bandwidth will induce timing variation owing to the dispersion effect of the added section of single mode fiber. There are two steps in Fig. 3-3 which explain the process for dispersion measurement. First, the output light of the wavelength-swept light source displays spectrogram on the optical spectrum analyzer, and the scanning range of wavelength is variable by adjusting amplitude of the low frequency signal generator. The value of  is then calibrated for certain applied signal and DC bias. Secondly, the light passes through a single mode fiber and produces a frequency spectrum on the RF

30

spectrum analyzer. The specific timing jitter

s

is therefore obtained, and the group-velocity dispersion coefficient of the test fiber is inferred with the known .

Fig. 3-3 The GVD identification process

3.2 Analyses with different lengths of fiber

The different lengths of single mode fiber introduce different magnitudes of dispersion effects. By making use of the periodic swept-wavelength light source, the group-velocity dispersion coefficient can be efficiently identified within minutes. The optical spectrum of the light source output is depicted in Fig. 3-4. The optical spectrum exhibits an inverse parabolic shape because the spectral edges have longer integration time than the central part due to the sinusoidal swept operation of the tunable filter. The scan speed of central band is faster than two edges. The largest wavelength scanning band that can be achieved with our current ASE source is selected to enhance the measurement sensitivity for the following measurements. The periodic wavelength-swept range is about 7 nm, and its half peak-to-peak value,

, is 3.5 nm.

31

1540 1545 1550 1555

-48 -46 -44 -42 -40 -38 -36 -34

Intensity(dBm)

wavelength(nm) 7 nm

Fig. 3-4 The optical spectrum around central wavelength 1550nm

The RF spectrum of the light source output is shown in Fig.3-5. The magnitude of sinusoidal timing variation will be changed when different types of external fibers are added. The frequency of sinusoidal timing variation is 1 kHz, which is set at first. The slow sinusoidal wavelength filtering applied onto the output light of the Mach-Zehnder modulator produces high order peaks around the central peak on the RF spectrum.

Spectra of different lengths of single mode fiber are plotted in Fig.

3-6 to Fig. 3-14. The timing jitters induced by these external fibers are calculated by Eq. (2.18) and from them the experimental value of group velocity dispersion parameter D is obtained. It obviously shows excellent consistency with the known value of standard SMF fibers, 17 ps/km-nm [2].

32

9.996 9.998 10.000 10.002 10.004 -100

-80 -60 -40 -20

Sapn 10kHz RBW 10Hz

In te n si ty ( d B m )

RF spectrum (GHz)

相關文件