A new technique to realize an array of multiple true-time-delay elements, which can be independently and continuously tuned, is reported. It is based on a WDM parallel signal processing approach in conjunction with a diffraction-based Fourier-domain optical signal processor. Programmable linear optical phase transfer functions are realized to obtain different electrical true-time delays. The technique can scale to a large number of wideband true-time-delay lines, with continuously tunable programmable delay. Results demonstrate multiple true-time-delay elements with independent tuning control and verify the concept by tuning the free spectral range of a microwave photonic notch filter. To our best knowledge, this is the first demonstration of multiple independently controllable true-time-delay lines for microwave photonic systems.
© 2012 Optical Society of America
True-time-delay elements are instrumental in the realization of phased array antennas that can operate with wide instantaneous bandwidth. Microwave photonic techniques for implementing true-time-delays are particularly attractive because of their unique advantages of low loss (independent of RF frequency), large instantaneous bandwidth, immunity to electromagnetic interference (EMI), and parallel signal processing capability [1–3]. These photonic beamforming techniques offer the prospect of future phased array antenna systems that possess a wide operating bandwidth, low loss, compact size, dynamic reconfigurability, remote antenna feeding, and immunity to EMI.
Various topologies for true-time delays [4–8] and phase shifters [9,10] have been reported previously. However, the former [4–8] have been restricted in the number of true-time delays that they can generate, which is an important limitation in multielement phased arrays. Moreover, they have difficulty in controlling the signal power, which is also required for beamforming. On the other hand, the latter [9,10] have been confined to phase shifting, which is not suitable for squint-free beamforming operation with large instantaneous bandwidth.
In this Letter, we report a new technique for realizing an array of multiple true-time-delay elements, which can be independently and continuously tuned. It is based on a WDM parallel signal processing approach in conjunction with a diffraction-based Fourier-domain optical processor (FD-OP). Our technique features the ability to scale to realize a large number of true-time-delay lines while sharing a single optical processing device. It also features the ability to control the signal amplitude at the same time as controlling the true-time-delays. Experimental results are presented that demonstrate four true-time-delay elements with independent programmable tuning control and minimal RF amplitude and group delay variation across a 4–20 GHz operation range.
The structure of the new microwave photonic multiple true-time-delay array is shown in Fig. 1. It comprises a WDM light source, an electro-optic modulator (EOM), an FD-OP, a demultiplexer, and an array of photodiodes. The WDM light source consists of multiple wavelength optical carriers, which can be realized by an array of distributed feedback (DFB) lasers or a multiwavelength fiber-ring laser. These carriers are modulated by the input RF signal using an EOM with a single sideband with carrier () scheme, which is preferred over conventional intensity modulation because of its spectral efficiency. The modulated carriers are distributed to different locations on the liquid crystal on silicon (LCoS) array, and the different true-time delays and RF powers are controlled via programming of the diffraction-based Fourier-domain processor. The FD-OP is based on a two-dimensional LCoS pixel array whose operation is as follows: the phase and amplitude of the optical signal spectral lines are controlled independently by programming the phase modulation patterns across both the horizontal and vertical axis of the LCoS pixel panel . Hence, true-time-delay RF modulated signals are obtained by programming the LCoS system to realize desired phase slopes. This is followed by a large channel count demultiplexer that routes the different wavelengths to individual photodetectors that output the RF signals.
Experiments were set up to verify the proof of principle for the new true-time-delay system. First, a single true-time-delay element was measured, using a single laser, an EOM constructed as in , an FD-OP (the LCoS system; Finisar Waveshaper, which had 4.5 dB insertion loss), and a 50 GHz photodiode ( Photonics). A vector network analyzer was utilized to measure the operation of the single channel. The laser was intensity modulated by the RF signal at a frequency using modulation, and then the output signal light was sent through the LCoS system, which was configured to provide a programmable optical field transfer function that had a linear optical phase response whose slope could be tuned. The transfer function of an optical channel with the central optical frequency can be expressed as
Figure 2(a) shows the measured RF phase versus RF frequency for different phase slopes, which was achieved by programming the linear phase slope continuously from to by means of the LCoS system. Excellent high-linearity phase responses can be observed from the measurement results. The corresponding time delays are shown in Fig. 2(b). The tuning range is from to . The standard deviation of the delay is less than 0.7 ps. The flat response in Fig. 2(b) indicates a true RF time delay. The measurement range spans from 4 GHz to 20 GHz. The lowest measurement frequency (4 GHz) was determined by the minimum operating frequency of the 90° coupler used for the modulation, and the highest frequency was limited by the bandwidth of the vector network analyzer used in the experiment.
The corresponding RF transmission was also measured, and the results are shown in Fig. 3. The amplitude response is relatively flat with a measured fluctuation of , which shows that the amplitude response of the delay line is independent of the frequency. Note that there is a dependence of the amplitude on the phase slope setting; however, this can be readily compensated by controlling the optical power by programming the power factor .
To demonstrate the ability of the structure to generate multiple independently controlled true-time delays, we set up four channels at central wavelengths of 1546.8, 1548.8, 1550.8, and 1552.8 nm using DFB lasers; e.g., the photodiode optical power was 3 dBm for a single channel for the case of zero phase slope. For each wavelength and true-time setting, measurements were taken over the frequency range from 4 to 20 GHz, and similarly flat responses were observed as for the single channel shown in Figs. 2 and 3. Figure 4 displays the mean time delay and amplitude versus central wavelength of each optical channel. The true-time delays for each channel were controlled independently with a delay change step of 8 ps in the experiment. The results show that the array of true-time-delay elements have a consistent performance and can be implemented independently and simultaneously at multiple optical wavelengths.
Finally, to demonstrate the independent tuning ability of the true-time-delay elements, we set up an experiment on a microwave photonic notch filter. The performance of the time change in the tunable true-time delays can be clearly assessed by observing the shift in the notch frequency of the filter. Figure 5 shows the setup of the microwave photonic notch filter. Continuous wave light from two laser sources at wavelengths of 1548.80 and 1550.80 was combined by an optical coupler and intensity modulated by an input RF signal using the EOM. A linearly chirped fiber Bragg grating (CFBG) with group delay slope () was used as the reference dispersive line, which was placed after the modulation and before the FD-OP. By programming the FD-OP, two true-time-delay elements were obtained at the two laser wavelengths. The electrical transfer function can be expressed as6, which corresponds to a wavelength spacing of 2 nm and the slope of the linearly CFBG. Next, we programmed the two true-time-delay elements to have several phase slopes between and . In the demonstration, both true-time-delay elements were configured to have the same magnitude of phase slope (, , , and ), but with opposite signs. Therefore, according to Eq. (3), the contribution from the time shifters adjusted the filter tap delay in either an increasing or decreasing manner. Correspondingly, the FSR of the notch would be detuned from the initial value toward lower frequencies or toward higher frequencies, as shown in Figs. 6(a) and 6(b), respectively. These results verify the tuning ability of the new multiple true-time-delay elements in microwave photonic signal processing systems.
In conclusion, a new technique to realize an array of multiple true-time-delay elements, which can be independently and continuously tuned, has been reported. It is based on a WDM parallel signal processing approach in conjunction with a diffraction-based FD-OP. Programmable linear optical phase transfer functions are realized to obtain different electrical true-time delays, while also providing the ability to control the signal amplitude. The technique features the ability to scale to a large number of wideband true-time-delay lines, with continuous tunable programmable delay (64 ps in this demonstration) while sharing a single optical processing device. Results have demonstrated multiple true-time-delay elements with independent programmable tuning control, and we have verified the concept by tuning the FSR of a microwave photonic notch filter by programming the true-time-delay elements. To our best knowledge, this is the first demonstration of multiple independently controllable true-time-delay lines for microwave photonic systems.
The work was supported by the Australian Research Council. The authors gratefully acknowledge valuable discussions with M. Roelens, S. Frisken, and S. Poole from Finisar Australia.
1. J. Capmany and D. Novak, Nat. Photon. 1, 319 (2007). [CrossRef]
2. R. A. Minasian, IEEE Trans. Microwave Theor. Tech. 54, 832 (2006). [CrossRef]
3. J. P. Yao, J. Lightwave Technol. 27, 314 (2009). [CrossRef]
4. D. B. Hunter, M. E. Parker, and J. L. Dexter, IEEE Trans. Microwave Theor. Tech. 54, 861 (2006). [CrossRef]
5. B. Vidal, T. Menqual, C. Ibanez-Lopez, J. Marti, I. McKenzie, E. Vez, J. Santamaria, F. Dalmases, and L. Jofre, IEEE Photon. Technol. Lett. 18, 1200 (2006). [CrossRef]
6. L. Jofre, C. Stoltidou, S. Blanch, T. Mengual, B. Vidal, J. Marti, I. McKenzie, and J. M. del Cura, IEEE Trans. Antennas Propag. 56, 1594 (2008). [CrossRef]
7. S. Blais and J. P. Yao, J. Lightwave Technol. 27, 1147 (2009). [CrossRef]
8. R. T. Schermer, F. Bucholtz, and C. A. Villarruel, Opt. Express 19, 5371 (2011). [CrossRef]
9. X. Yi, T. X. H. Huang, and R. A. Minasian, IEEE Photon. Technol. Lett. 23, 1286 (2011). [CrossRef]
10. T. Akiyama, H. Matsuzawa, K. Sakai, S. Itakura, and Y. Hirano, in Proceedings of IEEE International Topical Meeting on Microwave Photonics (IEEE, 2009), paper Fr3.1.
11. M. A. F. Roelens, S. Frisken, J. A. Bolger, D. Abakoumov, G. Baxter, S. Poole, and B. J. Eggleton, J. Lightwave Technol. 26, 73 (2008). [CrossRef]
12. G. H. Smith, D. Novak, and Z. Ahmed, Electron. Lett. 33, 74 (1997). [CrossRef]