All-optical and broadband microwave fundamental ... - OSA Publishing

Vol. 26, No. 6 | 19 Mar 2018 | OPTICS EXPRESS 7336

All-optical and broadband microwave fundamental/sub-harmonic I/Q downconverters YONGSHENG GAO,1,* AIJUN WEN,2 WEI JIANG,3 YANGYU FAN,1 AND YOU HE4 1

School of Electronics and Information, Northwestern Polytechnical University, Xi’an 710129, China State Key Laboratory of Integrated Service Networks, Xidian University, Xi'an 710071, China 3 National Key Laboratory of Science and Technology on Space Microwave, CAST, Xi’an 710100, China 4 Naval Aeronautical University, Yantai 264001, China * [email protected] 2

Abstract: Microwave I/Q down-converters are frequently used in image-reject super heterodyne receivers, zero intermediate frequency (zero-IF) receivers, and phase/frequency discriminators. However, due to the electronic bottleneck, conventional microwave I/Q mixers face a serious bandwidth limitation, I/Q imbalance, and even-order distortion. In this paper, photonic microwave fundamental and sub-harmonic I/Q down-converters are presented using a polarization division multiplexing dual-parallel Mach-Zehnder modulator (PDMDPMZM). Thanks to all-optical manipulation, the proposed system features an ultra-wide operating band (7–40 GHz in the fundamental I/Q down-converter, and 10–40 GHz in the sub-harmonic I/Q down-converter) and an excellent I/Q balance (maximum 0.7 dB power imbalance and 1 degree phase imbalance). The conversion gain, noise figure (NF), even-order distortion, and spurious free dynamic range (SFDR) are also improved by LO power optimization and balanced detection. Using the proposed system, a high image rejection ratio is demonstrated for a super heterodyne receiver, and good EVMs over a wide RF power range is demonstrated for a zero-IF receiver. The proposed broadband photonic microwave fundamental and sub-harmonic I/Q down-converters may find potential applications in multiband satellite, ultra-wideband radar and frequency-agile electronic warfare systems. © 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement OCIS codes: (060.2330) Fiber optics communications; (060.5625) Radio frequency photonics; (350.4010) Microwaves.

References and links 1. 2. 3. 4. 5. 6. 7. 8. 9.

J. Sevenhans, B. Verstraeten, and S. Taraborrelli, “Trends in silicon radio large scale integration: zero IF receiver! Zero I&Q transmitter! Zero discrete passives!” IEEE Commun. Mag. 38(1), 142–147 (2000). S. Vitali, E. Franchi, and A. Gnudi, “RF I/Q downconverter with gain/phase calibration,” IEEE Trans. Circuits Syst., II Express Briefs 54(4), 367–371 (2007). B. Lu, W. Pan, X. Zou, Y. Pan, X. Liu, L. Yan, and B. Luo, “Wideband microwave Doppler frequency shift measurement and direction discrimination using photonic I/Q detection,” J. Lightwave Technol. 34(20), 4639– 4645 (2016). H. Gheidi and A. Banai, “Phase-noise measurement of microwave oscillators using phase-shifterless delay-line discriminator,” IEEE Trans. Microw. Theory Tech. 58(2), 468–477 (2010). T. Yamaji, H. Tanimoto, and H. Kokatsu, “A I/Q Active Balanced Harmonic Mixer with IM2 Cancelers and a 45° Phase Shifter,” IEEE J. Solid-State Circuits 33(12), 2240–2246 (1998). C. Middleton, S. Meredith, R. Peach, and R. DeSalvo, “Photonic frequency conversion for wideband RF-to-IF down-conversion and digitization,” in Proc. IEEE Avionics, Fiber-Opt. Photon. Technol. Conf. (AVFOP), pp. 115–116 (2011). E. H. W. Chan and R. A. Minasian, “Microwave photonic downconversion using phase modulators in a sagnac loop interferometer,” IEEE J. Sel. Top Quantum Electron 19(6), 211–218 (2013). Y. Gao, A. Wen, W. Zhang, Y. Wang, and H. Zhang, “Photonic microwave and mm-wave mixer for multichannel fiber transmission,” J. Lightwave Technol. 35(9), 1566–1574 (2017). X. Yi, T. X. H. Huang, and R. A. Minasian, “Tunable and reconfigurable photonic signal processor with programmable all-optical complex coefficients,” IEEE Trans. Microw. Theory Tech. 58(11), 3088–3093 (2010).

#318747 Journal © 2018

https://doi.org/10.1364/OE.26.007336 Received 29 Dec 2017; revised 1 Mar 2018; accepted 6 Mar 2018; published 13 Mar 2018


10. J. Capmany, D. Domenech, and P. Muñoz, “Silicon graphene waveguide tunable broadband microwave photonics phase shifter,” Opt. Express 22(7), 8094–8100 (2014). 11. W. Liu and J. Yao, “Ultra-wideband microwave photonic phase shifter with a 360° tunable phase shift based on an erbium-ytterbium co-doped linearly chirped FBG,” Opt. Lett. 39(4), 922–924 (2014). 12. V. R. Pagán and T. E. Murphy, “Electro-optic millimeter-wave harmonic downconversion and vector demodulation using cascaded phase modulation and optical filtering,” Opt. Lett. 40(11), 2481–2484 (2015). 13. J. Yao, “Microwave photonics,” J. Lightwave Technol. 27(3), 314–335 (2009). 14. R. A. Minasian, E. H. Chan, and X. Yi, “Microwave photonic signal processing,” Opt. Express 21(19), 22918– 22936 (2013). 15. J. Capmany, J. Mora, I. Gasulla, J. Sancho, J. Lloret, and S. Sales, “Microwave photonic signal processing,” J. Lightwave Technol. 31(4), 571–586 (2013). 16. X. Zou, B. Lu, W. Pan, L. Yan, A. Stöhr, and J. Yao, “Photonics for microwave measurements,” Laser Photonics Rev. 10(5), 711–734 (2016). 17. T. Jiang, S. Yu, R. Wu, D. Wang, and W. Gu, “Photonic downconversion with tunable wideband phase shift,” Opt. Lett. 41(11), 2640–2643 (2016). 18. T. Jiang, R. Wu, S. Yu, D. Wang, and W. Gu, “Microwave photonic phase-tunable mixer,” Opt. Express 25(4), 4519–4527 (2017). 19. F. Zhang, D. Zhu, and S. Pan, “Photonic-assisted wideband phase noise measurement of microwave signal sources,” Electron. Lett. 51(16), 1272–1274 (2015). 20. S. Pan and J. Yao, “Photonics-based broadband microwave Measurement,” J. Lightwave Technol. 35(16), 3498– 3513 (2017). 21. R. Sambaraju, J. Palaci, R. Alemany, V. Polo, and J. L. Corral, “Photonic vector demodulation of 2.5 Gbit/s QAM modulated wireless signals,” in 2008 International Topical Meeting on Microwave Photonics, IEEE, pp. 117–120 (2008). 22. H. Emami and N. Sarkhosh, “Reconfigurable microwave photonic in-phase and quadrature detector for frequency agile radar,” J. Opt. Soc. Am. A 31(6), 1320–1325 (2014). 23. Z. Tang and S. Pan, “Image-reject mixer with large suppression of mixing spurs based on a photonic microwave phase shifter,” J. Lightwave Technol. 34(20), 4729–4735 (2016). 24. J. Zhang, E. H. W. Chan, X. Wang, X. Feng, and B. Guan, “High conversion efficiency photonic microwave mixer with image rejection capability,” IEEE Photonics J. 8(5), 3900411 (2016). 25. Z. Tang and S. Pan, “A reconfigurable photonic microwave mixer using a 90-degree optical hybrid,” IEEE Trans. Microw. Theory Tech. 64(9), 3017–3025 (2016). 26. J. Li, J. Xiao, X. Song, Y. Zheng, C. Yin, Q. Lv, Y. Fan, F. Yin, Y. Dai, and K. Xu, “Full-band direct-conversion receiver with enhanced port isolation and I/Q phase balance using microwave photonic I/Q mixer,” Chin. Opt. Lett. 15(1), 010014 (2017). 27. Z. Tang and S. Pan, “Reconfigurable microwave photonic mixer with minimized path separation and large suppression of mixing spurs,” Opt. Lett. 42(1), 33–36 (2017). 28. P. Li, W. Pan, X. Zou, B. Lu, L. Yan, and B. Luo, “Image-free microwave photonic down-conversion approach for fiber-optic antenna remoting,” IEEE J. Quantum Electron. 53(4), 9100208 (2017). 29. J. T. Gallo and J. K. Godshall, “Comparison of series and parallel optical modulators for microwave downconversion,” IEEE Photonics Technol. Lett. 10(11), 1623–1625 (1998). 30. Y. Gao, A. Wen, Z. Tu, W. Zhang, and L. Lin, “Simultaneously photonic frequency downconversion, multichannel phase shifting, and IQ demodulation for wideband microwave signals,” Opt. Lett. 41(19), 4484– 4487 (2016). 31. Y. Gao, A. Wen, W. Chen, and X. Li, “All-optical, ultra-wideband microwave I/Q mixer and image-reject frequency down-converter,” Opt. Lett. 42(6), 1105–1108 (2017). 32. Y. Gao, A. Wen, W. Zhang, W. Jiang, J. Ge, and Y. Fan, “Ultra-Wideband Photonic Microwave I/Q Mixer for Zero-IF Receiver,” IEEE Trans. Microw. Theory Tech. 65(11), 4513–4525 (2017). 33. C. Lim, M. Attygalle, A. Nirmalathas, D. Novak, and R. Waterhouse, “Analysis of optical carrier-to-sideband ratio for improving transmission performance in fiber–radio links,” IEEE Trans. Microw. Theory Tech. 54(5), 2181–2187 (2006). 34. J. Zhang, A. N. Hone, and T. E. Darcie, “Phase-modulated microwave-photonic link with optical-phase-lockedloop enhanced interferometric phase detection,” J. Lightwave Technol. 26(15), 2549–2556 (2008). 35. Z. Cao, H. P. A. van den Boom, R. Lu, Q. Wang, E. Tangdiongga, and A. M. J. Koonen, “Angle-of-arrival measurement of a microwave signal using parallel optical delay detector,” IEEE Photonics Technol. Lett. 25(19), 1932–1935 (2013). 36. D. Onori, F. Laghezza, F. Scotti, A. Bogoni, P. Ghelfi, M. Bartocci, A. Zaccaron, A. Tafuto, and A. Albertoni, “A DC offset-free ultra-wideband direct conversion receiver based on photonics,” in Radar Conference, IEEE, pp. 1521–1524 (2017). 37. A. A. Abidi, “Direct-conversion radio transceivers for digital communications,” IEEE J. Solid-State Circuits 30(12), 1399–1410 (1995). 38. Y. Y. Labs, “Modulator bias controllers,” http://www.yylabs.com/products.php 39. PlugTech, “Automatic bias control,” http://www.plugtech.hk/main/ 40. M. Martinelli, P. Martelli, and S. M. Pietralunga, “Polarization stabilization in optical communications systems,” J. Lightwave Technol. 24(11), 4172–4183 (2006).


1. Introduction Microwave in-phase/quadrature (I/Q) mixers, including up-converter and down-converter, are essential modules in modern electronic systems. For example, a microwave I/Q downconverter is the kernel module of a zero immediate frequency (zero-IF) receiver [1], an image-reject receiver [2], a Doppler frequency shift estimation system [3], or a microwave phase/frequency discriminator [4]. As shown in Fig. 1, a conventional microwave I/Q mixer contains two frequency down-conversion channels with 90 degree phase difference, which is often realized by an in-phase divider, a hybrid LO divider and two mixers. When it is used for down-conversion, the coming radio frequency (RF) signal can be down-converted by the local oscillator (LO) signal into I/Q parts (IF or baseband signals). However, limited by frequency-dependent electronic dividers and mixers, the commercial microwave I/Q mixer suffers serious I/Q imbalance in wideband applications and finally faces I/Q aliasing or image interference problems [5].

Fig. 1. Conventional microwave I/Q down-converter.

Photonic-based microwave mixers and phase shifters have attracted much attention and research in recent years [6–12], mainly due to the merits of large bandwidth, low frequencydependent loss, and electromagnetic immunity [13–16]. The integration of photonic microwave mixer and phase shifter have been also studied to realize phased array in a superheterodyne transceiver [17, 18] or phase detection in a phase noise analyzer [19, 20]. Combining the microwave mixer and phase shifter in the optical domain, photonic I/Q downconverters are consequently reported in recent years. The quadrature phase difference between the two frequency down-conversion channels can be achieved using dispersioninduced phase shift, which is verified in the photonic vector demodulator proposed in [21] and the radar detector in [22]. The quadrature phase difference can be also realized through differentiated RF/LO driving and proper bias control for the Mach-Zehnder modulator [23, 24]. However, due to frequency dependence of fiber dispersion and electronic couplers, the photonic I/Q down-converter is limited in the operating frequency and signal bandwidth. In [25, 26], an optical hybrid coupler is used to mix RF/LO signals and build the I/Q downconversion channels. However, the discrete parallel links suffer serious environmental disturbance. This problem may be mitigated by minimize the path separation [27], but not eliminated. The polarization division multiplexing technique can solve this problem, as described in the photonic I/Q down-converter based on tandem phase modulation and polarization modulation [28]. However, the tandem photonic converter suffers more optical loss and the conversion gain is significantly lower than the parallel one [29]. Recently, we proposed a photonic microwave I/Q down-converter using a polarization division multiplexing Mach-Zehnder modulator (PDM-MZM) [30]. However, this system suffers large dc offset and second-order intermodulation distortion (IMD2) due to the self-beating after photodetection. In addition, only fundamental microwave I/Q down-conversion can be obtained using the PDM-MZM. In [31], we demonstrate an all-optical microwave imagereject down-converter based on parallel modulation and wavelength division multiplexing. Then we describe a photonic microwave zero-IF receiver based on polarization and wavelength division multiplexing techniques [32].


In this work, we promote our work further and demonstrate an all-optical and broadband I/Q down-converter based on a polarization division multiplexing dual-parallel MachZehnder modulator (PDM-DPMZM). Using the modulator, two parallel optical signals modulate the RF/LO signals, and are then combined into a polarization-multiplexed signal. After optical filtering, power spitting, and independent polarization adjustment, I/Q frequency down-conversion channels are finally obtained. The LO modulation index is properly selected to optimize the conversion gain and noise figure (NF). Due to the all-optical manipulation, the proposed I/Q down-converter exhibits wide operating band and good I/Q balance. In addition, the I/Q down-conversion channels are both balanced detected to suppress the even-order term, which significantly reduces the dc offset and distortions and improve the system dynamic range. Based on the proposed system, image-reject superheterodyne reception and zero-IF I/Q demodulation are demonstrated in the experiment. In addition to fundamental I/Q down-conversion, efficient sub-harmonic I/Q down-conversion is also realized, which not only lowers the frequency requirement for the LO signal, but also reduces the LO leakage. 2. Principles The all-optical fundamental/sub-harmonic microwave I/Q down-converters are described in Fig. 2. An optical carrier, denoted by Ein ( t ) , is generated from a laser diode (LD) and injected into the PDM-DPMZM. The PDM-DPMZM contains two sub-DPMZMs (XDPMZM and Y-DPMZM) in parallel. Both X-DPMZM and Y-DPMZM have two submodulators (Xa, Xb, Ya and Yb). The Xa in X-DPMZM is driven by the RF signal while the Xb is left open. Symmetrically, the Ya in Y-DPMZM is driven by the LO signal while the Yb is left open. Xa, Xb, Ya and Yb are all biased at the null points to suppress the optical carriers.

Fig. 2. Schematic diagram of proposed all-optical fundamental/sub-harmonic microwave I/Q down-converters.

Assuming the RF signal is small and is expressed as VR sin ( Ω R t ) , where VR is the amplitude and Ω R is the angular frequency, the optical field after the sub-modulator Xa is

EXa (t )=Ein ( t ) {exp  jmR sin ( Ω R t )  − exp  − jmR sin ( Ω R t ) } / 4  ∞ n  = Ein ( t )   J n ( mR ) exp ( jnΩ R t ) 1 − ( −1)   / 4    n =−∞ ≈ Ein ( t ) J1 ( mR ) exp ( jΩ R t ) − exp ( − j Ω R t )  / 2

(1)

where mR = π VR / Vπ denotes the RF modulation index, and Vπ is the half-wave voltage of the

modulator. J n ( ⋅) is the n-th order Bessel function of first kind, and n is an integer. We can see that the optical signal contains RF modulated upper and lower sidebands, whose spectrum is shown in Fig. 2. As the sub-modulator Xb is biased at the null point and not driven by the


RF signal, there is no optical signal travelling out. Therefore, the optical field after XDPMZM is

EX (t )=EXa (t ) / 2

(2)

= Ein ( t ) J1 ( mR ) exp ( jΩ R t ) − exp ( − jΩ R t )  / 2 2

Similarly, assuming the LO signal is expressed as VL sin ( Ω L t ) , where VL is the amplitude and Ω L is the angular frequency, the optical filed after Y-DPMZM is

EY (t )=Ein ( t ) {exp  jmL sin ( Ω L t )  − exp  − jmL sin ( Ω L t ) } / 4 2  ∞ n  = Ein ( t )   J n ( mL ) exp ( jnΩ L t ) 1 − ( −1)   / 4 2    n =−∞

(3)

≈ Ein ( t ) J1 ( mL ) exp ( jΩ L t ) − exp ( − jΩ L t )  / 2 2

where mL = π VL / Vπ denotes the LO modulation index. This optical signal travels through a 90 degree polarization rotator (PR). Then the optical signals from the two sub-DPMZMs are combined by the followed polarization beam combiner (PBC), and a polarization-multiplexed optical signal expressed below is generated after the PDM-DPMZM

EPDM (t ) =

EX

= Ein ( t ) ⋅

EY

J1 ( mR ) exp ( jΩ R t ) − exp ( − jΩ R t )  J1 ( mL ) exp ( jΩ L t ) − exp ( − jΩ L t ) 

/2 2

(4)

The optical signal is amplified by an erbium doped fiber amplifier (EDFA) and filtered by an optical bandpass filter (OBPF). The OBPF is used to get one sideband of the optical signal. The polarization-multiplexed single sideband optical signal after the OBPF is

EOBPF (t ) = Ein ( t ) ⋅

J1 ( mR ) exp ( jΩ R t ) J1 ( mL ) exp ( jΩ L t )

/2 2

(5)

The corresponding spectrum is also plotted in Fig. 2. The optical signal is then divided into two channels (I/Q) by an optical splitter. In each channel, a polarization controller (PC) and a polarization beam splitter (PBS) is followed to form two optical signals:

E1 (t ) = Ein ( t )  J1 ( mR ) exp ( jΩ R t ) cos α +J1 ( mL ) exp ( jΩ L t ) sin α exp ( jϕ )  / 4 (6) E2 (t ) = Ein ( t )  J1 ( mR ) exp ( jΩ R t ) sin α − J1 ( mL ) exp ( jΩ L t ) cos α exp ( jϕ )  / 4 (7) where α is the polarization azimuth and ϕ is the phase difference between the two polarization components. In order to realize balanced detection, the PC is adjusted to set α = 450 . In the balanced photodiode (BPD), the two photocurrents subtract each other, and the following current is generated after the BPD:

i ( t ) ∝ E1 (t ) − E2 (t ) ∝ J1 ( mR ) J1 ( mL ) cos ( Ω R − Ω L ) t − ϕ  2

2

(8)

After balanced detection, the desired difference frequency term, or the IF signal, is obtained and the undesired dc offset and even-order distortions are cancelled out. We can also find that the phase of the IF signal is related to ϕ , which can be arbitrarily tuned through polarization control. In order to build I/Q frequency down-conversion channels, the PCs in the two channels should be independently adjusted as follows.


In the first channel, the PC is adjusted to set ϕ = 00 , and the two spectra after the PBS are shown in Fig. 2. The photocurrent after BPD can be written as

iI ( t ) ∝ J1 ( mR ) J1 ( mL ) cos ( Ω R − Ω L ) t 

(9)

So an in-phase down-converted IF signal is obtained. In the other channel, the PC is adjusted to set ϕ = 900 , and the two spectra after the PBS are also shown in Fig. 2. The photocurrent after BPD can be written as

iQ ( t ) ∝ J1 ( mR ) J1 ( mL ) sin ( Ω R − Ω L ) t 

(10)

So a quadrature down-converted IF signal is obtained. Finally, the proposed system becomes a microwave I/Q down-converter. In an RF receiver with a superheterodyne architecture, the two IF signals can be combined by a hybrid coupler for image rejection [30]. In an RF receiver with a zero-IF architecture, the LO frequency is equal to the RF carrier frequency, and the I/Q baseband signals are obtained after I/Q down-conversion. This system can also provide sub-harmonic I/Q down-conversion. In this operation mode, a relative large LO modulation index is applied and the Ya in Y-DPMZM is biased at the peak point to generate the second-order sidebands, along with the optical carrier. The Yb in Y-DPMZM is biased at a specific point, whose bias phase shift is denoted by θ . The main modulator Y-DPMZM is biased at the null point. The optical filed after Y-DPMZM is rewritten as

EY (t )=Ein ( t ) {exp  jmL sin ( Ω L t )  + exp  − jmL sin ( Ω L t )  − 2 cos (θ / 2 )} / 4 2  ∞  n = Ein ( t )   J n ( mL ) exp ( jnΩ L t ) 1 + ( −1)  − 2 cos (θ / 2 )  / 4 2    n =−∞ 

{

(11)

}

≈ Ein ( t ) J 2 ( mL ) exp ( j 2Ω L t ) + exp ( − j 2Ω L t )  + J 0 ( mL ) − cos (θ / 2 ) / 2 2

The amplitude of the optical carrier output from Yb is related to θ , which can be adjusted to equate with the optical carrier output from Ya. By biasing the Y-DPMZM at the null point, the optical carriers cancel each other out. The optical carrier suppression condition can be expressed as J 0 ( mL ) = cos (θ / 2 )

(12)

Then the optical field after Y-DPMZM is written as

EY (t ) = Ein ( t ) J 2 ( mL ) exp ( j 2Ω L t ) + exp ( − j 2Ω L t )  / 2 2

(13)

The following OBPF is now used to filter one of the second-order sidebands. Keeping other system configurations unchanged, the photocurrents after balanced detection are

iI ( t ) ∝ J1 ( mR ) J 2 ( mL ) cos ( Ω R − 2Ω L ) t 

(14)

iQ ( t ) ∝ J1 ( mR ) J 2 ( mL ) sin ( Ω R − 2Ω L ) t 

(15)

Therefore, a photonic sub-harmonic microwave I/Q down-converter is obtained.


3. Experiment and results 3.1 Experimental setup

An experiment is carried out in accordance to Fig. 2. The optical carrier output from an LD (Emcore 1782) has a wavelength of about 1552.2 nm, a power of 50 mw and a relative intensity noise of below −160 dBm. The optical carrier is injected into a PDM-DPMZM (FTM7977HQA) with a linear polarization. The modulator has a half-wave voltage of about 3.5 V and an extinction ratio of 30 dB. The RF and LO signals are generated from a vector signal generator (SMW200A) and a microwave signal generator (N5183A MXG). The maximum frequencies of the two signal generators are both 40 GHz. The modulator is driven by the RF and LO signals according to Fig. 2, and its dc biases are manually controlled using two dc stabilized power supplies (E3631A). The EDFA (KPS-STD-C-19-HG) has a fixed output power of 17 dBm and an NF of 4.5 dB. The followed OBPF has a 3 dB bandwidth of 0.6 nm and a 20 dB bandwidth of 0.93 nm. The response curve of the OBPF is measured by an optical spectrum analyzer (Advantest Q8384) with a resolution of 0.01 nm, and the result is plotted in Fig. 3 using a black dot line. The two BPDs are fabricated using four 1 GHz photodiodes with a responsivity of 1.1 A/W. The matched circuits after the photodiodes are designed symmetrically to reduce length mismatching. Due to the limited experiment condition, the length matching of the two BPDs is not measured in the experiment. 3.2 Fundamental I/Q down-conversion

First, the fundamental I/Q down-conversion is demonstrated and measured. The four submodulators in the PDM-DPMZM are all biased at the null points. A 26 GHz RF signal is directly down-converted to 100-MHz IF signals by the fundamental I/Q down-converter using a 25.9 GHz LO signal. The optical spectra before and after the OBPF are plotted in Fig. 3 using a red dashed line and a blue solid line, respectively. We can see that after optical filtering, the lower first-order optical sidebands are well suppressed by about 25 dB.

Fig. 3. Response curve of OBPF, and optical spectra before and after OBPF in fundamental I/Q down-converter.

Fig. 4. Measured conversion gain and NF versus LO power in fundamental I/Q downconverter.


By tuning the LO power from 2 to 17 dBm in 1 dB step, the conversion gain and NF are measured and plotted in Fig. 4. In this measurement, only a single photodiode is used. As the optical power injecting into the photodiode keeps unchanged (fixed output power of EDFA), a lower LO power will help optimize the power ratio between the LO and RF sidebands, and thus maximize the beating terms (or the down-converted IF signals) [7, 8, 32, 33]. Therefore, the conversion gain varies inversely with the LO power. However, as the conversion gain is improved by further pumping the EDFA, the amplified spontaneous emission noise from the EDFA is inevitably enhanced. What’s more, a low LO power will lead to a small input power of the EDFA. To ensure the desired output power, a large power gain is required, which may exceed the ability of the EDFA. This is the reason why the conversion gain increases little and the NF is obviously deteriorated with a small LO power. As a tradeoff, the LO power in the following experiment is set to 10 dBm. A two-tone test is then conducted and the spurious free dynamic range (SFDR) is measured. A two-tone RF signal (26 and 26.01 GHz) with a power of 0 dBm in each tone are down-converted to an IF signal with the fundamental frequencies of 100 and 110 MHz. The spectrum of the IF signal after one photodiode is measured by an electrical signal analyzer (FSW50) and shown in Fig. 5(a). Except for the two fundamental terms, small third-order intermodulation distortions (IMD3s) are also found at the frequencies of 90 and 120 MHz, which are due to the inherent nonlinearity of the modulator and photodiode. In addition, an IMD2 with a power of −18.9 dBm can also be observed at the frequency of 10 MHz, which originates from the self-beating of the two-tone RF signal. When both photodiodes are illuminated, balanced detection is achieved with proper polarization alignment. Figure 5(b) plots the IF spectrum after BPD. We can see the IMD2 at 10 MHz is reduced to −69 dBm, which is 50.1 dB lower than that without balanced detection. Meanwhile, the fundamental terms and IMD3s are both improved by 5.5 dB, close to the theoretical prediction (6 dB). The well suppressed IMD2 shown in Fig. 5(b) also indicates a good responsivity balance and length matching of the BPD. Then the power of the two-tone RF signal is gradually changed from −30 to 6 dBm and the down-converted IF fundamental term, the IMD2, the IMD3 and the noise floor are measured, as shown in Fig. 6. When a small RF power is applied, the IMD2 and IMD3 are lower than the noise, so the number of points of fundamental, IMD2 and IMD3 is not equal. A conversion gain of −2.4 dB and an NF of 40.6 dB are obtained. Note that the conversion gain and NF are slightly improved compared with the results without balanced detection in Fig. 4. The third-order input intercept point (IIP3) is 28.6 dBm. The second-order SFDR (SFDR2) and third-order SFDR (SFDR3) are 100.2 dB·Hz1/2 and 108 dB·Hz2/3, respectively.

Fig. 5. Down-converted signals with (a) single photodiode, and (b) BPD in fundamental I/Q down-converter.


Fig. 6. IF output power versus RF input power in fundamental I/Q down-converter.

Fig. 7. Waveforms of in-phase IF signal and quadrature IF signal with relative phase of 0, 90, 180, 270 degrees in fundamental I/Q down-converter.

The phase and power balance are then investigated. A 26 GHz RF signal is downconverted into 100 MHz I/Q signals using 25.9 GHz LO signal. The IF waveforms are observed using a dual-channel oscilloscope (DPO7254). The waveform of the in-phase signal (as the oscilloscope trigger) is drawn using a black line in Fig. 7. Through polarization adjustment, the relative phase of the quadrature signal is continuously and arbitrarily tuned. The polarization state (φ and α) is adjusted according to the down-converted I/Q waveforms. When α = 45 degrees, the dc offset after the BPD is eliminated. When φ is 0/90 degrees, the phase difference between the two IF waveforms are quadrature. As an example, the waveform of the quadrature IF signal with relative phase of 0, 90, 180 and 270 degrees are captured and plotted in Fig. 7. We can also find that the amplitudes of the waveforms with different phases are highly consistent.

Fig. 8. Measured I/Q phase and power imbalance versus (a) RF frequency and (b) IF frequency in fundamental I/Q down-converter.

In the followed experiment, the phase difference between the two IF signals is set to 90 degrees. The RF frequency is changed from 5 to 40 GHz in 1 GHz step. By changing the LO


frequency accordingly, the IF frequency remains at 100 MHz. The powers of the I/Q signals and their phase difference are measured and shown in Fig. 8(a). We can see the 3 dB RF operating bandwidth ranges from 7 to 40 GHz. Over this large bandwidth, the maximum power imbalance is 0.5 dB and the maximum phase imbalance is only 0.9 degrees. Then the RF frequency is set to 26 GHz. By gradually tuning the LO frequency, the IF frequency is changed from 0.1 to 1 GHz in 0.1 GHz step. As shown in Fig. 8(b), the power fluctuation over this operating bandwidth is 3.1 dB and the maximum power imbalance is 0.7 dB. The phase imbalance remains below 0.8 degrees. 3.3 Fundamental image-reject receiver

When the proposed I/Q down-converter is applied for an image-reject receiver, the I/Q IF signals will be combined using an analog or digital 90 degree hybrid coupler to eliminate the image product. A 26 GHz RF signal is down-converted to two 500 MHz IF signals using a 25.5 GHz LO signal. The I/Q IF signals are combined by an analog hybrid coupler and the coupled signal is sent to the oscilloscope and signal analyzer in turn. The waveform and spectrum of the desired 500 MHz signal are plotted using blue solid lines in Figs. 9(a) and 9(b), respectively. Next, the RF power is fixed and its frequency is changed to 25 GHz to act as the image interference. The waveform and spectrum captured by the instruments are drawn using red dashed lines in Figs. 9(a) and 9(b), respectively. A weak signal at 500 MHz, which is the image product, is observed. As compared with the desired IF signal, the image product is suppressed by 49.5 dB.

Fig. 9. Desired signal and image product after image-reject down-conversion based on fundamental I/Q down-converter: (a) waveforms and (b) spectra. LO frequency: 25.5 GHz; desired RF frequency: 26 GHz; image RF frequency: 25 GHz.

3.4 Fundamental zero-IF receiver

The proposed I/Q down-converter can be also configured as a zero-IF receiver. In the experimental demonstration, a 26 GHz RF vector signal with 100 MBaud 16 quadrature amplitude modulation (16QAM) is directly down-converted to baseband using a synchronized 26 GHz LO signal. The I/Q baseband signals are simultaneously analyzed to calculate the error vector magnitudes (EVMs). The power of the RF vector signal is tuned from −31 to 10 dBm, and the EVM curve is plotted using a blue line in Fig. 10. The EVM remains below 10% with the RF power sweeping from −25 to 7 dBm. As examples, the constellation diagrams with RF powers of −25 and 5 dBm are shown in the insets, and the calculated EVMs are 8.9% and 8%, respectively. Next, the RF carrier frequency is changed to 36 GHz then direct-converted using a 36 GHz LO signal. The EVM curve is drawn in Fig. 10 using a red line. We can see the two EVM curves with different RF carrier frequencies are quite similar, which indicates a good frequency tunability of the zero-IF receiver. The LO/RF isolation as a function of operating frequency is also tested and plotted in Fig. 11. It keeps below −30 dB over 0–40 GHz and below −40 dB over 0–20 GHz. The RF and LO ports are electro-optic shielded, and the residual leakage mainly results from the crosstalk between the two drive electrodes in the modulator. Nevertheless, this result is much better than conventional microwave I/Q mixers. In some special applications, a much higher LO/RF


isolation may be required to reduce the LO leakage, then the followed sub-harmonic I/Q down-conversion may be more advantageous.

Fig. 10. Demodulated EVM versus RF input power with 26 and 36 GHz carrier frequencies in fundamental zero-IF receiver. Insets: constellation diagrams. Modulation format: 16QAM; signal bandwidth: 100 MBand.

Fig. 11. Measured LO/RF isolation versus operating frequency.

3.5 Sub-harmonic I/Q down-conversion and zero-IF receiving

Next, the system is reconfigured as a sub-harmonic I/Q down-converter. A 36 GHz RF signal is converted to a 100 MHz IF signal by a 17.95 GHz LO signal. The dc biases in X-DPMZM remain unchanged. The Ya modulator in Y-DPMZM is operated at the peak bias point to suppress first-order optical sidebands and enhance second-order ones. According to the Bessel functions, the modulation efficiency of second-order sidebands is lower than the firstorder ones. So in order to improve the sub-harmonic conversion gain, a relative large LO power of 18 dBm is selected in the experiment. The dc biases of the sub-modulator Yb and the main modulator Y-DPMZM are tuned according to the analysis to suppress the optical carrier. Figure 12 shows the optical spectra before and after optical filtering. The upper second-order sideband becomes the dominant term and the harmonic suppression ratio is about 18 dB. The residual first- and third-order LO sidebands result from the limited modulator extinction ratio. Similar to the fundamental I/Q down-converter, the phase difference between the two IF signals can be arbitrarily adjusted through the PCs. The waveforms of the in-phase IF signal and the quadrature IF signal with different relative phases are drawn in Fig. 13. The phase difference between the two waveforms can be adjusted to the desired value while their amplitudes remain unchanged.


Fig. 12. Filter response and optical spectra before and after OBPF in sub-harmonic I/Q downconverter.

Fig. 13. Waveforms of in-phase IF signal and quadrature IF signal with relative phase of 0, 90, 180, 270 degrees in sub-harmonic I/Q down-converter.

The measured I/Q phase and power imbalance are shown in Fig. 14. In the 3 dB RF operating bandwidth (10–40 GHz), the maximum power imbalance is 0.6 dB and the phase imbalance is below 1 degree. With the IF frequency from 0.1 to 1 GHz, the IF signals have a maximum power fluctuation of 3.3 dB. The maximum power imbalance is 0.5 dB and the phase imbalance is below 0.8 degrees.

Fig. 14. Measured I/Q phase and power imbalance versus (a) RF frequency and (b) IF frequency in sub-harmonic I/Q down-converter.


Fig. 15. Demodulated EVM versus RF input power in sub-harmonic zero-IF receiver. Insets: constellation diagrams. Modulation format: 16QAM; signal bandwidth: 100 Mband; RF frequency: 36 GHz; LO frequency: 18 GHz.

The sub-harmonic I/Q down-converter is easy to be configured as an image-reject receiver with results similar to those demonstrated in the fundamental one. Next, only the zero-IF receiver is demonstrated using the sub-harmonic I/Q down-converter. A 36 GHz RF vector signal with a modulation format of 16QAM and a bandwidth of 100 MBand is directconverted to I/Q baseband signals using an 18 GHz LO signal. The measured EVM with different RF power is plotted in Fig. 15. The EVM is below 10% with a power range of 31 dB (−27 to 4 dBm). Two constellation diagrams with RF powers of −25 and 5 dBm are also shown in Fig. 15, with the calculated EVMs of 8.8% and 10.5%, respectively. 4. Discussion

Although the photonic techniques exhibit superiority in bandwidth, loss and electromagnetic isolation, how it can be efficiently applied in modern electronic systems still remains challengeable. Much work has proved the feasibility of the photonic-based single modules such as photonic-based microwave mixer, filter, phase shifter, frequency multiplier, etc. However, practical electronic systems, like an RF transmitter/receiver or microwave instrument, are often complex systems that involve several modules mentioned above. Due to the low efficiency in electro-optic modulation and photodetection, direct-integration of these modules will often lead to the accumulation of noise and nonlinearity, and finally cause performance degradation. The main work in this paper is to design a multifunctional but simple microwave photonic system that is closer to practical application. In the proposed fundamental I/Q down-converter, the microwave down-conversion and phase shifting are efficiently integrated in a simple and all-optical system. Additional frequency doubling for the LO signal is also introduced to the sub-harmonic I/Q down-converter. The highlight of the proposed system is that both the frequency down-conversion and phase shifting obtain a good performance in the experiment: high conversion gain, high IIP3 and large SFDR in frequency down-conversion; arbitrary and continuous tuning in phase shifting. 4.1 Observations

Thanks to the all-optical configuration, the proposed system exhibits an ultra-wide bandwidth: 7–40 GHz in the fundamental I/Q down-converter and 10–40 GHz in the subharmonic one. Furthermore, due to the frequency-independent and power-independent phase shifting, ultra-low I/Q phase imbalance (