Today’s world is a highly digital world. System designers are more and more inclined to digitize all processing processes because of many attractive features of digital technology, such as high speed, flexibility, serialization and high reliability. In this way, as a bridge connecting the real analog world and the digital world, the role of analog-to-digital converter is becoming more and more important, and the requirements for its performance are becoming higher and higher.
High performance analog-to-digital converter is an essential link between analog sensors (such as radar, communication equipment and electronic warfare equipment) and digital signal processing system. In recent years, on the one hand, with the popularization and application of electronic computers and the continuous improvement of detection automation, higher requirements are put forward for the performance of analog-to-digital converters (bit accuracy, sampling rate, etc.). Signal processing systems such as broadband radar, electronic reconnaissance, electronic countermeasure, nuclear weapon monitoring and spread spectrum communication require high conversion rate above GSPS. For example, a phase array antenna ideally requires hundreds or even thousands of low-power analog-to-digital converters. Typically, each requires 100 MHz bandwidth and 16 bit accuracy. Although these devices may account for only a small part of the whole system, they may be the bottleneck factor affecting the performance of the whole system. On the other hand, the development of parallel computing structure and its technology has produced a digital processor with 100GHz floating-point operation ability, but it can not be fully utilized due to the limitation of ADC performance. If an analog-to-digital converter with a sampling rate of 10 ~ 100 GSPS can be obtained, it can not only improve the performance of the existing system, but also have new application prospects.
At present, there are three kinds of analog-to-digital converters: electronic semiconductor analog-to-digital converter, superconducting material analog-to-digital converter and optical analog-to-digital converter (OADC). Superconducting materials need low temperature conditions, which greatly limits its application fields. At present, the most widely used electronic analog-to-digital converter has many advantages, such as wide application range, mature manufacturing technology and low cost. However, in the field of high-performance analog-to-digital converter, it has inherent shortcomings. When the sampling rate is greater than 2 MSPs, the sampling time is uncertain due to the influence of aperture jitter. The trend is that when the sampling rate is doubled, the bit accuracy decreases by about 1 bit. In the past 10 years, the average improvement of bit accuracy of electronic analog-to-digital converter is only 1.5 bits at a given sampling rate. At present, the fastest sampling rate that electronic analog-to-digital converter can achieve is 8 GSPS and the accuracy is 3 bit; At 8 bit accuracy, the sampling rate of 4 GSPS can be achieved. However, this is basically close to the limit of its theory. Even if the sampling rate can be improved, the corresponding bit accuracy will be reduced accordingly. Therefore, in order to meet the requirements of practical application, that is, the sampling rate is more than 10 GSPS and has appropriate bit accuracy (more than 4 bits), we must seek a new breakthrough. Using optical analog-to-digital converter technology has become the development trend of high conversion rate and high bit precision analog-to-digital converter.
2、 Main technical specifications of optical analog-to-digital converter
Like other analog-to-digital converters, the main technical indexes of optical analog-to-digital converters include: state resolution (represented by the number of bits n of encoded binary digits, usually represented by the number of bits), sampling rate (represented by the number of samples per second, samples / s or SPS), signal-to-noise ratio (SNR) and parasitic free dynamic range (SFDR, i.e. SPU rious free dynamic range) and power consumption (pdiss), in which calibration accuracy and sampling rate are the main performance indexes. Another common technical index of analog-to-digital converter is effective bit accuracy (Neff) , the effective bit accuracy refers to how many bits of the output calibration accuracy are actually effective. It can be expressed by signal-to-noise ratio, and its conversion relationship is
3、 Research progress of optical analog-to-digital converter
The optical analog-to-digital conversion technology was first proposed by S. Wright and others in 1974, and then two mainstream research stages have been formed in the adopted technology. One is from the middle and late 1970s to the mid-1980s, which mainly adopts integrated optical technology, and its main device forms are LiNbO3 waveguide Mach Zehnder interferometer array Balanced bridge modulator and channel optical waveguide Fabry Perot modulator array. Second, since the early 1990s, by drawing lessons from the technical schemes of time division multiplexing and wavelength division multiplexing of optical communication, the time division or wavelength division analog-to-digital converter of photoelectric hybrid mode has been adopted to reduce the required sampling rate through parallel processing.
The optical analog-to-digital converter proposed by Wright applies the analog signal voltage V to the interdigi tal electrode established on the electro-optic material substrate to spatially periodic phase modulate the wavefront of the laser beam passing through the substrate. As a result, different diffraction orders are obtained in the far field. By adjusting the zero order and first-order thresholds, the 2-bit green code output can be obtained, and the symbol bits can be included by applying the third comparator. Compared with the following schemes, this scheme is undoubtedly crude, but it pioneers the optical analog-to-digital converter. The principles of electro-optic modulator and optical detector proposed by it are still applicable today.
The analog-to-digital converter using Mach Zehnder interferometer array with integrated optics proposed by Taylor in 1975 has a wide impact on the development of optical analog-to-digital converter, as shown in Figure 1.
It uses several integrated Mach Zehnder interferometers to form an array. The digitized analog voltage V is applied to the electrode of each modulator at the same time, and the length ln of the electrode changes according to the binary sequence (2n). When a laser with an input intensity of I0 passes through one of the modulators, the output light intensity in synthesized by the two arms changes due to interference, which can be expressed as
Where ψ N is the additional optical phase difference between the two arms caused by the applied voltage V;
ψ N is the static phase difference between the two arms caused by the asymmetry of the two arms.
After the output light intensity of each modulator is received by the photodetector, the light intensity value is quantized into binary code “0” or “1” after comparing with the same threshold voltage. Another method that can be adopted is to slightly change the design of the comparator array, including setting a fixed phase for some modulators to produce the output in gray code format, and the output form is shown in Fig. 2. The reason why gray code is selected is that it produces only one bit change in each quantization level. Unlike shifted binary code, there are multiple bit changes in some specific quantization levels.
In equation (2) φ N can be expressed as
andThe electrode length of the least significant bit), so that when the bit bit increases, the half wave voltage will soon be reduced to the extent that the process level can achieve, which is also a main aspect that limits the improvement of bit accuracy of optical analog-to-digital converter.
Tayler’s scheme is simple in form, can directly generate Gray code output, and all devices can be integrated into one chip in principle. One of the devices using this scheme achieves 1 GHz sampling rate, 4-bit code conversion and 500 MHz signal bandwidth. However, a basic limitation of this scheme is that each additional bit needs to double the length of the modulator electrode of the least significant bit. Taking LiNbO3 as an example, when its significant bit is 6 bits, the sampling rate is about 1 GHz due to the limitation of transit time. And with the increase of the number of bits, the Y Splitter also increases accordingly, which will increase the total insertion loss and limit the improvement of bit accuracy.
The balanced bridge optical analog-to-digital converter uses a 3 dB coupler instead of the Y-branch waveguide (see Fig. 3) to reduce the transmission loss. Moreover, since the two inputs of the comparator behind the modulator are subject to the same effect, even if the intensity of the light source fluctuates, it will not cause obvious conversion error. However, this structure is more demanding in technology, and requires twice the comparator compared with Mach Zehnder analog-to-digital converter.
The channel optical waveguide Fabry Perot modulator (see Fig. 4) does not need to make a complex Y-branch waveguide, but only needs to make a straight channel waveguide, which avoids the technical complexity, reduces the total length of the device and reduces the optical insertion loss. However, each bit requires a laser, which affects the improvement of its bit
The above two devices are improved and evolved from Taylor’s scheme. In principle, they still can’t get rid of the limitations brought by half wave voltage. Generally speaking, their performance can’t exceed the performance limitations of Mach Zehnder optical analog-to-digital converter. However, Taylor’s scheme has a far-reaching impact. After entering the 1990s, it has been further improved in order to improve its performance. There are two methods worth mentioning here. One method proposes a symmetric digital system. Its core idea is to obtain multiple different quantization levels by adding a small number of comparators, so as to significantly increase the bit accuracy. Its coding scheme is shown in Fig. 5. This method uses 3 interferometers and 39 comparators, and can achieve 11 bit accuracy. However, this method improves the nominal accuracy, and the improvement of effective bits is far less than the nominal accuracy. Another method is to propose an optical folding flash analog-to-digital converter by optimizing the waveguide design, which eliminates the restriction that the electrode length needs to be doubled for each additional bit. The waveguide design is shown in Fig. 6. However, the design of its Y-branch waveguide Figure 6 optical folding flash analog-to-digital converter will undoubtedly be more complex. The above two methods have their own limitations, but their ideological methods are still worthy of our reference. Generally speaking, the research on the first generation optical analog-to-digital converter has basically stagnated since the 1990s. On the one hand, due to the principle limitation of the first generation optical analog-to-digital converter, on the other hand, due to the further development of electronic analog-to-digital converter, its performance has exceeded the level that the first generation optical analog-to-digital converter can achieve.
In the 1990s, people were faced with such a situation: on the one hand, analog-to-digital converters are still the bottleneck factor for further improving the performance of many systems, on the other hand, the performance of electronic analog-to-digital converters and the first generation optical analog-to-digital converters can not meet the requirements. This forces people to actively look for new analog-to-digital converter technology. At this time, the gradual maturity of optical communication technology and its rapid development provide a new idea for people to develop optical analog-to-digital converter technology. People began to use time division multiplexing and wavelength division multiplexing methods in optical communication for reference, use the high-speed rate and high time accuracy of laser for sampling, and use the multiplexing device of optical communication to parallelize the sampled signal, so as to reduce the high-speed rate required for quantization. Most of these schemes are combined with electronic technology, and electronic analog-to-digital converter is used for later quantization processing. Two relatively simple schemes were proposed earlier. The first is to use the time division multiplexing technology, use the high repetition rate pulse of the mode-locked laser to sample the electrical signal through the modulator, conduct optical time division multiplexing through the optical switch, distribute the signals in different time sequences to different optical paths, conduct photoelectric conversion, and then quantify through the electronic analog-to-digital converter (as shown in Fig. 7) 。 The second is to use multiple lasers to accurately control the timing of each different laser pulse, so that the laser pulse of each wavelength can sample the analog signal in turn, and then distribute the sampling signals of different wavelengths to different optical paths after wavelength division multiplexing, and the subsequent processing is the same at the same time (as shown in Fig. 8). These two analog-to-digital converters have higher sampling speed and bit accuracy than the first generation optical analog-to-digital converter, but both schemes need complex and accurate timing devices, which undoubtedly improves the complexity of the system. In addition, the improvement of sampling rate of time division multiplexing scheme also depends on the improvement of optical switch rate. The improvement of bit accuracy of wavelength division multiplexing scheme is at the cost of increasing the number of lasers, which are the bottleneck factors limiting the improvement of performance of these two schemes.
On the basis of the above two schemes, people continue to carry out research and propose an analog-to-digital converter based on optical delay, which absorbs the advantages of the above two schemes and eliminates the complex timing circuit. One implementation scheme is shown in Fig. 9. It adopts a supercontinuum wide spectrum EDFL fiber laser (the spectrum width is tens of nanometers, the pulse width is sub picosecond, and the repetition rate is about gigahertz). After a section of fiber is transmitted, it first passes through a polarization beam splitter (PBS), and then passes the polarized light through a WDM device, It is divided into several wavelengths. After each wavelength passes through different lengths of polarization maintaining fiber, the Faraday mirror reflects the light of each wavelength back. After passing through WDM and polarization beam splitter again, a pulse sequence containing light of different wavelengths is synthesized. The radio signal is sampled through a modulator, and the sampled pulse sequence passes through another WDM device, It is distributed to different optical paths according to wavelength to realize parallel processing. A device using this scheme achieves a sampling rate of 18 GSPS and a sampling accuracy of 7 bits.
Compared with the rapid development of foreign optical analog-to-digital converters, the research in the field of optical analog-to-digital converters in China started late, and began in the late 1980s. In the early 1990s, the Department of Applied Physics of Shanghai Jiaotong University studied Mach Zehnder integrated optical analog-to-digital converter. Shenyang Institute of technology and Changchun Institute of physics of Chinese Academy of Sciences jointly developed LiNbO3 proton exchange optical waveguide Fabry Perot 4-bit electro-optic analog-to-digital converter in 1994. At present, the second phase of optical analog-to-digital converter has not been studied in China.
4、 Application of optical analog-to-digital converter
Optical analog-to-digital converter has important applications in many aspects. At present, the research on optical analog-to-digital converter mainly focuses on the application in the system requiring high-speed information acquisition and processing, among which the most important application is microwave digital radar. As we all know, the current microwave digital receiver requires the received analog signal to be mixed and filtered in several steps to reduce the signal frequency to the baseband range of electronic analog-to-digital converter. This process is not only expensive, but also limits the reliability and instantaneous bandwidth of the system, but also increases the size and weight of the system. In addition, each mixing process will bring signal distortion and increase electromagnetic interference. If we can develop a high-speed, high dynamic range analog-to-digital converter to digitize the RF signal directly, it will greatly improve the performance of the digital receiver. According to Jane’s International Defense Review in June 1998, DARPA plans to spend about $40 million on “photoelectric analog-to-digital converter technology” in the next four years. Its purpose is to provide devices that can handle sampling rates up to 1000 GSPS. The “photoelectric analog-to-digital converter” program aims to overcome the limitations of electronic circuits used in the past by using advanced photoelectric components (such as lasers, modulators, detectors, microelectronics and optoelectronic devices). This will allow direct analog-to-digital conversion of signals at signal sources throughout the spectrum of interest to military systems, resulting in performance improvements in the following aspects: improving digital waveform shaping to suppress interference; It has a wide dynamic range to detect targets in clutter; It has a wide instantaneous bandwidth to improve target recognition. For example, when the sampling rate reaches 1000 GSPS, it may produce a new ability of direct broadband analog-to-digital conversion of millimeter wave signals.
In addition to the mainstream technologies mentioned above, there are various non mainstream and auxiliary technologies, such as optical analog-to-digital converter using seed, optical analog-to-digital converter using acousto-optic thermal modulation and optical oversampling Technology (∑) Δ Technology) to improve the effective bit accuracy of analog-to-digital converter, etc. The existence of these technologies, on the one hand, shows that the optical analog-to-digital conversion technology is still in the exploratory stage, which is not really mature, on the other hand, it also shows that the optical analog-to-digital converter has broad research prospects. From the development trend of optical analog-to-digital converter, the system tends to be complex. In order to realize the practical analog-to-digital converter with current sampling rate of more than 100 GSPS, we still need to make new breakthroughs in devices and materials.
Responsible editor: GT