NS software radio application based on 12-bit 3.6 GS/s analog-to-digital converter series

Summary: In view of the characteristics and limitations of the traditional hardware radio architecture, the application requirements of the analog-to-digital converter in the software radio design are analyzed, and the new product series of 12-bit giga-level sampling converters and the matching differential amplifiers and clock solutions are introduced. plan.
Key words: Hardware radio; software radio; ADC; differential amplifier; clock source

As the name suggests, software radio (SDR) means that the key features of radio equipment are determined by software rather than hardware. Such an architecture has many advantages: low complexity, smaller size, lower power consumption, convenient upgrades, and reduced hardware development and repeated design costs. But the software radio architecture requires very high-performance analog-to-digital converters, amplifiers, and precise clocks. This article will introduce the new product series of 12-bit Gigabit sampling converters. This series of converters integrates many advantages of software radio and has been widely used in military radar, communications, cable set-top boxes and other fields. This article will also introduce the matching differential amplifier and clock solutions.
Characteristics and limitations of traditional hardware radio architecture
The well-known receiver architectures such as heterodyne, homodyne, and low-IF, each has its own unique advantages and disadvantages. Their common feature is that whether it is a double-conversion or a single-conversion architecture, the RF signal is down-converted into a lower frequency and easier-to-manage IF signal through mixing or down-conversion. Due to the limitations of the existing analog-to-digital conversion technology, it is usually necessary to convert the analog radio frequency signal into a low intermediate frequency or baseband signal. Basically, the existing analog-to-digital converters are not enough to meet the requirements of many software radio applications in terms of sampling frequency, analog input bandwidth, and input sampling broadband noise tracking and holding. The superheterodyne architecture is shown in Figure 1. Regardless of whether there is frequency conversion, dual-stage (heterodyne) or single-stage (homodyne) architectures require a lot of challenging analog signal conditioning. The hardware radio (HDR) architecture requires a high-performance mixer, which requires a filter with excellent amplitude/phase matching ability, low local oscillator leakage level, high Q value and low insertion loss (preselection, image rejection) , Anti-mixing). This type of device has many troublesome effects: leakage, DC bias errors, flicker noise, I/Q mismatch, and even harmonic distortion. RF and analog system design is challenging-standing waves caused by impedance, harmonic distortion and reflections, I/Q mismatches, and equipment leakage are difficult to detect, and the effects are difficult to quantify. In addition, due to the emphasis on training “digital” engineers in educational institutions, the number of radio frequency and analog experts in the industry is decreasing.

NS software radio application based on 12-bit 3.6 GS/s analog-to-digital converter series

In addition to being technically difficult to implement, hardware radio architecture has some obvious shortcomings: dense analog design is highly complex, requiring a lot of power and circuit board area; in order to reduce electromagnetic interference (EMI), it is usually necessary to increase radio frequency shielding , And this has further increased the volume of the entire system; high energy consumption will inevitably lead to heat dissipation problems and cooling requirements; high cost, and the cost increases exponentially as the number of channels increases; fixed frequency points will cause rigidity; hardware systems The parameters (number of channels, channel bandwidth, etc.) are fixed, so system modification and redesign require a lot of research and development work and higher costs.
Comparison with software radio
The concept of software radio is not new. Although the term “software radio” was proposed by Joseph Mitola in 1991 and published a monograph in 1992, in fact, the defense department has been using this concept since the 1970s. The military’s goal is to develop a flexible and programmable radio architecture that can easily adapt to changing ground conditions. Basically, the characteristics of radio should be defined by software rather than hardware. Wireless infrastructure developers quickly realized the potential of software radio in reducing hardware development costs and increasing revenue. A software programmable base station can easily be adjusted to support emerging industry standards (such as WiMax, UMTS, MC_GSM), without the need to upgrade hardware or send maintenance engineers to the site. This requires that the characteristics of the radio be defined in the digital domain rather than in the analog domain. In order to realize this method, the digital module must be close to the antenna. Only minimal analog signal conditioning is required before analog-to-digital conversion. Figure 2 is only a simplified diagram, but generally speaking, only a preselection filter (used to eliminate out-of-band signal energy), a low-noise amplifier (LNA) and a differential variable gain amplifier (VGA) are required. Analog-to-digital converters often need an accurate clock source-because the analog-to-digital converter directly samples the radio frequency signal, the requirements for the clock source are more stringent than before. With this method, the entire signal band can be digitized, and complex non-linear mixers, local oscillators and filters (intermediate frequency and baseband) are no longer needed. At the same time, the analog-to-digital converter here also puts forward strict requirements on the front-end devices.

NS software radio application based on 12-bit 3.6 GS/s analog-to-digital converter series

Analog-to-digital converter requirements:
· Gigahertz sampling rate and Nyquist bandwidth;
· Low noise floor;
· High noise power ratio (NPR) and intermodulation distortion;
· Low power consumption;
· Recommended analog features: single power rail, automatic calibration, adjustable input gain and offset;
Recommended digital features: multi-chip synchronization function, programmable data interface (data rate, data acquisition clock, data/data clock phase relationship) and test mode.
Clock requirements:
· Ultra-low noise floor-sub-picosecond RMS jitter;
· Excellent parasitic noise performance;
· Recommended features: high integration, programmable output frequency and power.
A/D converter drive circuit requirements:
· The wide bandwidth is equal to the input bandwidth of the analog-to-digital converter;
· Flat out-of-band gain;
· Low noise and distortion;
· Recommended features: gain and common mode voltage control.
The advantages of software radio reflect the disadvantages of hardware radio. Fewer analog components means lower simulation complexity, and the simplification of RF signal processing means less RF shielding. This makes the design smaller, more compact structure, and lower power. It can save hardware and development costs immediately, but the main advantage comes from the inherent flexibility of software radio. This is indeed a significant advantage compared to hardware radios. Software programmability allows to change or completely change the radio specifications from a remote location, and such changes will not cause any changes to the hardware. By providing compatibility with the new 3G or 4G standards, network operators can upgrade communication base stations. Cable or satellite TV service providers can directly upgrade the client set-top box (STB) without adding additional tuners. Radar system manufacturers can also benefit from digital programmable frequency selection technology. Services are improved and operators’ costs are reduced, thereby benefiting customers.
It is worth mentioning that the rise of software radio does not mean the demise of analog systems. On the contrary, analog systems such as ultra-high-performance amplifiers, frequency synthesizers, and clock regulators have been widely used in the design of software radios.
Software radio component solutions
As shown in Figure 3, the ADC12D1800 is composed of two channels, the sampling rate is as high as 3.6 GS/s when operating in independent channels, and the sampling rate is as high as 1.8 GS/s when operating in dual-channel crossover. The device is 3.6 times faster than any existing 12-bit device when operating at 3.6 GS/s. The analog-to-digital converter has an analog input bandwidth of 2.8 GHz, a floor noise dynamic performance of -147 dBm/Hz, a noise power ratio (NPR) of 52 dB, and an intermodulation distortion (IMD) of -61 dBFS. Such specifications can meet Many software radio applications require.

ADC12D1800 is powered by a single 1.9 V power supply and is manufactured by a 0.18 μm pure CMOS process. The power per channel is only 2.05 W. Each channel of the device has multi-chip synchronization, programmable gain and bias circuits. Even when the input frequency exceeds 2 GHz, the internal track/hold amplifier and extended self-calibration mechanism enable the system to obtain a flat response to all dynamic parameters, while the bit error rate can be reduced to an incredible 10- 18. In addition to good performance in terms of noise floor, noise power ratio (NPR) and intermodulation distortion (IMD), the ADC12D1800 also has a signal-to-noise ratio (SNR) of 57.8 dB and a spurious-free dynamic range of 67 dBc at 125 MHz ( SFDR) and 9.2 effective digits. Low-voltage differential signaling (LVDS) output can be configured as 1:1 or 1:2 demultiplexing mode. The test mode can be used for system debugging.
LMX2541 is an ultra-low noise frequency synthesizer, which integrates a high-performance delta-sigma fractional-N phase-locked loop (PLL), a fully integrated voltage-controlled oscillator (VCO) with a tank circuit, and an optional frequency divider Device. The phase-locked loop (PLL) can generate frequencies from 31.6 MHz to 4 GHz, while the normalized floor noise reaches an unprecedented -225 dBc/Hz. When matched with a high-quality reference oscillator, the LMX2541 can generate a very stable low-noise signal. So LMX2541 becomes the ideal clock source of ADC12D1800.
The LMX6554 has a unity gain small signal bandwidth of 2.8 GHz, and can work in an environment with a gain greater than 1 without sacrificing response flatness, bandwidth, harmonic distortion or output noise performance. It has a gain flatness of 0.1 dB at 830 MHz, a noise figure of 8 dB and an intermodulation distortion of -99 dB at 150 MHz. For DC-coupled applications, LMH6554 has a common-mode output voltage pin that is used to correctly set the common-mode voltage of the ADC121X00 series.
Many systems can benefit from software radio architecture, such as test equipment (spectrum analyzer, digital oscilloscope), radar, communications (satellite, microwave backhaul, optical link), multi-channel set-top box (STB), signal intelligence, and laser detection and measurement Distance (LIDAR). These systems used hardware radio architectures in the past. For the above-mentioned systems, software radio can reduce the number of components and the cost of the bill of materials, reduce the scale of the solution and the power consumption, while providing unlimited flexibility and programmability. Upgrading equipment by reusing common analog front-end modules can also help reduce research and development costs.
For more details about software radio and the above products, please refer to: http://www.national.com/analog/adc/ultra_high_speed_adc.

The Links:   2DI100Z-100 NL6448BC26-17

Related Posts

Leave a Reply

Your email address will not be published. Required fields are marked *