Use signal averaging technology to eliminate noise interference and improve the accuracy of repeated signal sampling

　　Normally, in the test of analog signals, the collected data is often mixed with some unnecessary and random content. These data are caused by surrounding interference or test errors. We call them random noise. This noise may affect our target signal, that is, the data we need to collect. The use of signal averaging technology can reduce the impact of random noise, improve the signal-to-noise ratio (SNR), and minimize the impact on the target signal, thereby improving the accuracy and dynamic range of data collection. Specifically, ADLINK 's Data Average Mode (DAM) provides such a high-level signal averaging function.

　　1  Advantages of FPGA

　　There are two solutions to eliminate noise: one is a DSP -based solution, and the other is an FPGA -based solution. When the sampling rate required by the measurement test system is less than a few kilohertz, a DSP- based solution is usually used. However, when the sampling rate required by the measurement test system is relatively high, a solution based on FPGA (Field-programmable Gate Array) is a better choice. Because DSP is a method based on code or instructions, it inevitably involves the system architecture and core processor, which will lead to excessive occupation of system resources and increase processing time. FPGA provides multiple gate arrays (Gates) and memory blocks (RAM Blocks), which can form multipliers (Multipliers), registers (Registers) and other logic units, thereby achieving fast calculations. Therefore, many high-performance applications currently use FPGA-based solutions.

　　Most of ADLINK 's high-speed digitizers offer onboard FPGA functionality, which is ideal for real-time acquisition applications that require high speed and bandwidth. FPGA-based boards support onboard real-time data processing functions, such as signal averaging, which can reduce the signal averaging tasks running on the host. And in terms of processing speed, performing signal averaging on the FPGA is much faster than executing it on the host, and it does not take up any resources from the host CPU.

　　2 Eliminate noise interference and improve the accuracy of repeated signal sampling

　　In the test of analog signals, the collected data often contains some noise (such as harmonic components, modulation sidebands, etc.), which may mask the signal of interest or its harmonic components, modulation sidebands, etc. As we all know, since the expected value of random noise is zero, signal averaging technology becomes a simple and effective solution to eliminate random noise in periodic or repetitive signals. The signal averaging mode of PCIe-9852 works according to the following principle. In the re-trigger mode, the N sampling points are repeatedly collected R times, and the trigger source is an external digital or analog signal. Each collected data will be stored in the same onboard buffer and automatically accumulated by the FPGA without software intervention. After the R re-trigger is completed, the FPGA divides the accumulated data by R to obtain an average and sends the average to the host computer. All data (including the average) will be represented by traces one by one, thereby reducing noise and making the data closer to the target data. The following figure shows the comparison results of the data averaging mode of PCIe-9852 in data acquisition and noise reduction.

　　One of the biggest advantages of the PCIe-9852 data averaging mode is that it saves a lot of storage space. Using software to average signals on the host requires memory space to be reserved for N (sample points per trace) x R (number of retrigger times), while the PCIe-9852 data averaging mode only requires memory space to be reserved for N samples, because each sample will be stored in the same onboard buffer and will be handed over to the host after averaging. Since the data becomes smaller, the data transmission time of the signal averaging mode is less.

　　At the same time, compared with using software to process signal averaging, the signal averaging mode can reduce the CPU load. Since the signal averaging mode is based on the FPGA , the entire calculation process is completed independently of the CPU.

The table below compares the test results 　　of different solutions (one software-based and one FPGA- based). The PCIe-9852 acquires a 2.0Vpp, 200kHz continuous sine wave with 100 retrigger events. The PCIe-9852 has a sampling rate of 200MS/s and the total amount of data acquired each time is 100kS. The results show that the signal averaging mode is better than the software-based solution. The advantage of the signal averaging mode will be more obvious when the amount of data and the number of retrigger times increase, or when the test platform uses a low-power processor.

　　The following block diagram shows the processing rules using the signal averaging mode and software-based signal averaging.

　　3 Successful Applications of Distributed Fiber Temperature Measurement (DTS)

　　Distributed fiber temperature measurement (DTS) is a very typical application that benefits from signal averaging technology. DTS uses an optical time-domain reflectometer (OTDR)-based measuring instrument to measure temperature through optical fiber, thus replacing traditional thermocouples or thermistors. In addition to obtaining accurate temperature data, the DTS solution can also save a lot of costs. By using a pulsed laser coupler, DTS can measure optical fibers up to 30km long. When the temperature in a specific area changes, the wavelength of the light changes and propagates in the optical fiber in the form of backscattered light. Accurate temperature change data can be obtained by accurately measuring the backscattered light.

　　Such a high-speed repetitive signal carries a prohibitive level of noise, and a high-speed, high-precision digitizer is an ideal solution for processing such a signal, and for such an application, the signal averaging function is a very important factor. ADLINK 's PCIe-9852 is a 2-channel 200MS/s 14-bit high-speed digitizer that is very suitable for DTS applications. The 2 analog inputs of the PCIe-9852 can receive Stokes and anti-Stokes light synchronously, and its high-precision sampling rate can easily meet the detection distance of more than 30km. In addition, the onboard signal averaging provided by the data averaging mode can extract extremely small detection data from complex environments.

　　Tips: High-speed digitizers, also known as computer-based oscilloscopes, have open architectures and flexible software, offering several advantages over traditional standalone oscilloscopes. High-speed digitizers can perform measurements on instruments such as oscilloscopes, spectrum analyzers, and transient recorders, but at a higher rate than traditional instruments. With resolutions ranging from 8 to 21 bits, digitizers have low-jitter clocks that improve measurement accuracy and stability.

　　In addition, high-speed digitizers feature advanced timing and synchronization hardware, allowing you to easily develop high-channel-count systems and integrate them with other instruments, such as arbitrary waveform generators, digital I/O, and image acquisition devices.