Design of a high-precision portable all-digital oscilloscope

0 Introduction
The oscilloscope is an electronic measuring instrument with a wide range of uses. The rapid development of electronic technology has made various types of electrical signals more and more complex. In engineering applications, there are also high requirements for the real-time sampling rate and waveform capture rate of signals. Digital oscilloscopes have become an indispensable tool for hardware developers and testers in various positions. In response to the needs of the current flexible industrial measurement system, this paper presents a DSP+CPLD solution to achieve the design of a high-precision, highly integrated, portable digital storage oscilloscope. The system uses digital integrated circuits as much as possible, has a simple structure, high reliability of measurement results, and a friendly human-machine interface. It also has the characteristics of high sampling rate, high resolution and low error.

1 System design scheme
The digital oscilloscope mainly consists of two parts: measurement control and display output. At the front end of the measurement circuit, the input signal is processed by the signal conversion circuit into an equivalent signal that can be processed by each secondary unit. It is mainly processed by the comparison circuit to output a positive square wave and an equivalent operational amplifier output signal with a peak-to-peak value of 2.5 V. The square wave signal is used as a counting pulse to trigger the programmable logic device CPLD to achieve the measurement of the frequency value. At the same time, the signal output by the operational amplifier is input to the sample and hold device, and the main controller DSP sends a latch signal to its related pins to realize the function switching of sampling and latching output of the signal to be
tested. When the control terminal is set to "1", it is a latch output. At this time, the output signal can be used for data acquisition by the A/D converter; when it is set to "0", the signal is acquired. The data collected by the A/D is sent to the DSP, and then the DSP first stores the external data in the external memory, and then analyzes it. Finally, the processed data is sent to the LCD display in the form of a data packet through the RS 422 standard interface for display output. The system principle block diagram is shown in Figure 1.

2 Hardware Design
The hardware part of the system consists of signal input conversion circuit, sampling and holding circuit, main control circuit, intelligent terminal equipment and other parts.

2.1 Signal input conversion and sampling and holding
circuit The signal input conversion circuit is mainly used to realize the equivalent conversion of signals. The design uses high-speed OP37 for signal conversion and sampling and holding. It is an important component of the data acquisition system and acts as an isolation buffer for the signal. If the analog signal with a high change speed is to be converted into A/D, the conversion accuracy requirement is relatively high. In order to prevent the signal from changing during the A/D conversion process, the S/H circuit must be used. The S/H circuit cooperates with the A/D to eliminate the output pulsation of the A/D and realize multi-channel sampling control through MUX.

The high-performance single-chip sample/hold device LF398 is used here, which has high DC accuracy, fast sampling time and low drop rate. The dynamic performance and holding performance of the device can be optimized through the appropriate external holding capacitor. The signal conditioning circuit is shown in Figure 2. 

2.2 Frequency measurement
circuit The frequency measurement of the signal by the oscilloscope is designed based on the principle of equal precision frequency meter. It is completed by the programmable logic device EPF10K50V, and the standard frequency signal of 100 MHz directly enters EPF10K50 V. The device uses the square wave pulse output by the signal input conversion circuit as the clock input signal of the counter, counts with the standard 100 MHz, and finally calculates the frequency of the input signal.
Through the graphical method and VHDL language to program EPF10K50V, in this design, CPLD completes the measurement of signal frequency. The principle of frequency measurement is as follows: In the unit measurement time Tp, the count value of the measured signal is Nx, and the count value of the standard signal is Ns. Based on the known standard frequency fs, the frequency value fx of the measured signal satisfies:

2.3 Amplitude signal acquisition
In order to meet the acquisition of high-frequency signals, AD7667 launched by ADI is selected to realize the amplitude measurement of the measured signal. AD7667 is a 16-bit A/D conversion chip with an internal 2.5 V reference voltage, an operating range of 0 to 2.5 V, an LSB less than ±2 b, a conversion rate of 800 Kb/s, a conversion time less than 1 μs, and a single +5 V power supply. The signal conversion circuit converts the measured signal into an effective value within the working range for accurate measurement.

2.4 Design of human-computer interaction
The display and command input of the oscilloscope are realized by the intelligent terminal device LJD-ZN-3200K. LJD-ZN-3200K is an intelligent graphical interface output device integrating input and output with a resolution of 640×480, which can meet the requirements of system design. The device terminal communicates with the main controller through a serial interface to complete data transmission.

Load the designed graphical interface into the storage unit of the intelligent terminal, and then identify the coordinate value according to the setting to realize touch control input. The oscilloscope has a total of 9 function keys, namely: 3 vertical discrimination selection buttons for vertical sensitivity selection; 3 horizontal discrimination selection buttons for horizontal scanning speed selection; sampling mode switching button for selecting real-time sampling and equivalent sampling; waveform storage button and waveform call-out button for current waveform acquisition, storage and call-out; single trigger button, which can perform single acquisition and storage of signals that meet the trigger conditions.

3 Signal acquisition and processing analysis
3.1 Signal acquisition principle
When measuring different frequency signals, selecting a reasonable sampling method will directly affect the measurement accuracy of the system. In digital signal analysis technology, there are two commonly used signal sampling methods: real-time sampling and equivalent sampling.

Real-time sampling is usually performed at equal time intervals, and its highest sampling frequency is the Nyquist limit frequency. Its characteristic is that the duration of the pulse sequence obtained by sampling a waveform is equal to the actual time experienced by the input signal, so the spectrum of the sampled signal is wider than the original signal. In this design, the A/D conversion device frequency used is 400 kHz. According to the sampling characteristics, it can be calculated that the digital oscilloscope can sample and output input signals not greater than 50 MHz.

Equivalent Sampling refers to a data acquisition method that takes advantage of the time-domain repetitive characteristics of periodic signals, performs multiple samplings at lower sampling rates in different time periods, and then combines these low-sampling rate samples into high-sampling rate data samples to truly reconstruct the original signal waveform. It uses the periodicity of the signal to reduce the pressure on the high-speed sampling circuit at the cost of increasing the acquisition time, and restores the original signal through reorganization.

This paper adopts the extraction of equivalent sampling time sampling, which uses the special relationship between the signal repetition frequency fi and the sampling rate fs to increase the equivalent sampling rate by D times.

First, the repetition frequency fi of the input signal is appropriately selected, and the signal waveform of D cycles is sampled. Then the recorded data is rearranged and combined through a simple algorithm to obtain a complete input signal waveform. In this way, the equivalent sampling rate is D times the actual sampling rate.

In actual implementation, the selection of D depends on the required equivalent sampling rate fe, so that fe=Dfs. L is the number of actual sampling points in a single cycle, L=int(M/D), and M is the sum of the recorded sampling data. The repetition frequency of the output signal is:

The method of extracting equivalent time sampling can improve the sampling rate, but requires the repetition frequency fi of the input signal to be controlled with accuracy, while the equivalent sampling rate is Dfs, which has nothing to do with the input signal. When the repetition frequency of the input signal deviates from the value given in formula (2), the maximum time deviation of the equivalent sampling rate is:

It is equivalent to widening the frequency band. At this time, the width of the frequency band has almost nothing to do with the speed of A/D conversion and the speed of the microprocessor. This method combined with the designed digital oscilloscope can easily measure the frequency and amplitude of high-frequency signals.

Finally, the sampled data is stored and then analyzed uniformly to reproduce the function curve of the signal and calculate the amplitude.

Since the voltage signal sampling and analysis in the design process adopts an equivalent method, the discrete sequence with time as the independent variable is collected. These sampled data reflect the change process of the measured parameters, but with a certain degree of error, which will inevitably cause the distortion of the collected data. In order to avoid the interference of values ​​within the error allowable range on the measurement results, the software is used to correct the data by curve fitting the measurement results to ensure the relative accuracy of the measurement results.