Application of filter MAX274 in power parameter measurement

In the power parameter measurement device, the power parameters come from the power grid with a strong interference source. At the same time, due to the influence of the mutual inductor and the amplifier circuit itself, before the signal enters the A/D converter, the collected signal is mixed with signals of various frequencies, but many of them are unnecessary signals. In practical applications, to obtain accurate FFT operation results, the requirements of the Nyquist sampling theorem must be met to prevent the occurrence of spectrum aliasing. Therefore, the pre-analog low-pass filter plays an irreplaceable role in the power parameter measurement device, and its performance will directly affect the performance of the entire system.

DSP-based power parameter measurement devices have become a hot topic of research, but the pre-analog low-pass filters of most devices are still implemented by operational amplifiers and R, C. Although this type of filter is relatively easy to implement, it is difficult to adjust the parameters, and when the operating frequency is high, the stray capacitance around the components will seriously affect the characteristics of the filter, causing it to deviate from the predetermined working state, and the final effect is not very good. The 8th-order continuous-time filter chip MAX 274 launched by Maxim is currently a more ideal filter chip. This article focuses on the process of implementing a low-pass filter in a power parameter measurement device using MAX 274.

1 Working process of analog signal conditioning circuit

The measuring device is based on the DSP chip TMS320LF2407A, and is mainly composed of analog signal conditioning circuit, A/D conversion circuit, frequency measurement circuit, keyboard display circuit and communication interface circuit. It can accurately measure the three-phase voltage, current and frequency, and calculate the three-phase active power, reactive power, power factor and electric energy through the DSP FFT algorithm. It can also perform 16th harmonic analysis. The device structure is shown in Figure 1. The A/D converter in Figure 1 uses a high-speed multi-channel 14-bit data acquisition chip MAX 125 with a synchronous sample and hold. It has an 8-channel multiplexer and 8 input channels. Its sampling voltage range is -5 to +5 V, and the conversion time is 3μs.

The voltage and current signals in the power system cannot generally be sent directly to the input end of the A/D converter. The 3-way voltage and 3-way current analog signals are first reduced by PT and CT, and then converted into a voltage signal with a rated value of 5 V by the voltage/current converter. After signal conversion and low-pass filtering, they are sent to the A/D converter MAX 125.

The current signal conversion circuit is similar to the voltage signal conversion circuit. The voltage signal conversion circuit is shown in Figure 2. The primary side of the voltage transformer is 100 V, and the secondary side outputs 2 mA current, which is converted into a voltage of -3.5 to +3.5 V by OP07. D1 and D2 provide input limiting protection for the op amp OP07. C2, R2, R3 and R4 are the transformer phase shift compensation circuit. C1, C4 and C5 are used for decoupling and filtering. The op amp TL084 constitutes a voltage follower, which plays the role of impedance matching, as well as buffering, isolation and improving load capacity. The design of the low-pass filter is implemented using MAX 274, and the output of TL084 is connected to the INA pin of MAX 274. The following focuses on the process of designing a low-pass filter using MAX 274.

2 Low-pass filter design process

2.1 Determine filter performance indicators

In order to obtain accurate voltage and current effective values, and considering the amount of DSP calculation, the sampling points per cycle of the power parameter measurement device are set to 32, which not only realizes the accurate measurement of conventional power parameters such as three-phase voltage and current, but also can perform 16th harmonic analysis. Therefore, the cutoff frequency Fc of the filter is 800 Hz, and the sampling frequency Fs is 1.6 kHz. As the most commonly used filter to date, the Butterworth filter has a very flat amplitude-frequency response in the passband, a faster attenuation rate in the transition band than the Basel filter, and no ripples in the stopband range, so it is very suitable as an anti-aliasing filter in the data acquisition system. For the Butterworth filter, the maximum attenuation Amax in the passband is 3 dB, and the minimum attenuation Amin in the stopband is determined by the order.

Designing filters using MAX 274 software mainly involves two steps:

(1) Determine the poles, Q values ​​and zeros based on the filter indicators;

(2) Complete the implementation of the filter on MAX 274 hardware.

2.2 Determining poles, Q values ​​and zeros from filter indicators

This step is mainly based on the performance indicators that the filter needs to achieve, such as maximum attenuation in the passband, minimum attenuation in the stopband, cutoff frequency, sampling frequency, Q value, etc., to quickly calculate the poles, order and Q value of the classic Butterworth, Chebyshev, Bessel or elliptic filter.

(1) After selecting the filter type as Lowpass, enter the Amax, Fc, and Fs values ​​respectively to determine the filter order as 4th order, and the Amin value becomes 24.079 dB;

(2) Press the [V] key to open the window; use the [T] key to select the option to view; [Enter] to display the filter effect diagram and view its amplitude-frequency characteristics, phase-frequency characteristics, and transmission delay characteristics (see Figure 3). It can be seen that its waveform is flat in the passband and drops sharply in the stopband, which meets the actual requirements for filtering characteristics, and the drop rate can reach -80 dB/10 times the frequency. If the effect is not ideal, you must return to the previous step and modify the parameters again until you are satisfied;

(3) Press the [P] key to view its Q value. It is important to note that both the Q and F0 (i.e., Fc) values ​​are within the allowable range. Also note the pitfalls in filter design, i.e., the order of the filter must be an even number, which is determined by the chip structure. [page]

2.3 Filter implementation on MAX 274 hardware

This step is mainly to calculate the filter output gain and programming resistance, and give the circuit schematic. The specific design steps are as follows:

(1) Enter the filter operation interface implemented on the MAX 274 hardware and use [L] to load the result of step 1;

(2) According to the requirements of the subsequent circuit, press [ALT+G] to set the total gain of the last stage output to the first stage input (you can use [U] to switch the gain unit). The final gain of this design is 1.000;

(3) [V] can be used to check whether the gain effect diagram of each filter unit is satisfactory;

(4) If you are satisfied, press the [R] key to configure the resistors. There are three main steps to configure the resistors: In the first window, you can select the connection method of the Fc pin and the chip model. For R2, R3 and R4, if they are greater than 4 MΩ, the Fo and Q values ​​will have large deviations due to the influence of resistor accuracy and parasitic capacitance. The solution is to replace the large-value external resistors with a T-type resistor network. Press the [2], [3], [4] number keys to automatically convert it into a T-type resistor network; then enter the second window to compensate for the Fo/Q value; then enter the third window to standardize the resistor values ​​to obtain the final result.

If an error message appears during the design process, it means that the parameters do not meet the requirements and should be redesigned.

The design parameters of the 4th-order low-pass filter are shown in Table 1.

Since the resistance values ​​listed in Table 1 are calculated by software and may not be available for purchase, a standardized design is required. Most designs use a standardized method of selecting resistance values. Although this method does not affect the filter spectrum too much within a 5% error range, it is still not accurate enough. Therefore, this design uses a method of splitting and combining the calculated resistance values, that is, the calculated resistance values ​​are composed of several actual resistance values, which reduces the error caused by selecting resistance values. For example, the theoretical value of the external resistor R1 of the first group of second-order filters is 445.625 kΩ, but there is no such resistor in reality. Here, the theoretical R1 value is obtained by the series connection of resistors R1A1 (430 kΩ), R1A2 (15 kΩ) and R1A3 (620). The low-pass filter circuit is shown in Figure 4, and the specific values ​​of R1, R2, R3 and R4 of each group of filters are shown in Figure 4.

3 Conclusion

In the power parameter measurement device, there are many interference components from the power grid. This paper discusses the method of designing an anti-aliasing low-pass filter using MAX 274 and gives experimental data.

Practice has proved that the Butterworth low-pass filter designed with the continuous-time integrated filter MAX 274 has a simple structure, is easy to design, has reliable performance, and meets the design requirements, providing an effective solution for the design of anti-aliasing low-pass filters in power parameter measurement devices. Replacing the RC low-pass filter in the power parameter measurement device with the low-pass filter designed with MAX 274 will generate at least 500,000 yuan in economic benefits.