Design of a low-cost noise meter

Noise pollution, gas pollution and solid matter pollution have been called the three major pollutions in the world today. Excessive noise can damage people's health, such as neurasthenia and neurosis. Noise monitoring is an important way to improve people's living standards. The hearing threshold of the human ear is generally 20μPa, and the pain threshold is generally 200 Pa, which differs by 107 times. Such a wide sound pressure range is not easy to measure, and the human ear has nonlinear characteristics in distinguishing the relative change of sound pressure. The inverse curve of the 40-square equal loudness curve is usually used to weight and correct the sound pressure level, that is, the A-weighted network is used to measure the A sound level, and its unit is dB. At present, foreign companies have been occupying the domestic high-precision noise meter market, and its price is generally high, which seriously restricts the widespread use of noise meters. However, environmental noise measurement equipment used in daily life does not require complex and very accurate measuring instruments, so it is very important to design a low-cost, portable noise meter that can meet daily life use.

1 Circuit Design
In acoustics, the sound pressure level Lp is often used to reflect the change of sound pressure. The sound pressure level of sound pressure P is expressed as:
LP=20lg(P/P0)
where: the reference quantity P0 is in μPa. When P=P0, Lp=0 dB;
when P=107P0, LP=140 dB.
The system schematic diagram is shown in Figure 1. An electret microphone is used in the signal acquisition and amplification circuit, and an inverting amplifier is used to amplify the signal to a suitable processing range; the A-weighted network well meets the sensitivity of the human ear to sounds of different frequencies; the AD536 chip can easily obtain the effective value and its corresponding decibel value, avoiding the use of a single-chip microcomputer for large-volume logarithmic processing and improving the response speed; the adjustment circuit can determine the required different measurement zero points; finally, the digital tube is used to display the results in real time, with a fast and slow display function.

1.1 Noise signal acquisition and amplification circuit implementation
The electret microphone is connected by source grounding and drain output, with wide dynamic range and high sensitivity. Under the voltage drive of VCC=9 V, the dynamic range can reach -2~+2V. After debugging, when the microphone voltage V1=VCC/2, the microphone sensitivity reaches the maximum. In this circuit, the drain load resistor R1=20 kΩ is taken, and the circuit diagram is shown in Figure 2. After the sensor converts the noise signal into an electrical signal, it continues to amplify it. The amplifier circuit uses an operational amplifier with a bandwidth greater than 2 MHz.

Circuit parameters: R2=20 kΩ, R3=51 kΩ, C=0.22μF;
Gain: A=R3/R2=2.55;
Approximate range of output voltage amplitude: 0~5 V.
1.2 Implementation of filter circuit scheme
According to the loudness perception of different frequencies of human ears, in noise measurement, A-weighting network is often used to measure A-level. Table 1 shows the relationship between the center frequency of the octave band and the A-level correction amount.
The weighting network consists of two parts: passive high-pass filtering and active low-pass filtering. The parameters of related peripheral components are carefully calculated and debugged to make its amplitude-frequency characteristics close to the A-weighting curve. The circuit diagram is shown in Figure 3.

[page]

The corresponding parameters of the high-pass filter are calculated as follows:
Take C1 = 10μF, use the equation to obtain the rear-stage resistor R2 = 100Ω, and the front-stage resistor R1 = 100 kΩ to obtain a higher input impedance. The corresponding parameters of the second-order Butterworth active filter circuit are calculated as follows: (1) Select the capacitance of capacitor C2 and calculate the resistance of resistors R3 and R4. Usually, capacitors C2 and C3 should be below the microfarad level, and the values ​​of resistors R3 and R4 are generally within a few hundred kilo-ohms. Set C2 = C3 = C = 1 000 pF, so:

Considering the -3dB cutoff frequency ωH=ωn, then:

Let: S = s/ωH, according to the case of n = 2 in the Butterworth polynomial

Taking into account, we get:

(2) Calculate the resistance of R5 and R6
Considering that the external resistors at the two input terminals of the op amp must meet the balance condition, that is, R5||R5=R3+R4=16kΩ, we can calculate R5=25kΩ and R6=43kΩ.
Since the filter performance is sensitive to the error of the components, stable and precise resistors and capacitors should be selected in the circuit.
1.3 Implementation of the effective value and logarithmic circuit scheme
The effective value circuit uses the true effective value/DC converter AD536A. The performance of AD536A is equivalent to or even better than that of hybrid or analog-to-digital devices, but its price is much lower and its connection is very simple. Only an external capacitor is needed to set the average time constant. Here, the input time is taken as 0.25s. When the time constant of the low-pass filter formed by CAV and the input resistance of the signal is greater than the period of the signal, the effective value and logarithmic value can be well calculated. The
logarithmic output is led out from terminal 5, and the voltage at this point is proportional to logVin. The voltage can be buffered by an emitter follower and can be shifted by an external adjustment circuit. The voltage adjustment circuit is composed of the voltage regulator chip AA580. An amplifier with adjustable gain is cascaded at the output of the buffer to change the step size.
1.4 Adjustment and display circuit solution
1.4.1 Adjustment circuit
The output voltage of the effective value and logarithmic circuit is proportional to -log Vin. Through the step size adjustment of the buffer, the first term on the right side of the equation Lp=20lg(P/P0)=201g P-20lg P0 can be obtained. In the design, the reference voltage, that is, 0 decibel value, only needs to change the second term on the right side of the equation. Therefore, a compensation circuit based on a subtraction circuit is added after the logarithmic circuit, which can be used for zero adjustment. At the same time, the relative reference zero point can be set by adjusting the compensation voltage to measure the relative value of the decibel change. By changing the size of the reference voltage, the voltage signal representing the decibel value of the input display circuit can be adjusted.
The specific derivation formula is as follows:

Therefore, we can get Vout=Vin-Va, and Vin is the voltage signal representing the decibel value. By changing the size of the potentiometer RX to adjust the size of the reference voltage Va, the voltage signal Vout representing the decibel value of the input display circuit can be adjusted. Va is the input voltage of R3, Vb is the reverse input voltage of the op amp, Vout is the same direction input voltage of the op amp, Vin is the input voltage, and Vout is the output voltage.

[page]

1.4.2 Hold Circuit
In order to continuously display the current maximum decibel value within the measurement period, a peak detection hold circuit must be used, as shown in Figure 4.

Here, a diode capacitor detection circuit is used, and the capacitor is 470μF. Its charging time constant is small and its discharging time constant is large. Of course, the high input resistance and low output resistance of the op amp can also be used for optimization. When the time weighting switch is set to "slow", the instrument can measure the maximum sound level value within a period of time; when it is set to "fast", the environmental noise can be measured in real time.
1.4.3 Display circuit
The ICL7107 chip is used as the head driver. It is simple and economical and requires a bipolar power supply voltage drive. It contains a bit digital A/D converter, which can directly drive the LED digital tube. It has a reference voltage Vref, an independent analog switch, logic control, display drive, automatic zero adjustment function, etc. Here, its A/D conversion function and the driving function of the LED digital tube are applied. Its display principle is as follows:

In the formula: Vref is adjusted by pin 36, and its dynamic range is (-Vref, +Vref). In this design, bit 10 is taken as the lowest bit, so the range is reduced by 10 times.
The 5 V voltage required in this design is obtained by using the 7107 chip to generate an oscillation signal, and then connecting a 4069 inverter in series to its pin 38, and then the signal is output through two 4μF capacitors and two 1N4148 diodes to form a voltage doubler rectifier circuit.

2 Test experiment
The most important part of the noise meter is the A-weighting network design. The system is implemented by a passive high-pass filter and an active low-pass filter. The type II passive high-pass filter circuit is used to attenuate the input voltage signal below 1 kHz to achieve a response similar to that of type A weighting. The low-pass filter circuit filters out signals with frequencies higher than 20 kHz. The amplitude-frequency response of the second-order Butterworth active filter circuit has the maximum flatness in the passband.
The amplitude-frequency curve obtained by simulating the A-weighting network with Workhench software is shown in Figure 5. As can be seen from the figure, it has a large attenuation for signals below 100 Hz, which is in line with the characteristics of the human ear's sensitivity to noise signals, and can also filter out signals above 20 kHz.

The comparison of the amplitude-frequency characteristics of the test filter and the requirements of A-weighting is shown in Figure 5. The input voltage of the test is 1V, with a 1kHz signal as the reference.

3 Conclusion
Humans have different sensitivities to sounds of different frequencies, with the lowest sensitivity being in the low frequency band and the highest sensitivity being around 3 kHz. Therefore, noise measurement does not measure the actual sound pressure, but rather attenuation. Common methods include A-weighting, B-weighting, C-weighting, and D-weighting. A-weighting is used in this article. In fact, a noise meter should use different weighting networks for different loudness bands.