Ac Backward Centrifugal Fan ,Backward Curved Blower,Backward Inclined Centrifugal Blower,Reverse Incline Fan Hangzhou Jinjiu Electric Appliance Co Ltd. , https://www.jinjiufanmotor.com
ã€Keywords】 Weigh indicator noise I. Overview The word “noise†is defined in the “modern electronic science and technology dictionary†as “irregular electrical disturbances that interfere with useful signals in circuits and systems.†The performance of noise in non-automatic weighing instruments is the jumping of weighing indications, which makes it impossible for operators to judge the accurate measurement results. The performance in automatic weighing instruments is the error of quantitative values ​​in quantitative scales and the measurement values ​​in sorting scales. Error and so on. In short, noise is a harmful phenomenon in the weighing system, and its influence should be eliminated or reduced as much as possible. Weighing system noise sources are internally generated by the weighing indicator and external to the indicator, such as mechanical system vibration, air flow disturbances, industrial interference and electromagnetic interference due to electrical disturbances, noise from load cells and signal transmission lines, etc. of.
This article focuses on how to correctly define and evaluate the noise of the analog signal input weighing indicator, and also discusses how to reduce the noise of the weighing indicator. For the digital signal input weighing indicator, theoretically it will not produce noise itself, in fact the digital operation process will also produce a certain error, in the dynamic conditions, it can also be seen as a noise, but if the algorithm is appropriate The calculation accuracy is high enough and its influence can be ignored. Therefore, this article has not been analyzed.
Second, the impact of noise on the measurement performance of the weighing system The noise of the weighing system causes the output value of the weighing signal to fluctuate irregularly within a certain range. When the zero setting operation is performed, taking the current output value as the zero point will make the zero point of the weighing instrument. There is a certain amount of deviation. The output value here refers to the data flow processed by the weighing signal data in the weighing indicator. According to OIML R76 "non-automatic weighing instrument" requirements, after zeroing the instrument, the zero deviation should be less than 0.25e. According to the error distribution coefficient Pi = 0.5 of the weighing indicator zero deviation, the peak-to-peak noise value of the weighing indicator output value should be less than 0.25e. When carrying out the discriminating force test of the weighing apparatus, under the specified load, first add 10 small weights of 0.1e to the carrier, the weighing apparatus shows the value of I, and then remove the small weights one by one until the value I shows a definite reduction. An actual division value is changed to Id, a small weight is put back, and the load corresponding to 1.4d is gently placed on the carrier. The weighing indicator should clearly indicate the original value. The result of adding an actual division value is I+d. Obviously, in the discriminating power test, the peak-to-peak value of the scale indicating value can't be more than 0.4d (here d=e). According to the error distribution coefficient Pi=0.5, the peak-to-peak noise value of the weighing indicator output value should be less than 0.2e. .
It can be seen that if the weight of the weighing indicator is too high, the weighing instrument will overshoot and the noise of the weighing indicator must be controlled.
Third, the noise sources and their spectral characteristics We all know that the output signal of the strain sensor is a weak signal of the millivolt level, especially some large scales using multiple sensors, the output of the sensor is only a few mV/V at the maximum weighing. The signal voltage of each grading value e is below a few microvolts and even below 1 microvolt, so the weighing indicator is very sensitive to noise in the circuit. The noise of the weighing indicator is mainly generated by the preceding stage, especially the first-stage operational amplifier, A/D converter, etc. Therefore, to reduce the noise of the weighing indicator, the key is to select a low-noise operation amplifier and A/. D conversion device.
The noise in the weighing indicator mainly comes from the thermal noise generated by the thermal motion of the electrons in the resistor and the thermal noise, shot noise, distribution noise, and 1/f noise generated by the irregular movement of the charged particles in the transistor. Some of these noise sources are white noise, such as thermal noise, shot noise, and their spectral density has nothing to do with the frequency; some are colored noise, such as 1/f noise is pink noise, and its spectral density is larger at low frequency; The total spectral density of noise in the weighing indicator circuit is shown in the figure below:
Spectral Density of Noise Voltage in a Circuit In the device data sheet, the noise voltage is sometimes given as a rms value, sometimes as a peak-to-peak value, and sometimes as nV/√Hz. Available noise voltage peak-to-peak value ≈ 6.6 × rms noise voltage rms value. The use of these data must pay attention to their relationship with the passband, and the actual noise should be the integral of the noise spectral density in the passband.
Fourth, the evaluation method In GB "weighing display" provides the indicator of the noise indicator requirements. The author believes that because the requirements of zero error, discriminating ability, and other requirements in OIML R76 have implied the limitation of noise, there is no need to add additional requirements and tests for noise.
From the spectrum distribution of the noise, it can be seen that the total noise is closely related to the cut-off frequency of the pass band of the weighing indicator or the output of the weighing data, the noise increases with the passband, and the noise can be suppressed with a narrow passband. If a weighing indicator is used for static weighing, the cut-off frequency can be set below zero-few hertz for dynamic weighing. For example, high-speed sorting scales for sorting 200 objects per minute require a cutoff frequency of tens of Even hundreds of hertz or more. The same instrument, used in different occasions, due to the difference of passband width hundreds of times, the output noise amplitude can be more than 10 times. Therefore, for weighing indicators used in automatic weighing machines, it is not meaningful to ignore the noise index value of the passband width.
OIMLR76 does not specify the passband width of the weighing indicator. Only one requirement is related to the cutoff frequency. R76 requires that when the load on the carrier changes, the original value should not be held longer than 1 second. Note that only the maximum time limit for the original value retention is specified here, and the occurrence time of the accurate indication value is not specified. This leaves more space for the passband width of the weighing indicator. Designers can reduce the passband cutoff frequency to reduce system noise. The software is designed to allow the indicator to use the lowest possible passband limit while satisfying the above R76 requirements. ADI's AD7730 A/D conversion device gives such an example. The AD7730 has two stages of digital filters: the first stage is a sinc3 low-pass filter, and the second stage is a FIR filter (finite impulse response digital filter).
The first filter is always on, and the second filter can choose three different operating modes. The first is an ordinary FIR filter mode. The filtering effect is good but the response to the step signal is slow. The second is the fast mode. In this mode, the output of the FIR filter and the output of the first stage filter are When the difference is more than 1% FS, the FIR filter automatically enters the moving average filter state. The average number of calculated data automatically increases from 2 or 1 to 4, 8, and 16, and the filter automatically returns to the FIR filter after reaching 16. Device status, this feature speeds up the A/D device's response speed without increasing the output noise when the input signal is stable, the contrast between the normal mode and the fast mode response speed; the third is the SKIP mode, in this mode Next, the second-stage filter is bypassed. Although the output response is fast, the output noise is about three times larger. Weighing indicator software of some domestic weighing and manufacturing companies also adopts an algorithm similar to the second mode, which effectively reduces the noise when the balance is stable, and enables the indicator to meet the requirements of the response speed of the R76.
Comparison of normal mode and fast mode response speeds For automatic weighing devices, most of them are weighed dynamically. The contradiction between the noise of the weighing indicator and the response speed (or the cut-off frequency of the output data) is prominent, and only from the meter's The indication of the relationship between the two parameters is not well quantified, and these two parameters of the weighing indicator have an important influence on the dynamic performance of the automatic weighing instrument such as the quantitative scale and the sorting scale. Standards and verification procedures for existing weighing indicators are also a blank for this. In order to make the manufacturer of automatic weighing instrument have a quantitative evaluation method for the frequency and noise performance of the weighing indicator, and to increase the performance of the automatic weighing instrument, I tried to propose the following method to solve this problem, in order to play a role in attracting attention. . This method requires that the weighing indicator has an interface that can continuously output weighing data. This should not be a big problem for existing weighing indicators, especially weighing indicators for automatic weighing instruments, but it may be necessary to modify the indicator. The software allows the data rate output by the interface to match the speed of the internal data processing of the meter and has a sufficiently high data accuracy, ie, resolution. In addition, there is a need for a PC and a PC-controlled weighing signal generator, such as the XY1 sensor simulator produced by Shanghai Yaohua Weighing System Co., Ltd. The test system is connected as shown in Fig. 3.
Test scheme for noise and response speed An RS-232C serial interface of the computer is connected to the XY1 sensor simulator, another RS-232C interface collects the weighing data output from the weighing indicator, and the analog signal output of the sensor simulator The sensor interface of the weighing indicator is connected. Program a small program that is executed in a loop using the programmable function of the XY1 sensor simulator. When receiving the start signal from the computer, the sensor simulator generates a square wave signal. The computer simultaneously collects the output data of the weighing indicator for analysis. The response speed and noise level of the weighing indicator can be measured. If the weighing parameters of the weighing indicator are adjustable, tests can be performed under different parameters to obtain the relationship between the two. It is also possible to use the Fourier transform on the computer to analyze the spectrum of the noise of the whole scale of the automatic scaler, analyze the noise source, and improve the overall structure.
References [1] "Dictionary of Modern Electronic Science and Technology", Electronic Industry Press, first edition, June 1992.
[2] "AD7730 DATASHEET", Analog Devices, 1998.
Weigh indicator noise
ã€Abstract】This article discusses the definition, source, influence and evaluation method of weighing indicator noise.