| It is challenging to measure the speed of rotating bodies in harsh application environments such as automobiles. It is necessary to ensure the reliability and accuracy of the measurement while considering factors such as radiation interference, electromagnetic interference and conducted interference. This article comprehensively considers these practical environmental factors and introduces the implementation method and practical circuit design of gear speed measurement using TLE 4921-3U. Infineon Technologies has integrated two Hall sensors, a differential amplifier and evaluation circuits on the TLE 4921-3U. The device only evaluates the difference in field strength, not the absolute field strength value, which means that the influence of factors such as temperature drift, manufacturing tolerance and magnetic field environment on the measurement results can be minimized. If a high-pass filter with an external capacitor is used to dynamically process the measurement signal, the interference effect can be further reduced.
 The device is intended for use in harsh automotive environments. A small permanent bias magnet is required to sense ferromagnetic gears of various shapes. Correct switching ensures detection of minimum field strength differences between gear teeth and tooth gaps. For a 470nF filter capacitor, a typical lower switching frequency is about 10Hz. The TLE 4921-3U device has a package thickness of 1mm and only four pins.
Functional Design When the Hall device is in a constant magnetic field of any polarity, the two Hall sensors on it will produce the same output signal. No matter how large the absolute strength of the magnetic field is, the difference between them is always zero. However, since one unit faces the gear tooth where the magnetic field is concentrated and the other unit faces a tooth gap, if there is a magnetic field gradient between the two Hall units, a difference signal will be generated and amplified on the chip. In fact, this difference represents a small offset, which can be corrected by the corresponding integrated control mechanism. This dynamic differential principle allows high sensitivity to be maintained even when there is a large air gap between the sensor surface and the gear. As shown in Figure 1, a Schmitt trigger is used to digitize the signal, and an open collector output circuit provides the output signal. Overvoltage protection, reverse polarity protection, and anti-electromagnetic interference functions are also integrated on the chip, so it can be used in harsh environments commonly found in automobiles. Working methods The generation and evaluation of the difference signal can be explained here with the help of a typical application such as a ferromagnetic gear induction. A permanent magnet mounted on either side of the back of the device generates a constant bias magnetic field. The two Hall sensors in the device are 2.5 mm apart. If one sensor is temporarily facing a tooth and the other facing the tooth gap, the tooth will act as a flux concentrator, which increases the flux density through the Hall sensor and thus generates a differential signal. As the gear rotates, the polarity of the differential signal will change at the same speed as the change between the tooth and the tooth gap. When the zero crossing occurs directly in the center of the tooth or tooth gap, the maximum difference occurs at the edge of the tooth. When the difference exceeds the upper threshold, the output transistor of the TLE 4921-3U turns off. This is the case in Figure 2 where Hall sensor 2, closest to pin 4, senses the tooth. As the difference falls to the lower threshold, the transistor turns on, and this is the case where Hall sensor 1, closest to pin 1, senses the tooth. The integrated high-pass filter adjusts the difference signal to zero with a time constant that can be set by an external capacitor. In this way, only the differences that change at the lowest rate are evaluated (the lowest rate depends on the capacitor value). The output signal is not limited in the steady state and the accuracy achieved will allow for small switching hysteresis and large air gaps (maximum 3.5 mm). 
Gears, Sensing Distance and Angular Accuracy
A gear can be characterized by its module: m=d/z. Where d is the tooth diameter and z is the number of teeth. The distance between teeth is T, and the pitch is calculated as T=π×m. The maximum difference is sensed when one Hall sensor faces a tooth and the other Hall sensor faces a tooth gap. The device has a 2.5 mm spacing between the two Hall sensors, which allows the device to sense the difference at a module of 1, corresponding to a tooth pitch of 3.14 mm. If the module is greater than 3 or the gears are irregular, sufficient difference may not be detected for a long time, which means the output signal will be uncertain. The maximum distance allowed between the sensor and the gear is a function of temperature, module, magnet and speed, which can be characterized by the presence of a pulse at the output for each tooth/backlash transition. If the distance is reduced, a larger useful signal will be generated. Therefore, the switching accuracy can be increased by increasing the number of low/high transitions of the sensor, which can represent one rotation angle of the gear. Filter capacitor The filter capacitor plays an important role in the correction function of the Hall device. If an application needs to operate at temperatures above 100 degrees Celsius, ceramic capacitors (X7R) are recommended. The connection between the filter capacitor pin and the GND pin should be as short as possible. Leakage current at the filter capacitor pin will cause a drift in the switching threshold and false switching. The drift of the switching threshold can be calculated as follows:
Where I L , S C and R C are leakage current, filter sensitivity to ΔB and filter input impedance, respectively, which are given in the data sheet. Special attention should be paid to the selection of capacitors with high DC impedance and to their packaging. Leakage currents may appear on the printed circuit board between the connections or in a defective capacitor and may be a reason for sensor malfunction. Suppress transient interference in power supply lines Figure 3 shows the measurement circuit with the TLE 4921-3U. The filter capacitor CF ( 470nF) is connected directly to pin 4, a bypass capacitor (CS = 4.7nF) is added to the supply line and a 300Ω series resistor is used (400Ω for test pulse 5). Some applications do not allow such a high series resistor because of the resulting supply voltage drop. If a smaller series resistor of 50 to 150Ω is used, this will hardly have any effect on the following results, only on pulse 5. Due to its long duration of 400ms, this so-called load-dump pulse causes a high power dissipation in the TLE 4921-3U device. Without a sufficiently large series resistor, the device may be destroyed. Therefore, the minimum resistor must be adapted to the load dump requirements of the application in question and vice versa. An alternative approach is to add a suppression diode to the supply line so that a large series resistor does not have to be used.
Radiated interference The test was carried out in a TEM cell with an optimized printed circuit board with mounted Hall sensors. The measurement results show that the operation of the TLE 4921-3U is not affected in magnetic fields up to 160 V/m over the entire frequency range. To optimize the EMI performance of the TLE 4921-3U, the first thing to consider is the PCB layout. The following recommendations are based on EMI measurement results obtained while testing the device internally. Component parameter value C F =470nF High-pass filter capacitor
C S =4.7nF Optional high-frequency bypass capacitor
R P =0~330Ω Forms a low-pass filter on the power line with C P
C P =4.7nF Prevents conductive coupling and fast interference pulses
R q =33Ω Used with C q to smooth the falling edge of the signal
C q =4.7nF
R L =330Ω Load resistor The following are the optimization measures to reduce radiated interference:
1. Grounding (GND): The reference point on the circuit board is the GND pin of the Hall device. In order to avoid conducted interference, all connections to the GND pin should form a star connection, otherwise the anti-interference electromagnetic performance will be reduced.
2. Connection of the filter capacitor: The connection between the filter capacitor CF and the GND pin should be as short as possible (ideally, the location of CF is close to the Hall device), and the star connection structure to the GND pin mentioned above should be taken into account. Otherwise, it is recommended to use a second smaller capacitor (e.g. 82nF) between CF and TLE 4921-3U, the purpose of which is to shorten the connection between CF and the corresponding pins. This measure is only applicable when there is little available space near the Hall device.
3. Ground shielding: It is recommended to extend the GND connection of the filter capacitor as a ground shield for the connection of the capacitor to the C pin.
4. Additional RF bypassing: The RF bypass capacitor CS can further improve the anti-interference performance. The effectiveness of the optimization steps listed above (in descending order of importance) will vary depending on the specific characteristics of the system (sensors, cables, and control units). Not all of these measures need to be adopted, depending on the specific application requirements. Detecting rotation speed The output signal of the gear tooth sensor is a rectangular wave, where each change of the switch state represents a transition from a tooth to a tooth gap. For rectangular teeth (e.g. modulo 2) and a sensing distance of 1 to 2 mm, the duty cycle of the signal is almost 1:1. The form of the speed information depends on the specific application and may be in digital form or in the form of an analog voltage. 1. Simulation Evaluation
Speed control is the most common task in traditional control engineering. The controlled variable for analog controllers (proportional, proportional integral and proportional integral derivative) is a voltage proportional to the speed. The first step to obtain this voltage proportional to the speed is to convert the sensor output signal into a square wave signal with a fixed "on" time and a variable "off" time (depending on the speed) through a monostable edge trigger. The second step is to perform linear averaging and apply a conversion factor to make it proportional to the speed. A moving coil instrument is particularly suitable for analog display of speed, as it is an ideal averager at a low cutoff frequency (typically 10 Hz). 2. Digital evaluation
If the voltage proportional to the speed is generated in digital form or if a microcomputer is available in the system as a digital controller, the speed can be easily calculated. Connecting the gear sensor to a counting input of a microcontroller (e.g., the external input of Counter 0 on an 8051 microcontroller), the rotational speed of the gear can be determined by counting the number of high/low transitions of the sensor output within a certain time window (T window ). By carefully determining the width of this time window, the speed can be directly converted into a "revolutions per minute (rpm)" value without conversion. For example, the required time window for a gear with 15 teeth is 4 seconds. If one pulse is counted within this time window, this will correspond to 1 rpm. At the same time, this is also the highest resolution that can be achieved.
However, due to the high operating frequency of the microcontroller, it is difficult to set a long time window. If you choose a shorter time window, the count value must be multiplied by a correction factor, the ratio of which ideally matches the actual window. However, the measurement accuracy and resolution achieved is at most equivalent to this factor. Example:
A gear with 15 teeth has a time window of 4 seconds.
The actual time window is 40 milliseconds and the correction factor is 100.
If a pulse is counted within this set real-time window, the corresponding speed is 100 rpm.
If there is no pulse, the displayed value will be 0.
From this, it can be seen that the measurement limit can be determined by the selection of the time window. Inductive detection of rotation 1. Implementation with logic circuits
The detection of rotational induction can be simply realized with two sensors. These sensors should be placed on the circumference of the gear so that their output signals are 90 degrees out of phase. The switching order of these sensors is converted into a static direction signal by an edge-triggered D flip-flop, because according to the rotational induction, one sensor will switch earlier than the other sensor. The output signal of the dynamic gear sensor is only valid when the speed is above the minimum speed, and the same is true for the direction signal. So when a gear is braked and then starts to rotate in the opposite direction, the output signal near the stop point and the direction signal are not necessarily correct. 2. Implementation of the switching sequence in software
The evaluation can also be implemented by a microcontroller and software. The sensor signals are connected to two interrupt inputs of the microcontroller and the lower cutoff frequency can be monitored in software. If the sensor signals do not exceed the lower cutoff frequency, they are not evaluated. By Ernst Katmaier
Infineon Technologies | 
提升卡
变色卡
千斤顶
京公网安备 11010802033920号
