Resolver software decoding solution based on C2000
[Copy link]
In the "New Energy Vehicle Industry Development Plan (2021-2035)" (Draft for Comments) issued by the Ministry of Industry and Information Technology, it is proposed that by 2025, the sales volume of new energy vehicles will account for about 25% of new car sales, and the sales volume of new intelligent connected vehicles will account for 30%, and highly automated driving intelligent connected vehicles will achieve commercial applications in limited areas and specific scenarios. New energy vehicles are mainly powered by electricity and driven by electric motors. In order to obtain a better driving experience, engineers often need to know the current angular position and speed information of the motor, and provide corresponding torque and power in the algorithm. The driving environment of automotive applications is complex, and resolvers are sensors that are often selected for use in this application scenario.
Working principle of resolver:
A resolver is an electromagnetic sensor, also known as a synchronous resolver. It is a small AC motor used to measure angles. It is used to measure the angular displacement and angular velocity of the rotating shaft of a rotating object. It consists of a stator and a rotor. The stator winding serves as the primary side of the transformer and receives the excitation voltage. The rotor winding serves as the secondary side of the transformer and obtains the induced voltage through electromagnetic coupling. Usually, the secondary side uses two winding coils, which are placed on the rotor at 90° to each other, as shown in Figure 1.
Figure 1
In actual use, the rotor will rotate coaxially with the motor, that is, the angular speed and position of the rotor represent the corresponding state of the motor. If we apply a sinusoidal excitation signal VR to the stator, the AC energy will generate a magnetic flux Φ through the primary coil. Under ideal conditions, this magnetic flux will generate an induced voltage, VS and VC, on the secondary side. Then, according to Faraday's law of electromagnetic induction, the relationship between VS, VC and the angle Θ is as follows:
From this, we can know that if we know the real-time information of the applied stimulus VR and the obtained responses VS and VC, we can get the angle and speed information according to the above formula. After knowing the basic working principle of the Resolver, in order to obtain the angle and speed information and provide it to the DSP for algorithm reference, we need the following functional circuit resources to assist the Resolver to achieve the expected function:
DAC (Digital to Analog Converter) circuit: provides excitation sinusoidal signal VR. The excitation frequency is usually between 10kHz and 20kHz.
Boost circuit: Increase the voltage amplitude of the excitation signal. Usually the excitation signal received by the resolver is 4Vrms, 7Vrms, etc. At the same time, a common mode voltage needs to be provided to the system during the application process, so it is necessary to amplify the output signal of the DAC to a certain extent.
Excitation amplifier pre-stage circuit: Before the excitation signal output by the DAC is power amplified, it is often necessary to use an op amp to build a circuit to filter the DAC output and apply a common-mode voltage.
Excitation power amplifier circuit: amplifies the excitation signal driving capability. The specific driving capability depends on the specifications of the resolver. Usually 100mA~300mA is required.
Secondary side signal conditioning circuit: filters the rotor-sensed signal VS/VC and conditions it to a signal range acceptable to the ADC.
ADC (Analog-to-digital converter): As introduced in the basic principle, we need to convert the analog signals of VS/VC/VR into digital signals for RDC to calculate the angle and speed.
RDC (Resolver-to-digital converter): Executes the algorithm on the output and sensed digital signals of the rotor and stator, calculates the speed and angle information, and outputs it to the DSP's CPU for motor algorithm reference.
It can be seen that it is not an easy task to implement resolver decoding. TI has extensive experience in automotive and industrial motor solutions and provides a variety of solutions. This blog post will mainly introduce two widely used solutions. The first is a resolver software decoding solution based on the C2000 series DSP, and the second is a highly integrated resolver interface chip solution based on TI PGA411-Q1.
Figure 2 shows the hardware block diagram of the discrete resolver soft decoding based on the C2000 architecture .
Figure 2
Boost circuit: As mentioned above, in order to achieve the driving voltage of the resolver, it is usually necessary to amplify the excitation signal. In the development of electric vehicle applications, a two-level architecture is usually used. First, a 12V (low-voltage lead-acid battery) is converted into a 5V primary power supply. Then a BOOST boost power chip is used to convert 5V into a 15V (7-VRMS Mode) power supply. There are many choices here, and there are not many restrictions for this application. Excellent engineers often combine other applications and requirements in the circuit to find suitable power chips on ti.com. Here we recommend the first-level buck power supply LM63635-Q1 and the second-level boost BOOST power supply TPS61175-Q1.
Excitation amplifier pre-stage circuit: The EMI environment of automotive applications is complex. In order to ensure that the excitation power amplifier circuit is not interfered with, maintain signal integrity and distortion, and increase a certain common mode, engineers often need to use op amps to build excitation amplifier pre-stage circuits. The selection of op amps here mainly requires a wider bandwidth and a higher open-loop gain to ensure that the signal is not distorted. The OPA197 series op amp is recommended here. It has a 10-MHz GBW, and the OPEN-LOOP GAIN can reach 143dB, which can ensure the accuracy requirements of the resolver decoding system.
Excitation power amplifier circuit: The excitation primary coil of the resolver usually has a very low DCR (DC resistance is usually less than 100Ω), so a certain current output capacity is required to drive the resolver, usually 100-300mA. At the same time, in order to make the resolver obtain better accuracy and linearity, a higher SR (slew rate) is also required in the application here. The traditional solution is to use transistors to build a CLASS AB power amplifier circuit, which has a complex circuit, low reliability, and unsatisfactory cost and performance. In response to the design pain points of engineers here, TI developed the ALM2402F-Q1, a dual-channel op amp designed for resolver excitation applications. The ALM2402F-Q1 chip has the following features:
Very high current output capability, supporting up to 400mA continuous current output. Fully meets the needs of various resolvers.
3.4V/us SR. Ensures that the excitation signal is not distorted.
Built-in RF/EMI filter. Works better in complex noise environments such as inverters.
Using ALM2402F-Q1 can greatly reduce engineers' system BOM and improve system reliability. In addition, the current capability and SR provided by ALM2402F-Q1 can meet the needs of most resolvers. ALM2402F-Q1 will also launch products of the same series for resolver applications in the future, so please continue to pay attention to ti.com.
Resolver primary winding input signal and secondary winding output signal conditioning solution: As shown in Figure 3, in typical applications, we need to collect the primary excitation input signal of the resolver and the output signal of the secondary Sin/Cos winding, and convert the differential signal into a single-ended signal to provide it to the ADC for subsequent algorithm processing. Therefore, this part requires the op amp used to have differential signal input capability and in order to obtain a more accurate analog signal, the system requires the op amp to have a lower gain error and offset. In addition, it should be noted that due to the complex electromagnetic environment of automotive motors, in order to obtain more accurate sampling information, the op amp used here must have a higher CMRR (Common-mode rejection ratio). Engineers can go to ti.com to select a suitable op amp according to their application requirements. Here we recommend using TLVx197-Q1, TLC2272-Q1.
Figure 3
ADC, DAC&RDC: TI C2000 integrates a wealth of resources for developers to use. The resources mentioned above include 3 ADCs, 1 DAC, and RDC. In this example, TI C2000 TMS320F28069 is used. The TI C2000 microcontroller integrates up to 4 12-bit/16-bit ADC units and 3 12-bit DACs. The 12-bit ADC has a maximum sampling rate of 12.5Msps. The 32-bit C28x DSP core and coprocessor CLA can be used to implement the resolver decoding algorithm. TI C2000 integrates a wealth of resources for developers to use. Any C2000 product can implement the resolver decoding function, and the specific selection can also be combined with other requirements of the circuit being developed.
TI's discrete soft decoding solution has the advantages of small size, low cost, high accuracy, and flexible design. The powerful performance of TI's DSP C2000 processor can be directly used for motor control algorithms and drive implementation. For the discrete solution's resolver decoding front-end design, TI provides a system reference design, TIDA-01527. Smart engineers can go to ti.com and search for TIDA-01527 to download relevant information about the design.
| 
提升卡
变色卡
千斤顶
京公网安备 11010802033920号
